Chapter

2 Managing Money for Human Development Results

Author(s):
International Monetary Fund
Published Date:
May 2006
Share
  • ShareShare
Show Summary Details

The wave of new donor commitments in 2005 comes on top of sustained increases in global funding for education and health over the previous five years. For the lowest-income countries, especially, the stage is set for resource transfers on an unparalleled scale and the chance to make the Millennium Development Goals (MDGs) a reality of better lives for hundreds of millions of people. The challenge now is to turn resources into results. This chapter highlights how some countries are doing just that, and identifies the key constraints to faster and more equitable MDG progress elsewhere.

Broad regional trends of MDG progress have not changed significantly since publication of Global Monitoring Report 2005. All regions are off track on the child mortality goal and on at least some of the other goals. The two regions lagging most seriously behind—South Asia and Sub-Saharan Africa—are off track on all of the goals. Children’s nutrition is worsening in many parts of Africa, a majority of countries are not making sufficient progress on maternal mortality, and HIV/AIDS continues to spread across the world.

But the latest data on outcomes since adoption of the MDGs provide some encouraging signs of progress. A large set of countries, including many in Sub-Saharan Africa, have accelerated their progress on primary school completion in the past few years. Supplemental immunization programs in Africa have produced substantial declines in measles cases since 2000. Girls’ enrollments are growing faster than boys’ in every region, and the number of women elected to parliaments is increasing in many places. The first evidence that HIV prevention programs are beginning to work is emerging. And the number of AIDS patients under treatment in the developing world has increased rapidly: from less than 100,000 five years ago to 1 million in 2005.

Encouraging trends are also observed in policies and program design. Countries in the global partnership Education for All Fast-Track Initiative (EFA FTI) are beginning to see clear benefits from more harmonized donor practices. Insecticide-treated bed nets (ITNs) are being distributed on a larger scale than ever, and national malaria strategies are getting funding and visibility. Global partners are closely monitoring child survival where it is not improving. Immunization programs are making strides in reaching the poor. And in every region, countries are adopting policies to make education and, to some extent, health systems more effective and responsive to the people they serve. They are increasing community voice in the management of front-line schools and health facilities. They are allocating funds more transparently. They are beginning to link providers’ pay to performance. And they are conditioning income transfers to families on the families’ use of education and health facilities.

The latest data also confirm that external funding to support the health and education MDGs has increased sharply since 2000, and donor support appears better targeted to the lowest-income countries that have country-owned poverty reduction strategies and strong commitment to the MDGs. Recent studies also suggest that at least some developing countries have successfully managed to scale up immunization and schooling coverage while carefully managing unit costs.

Signs of progress are clear. But the world is still far from achieving the human development MDGs; donors and countries must keep working on ways to speed progress. This chapter analyzes five core challenges in ensuring that increased financing translates into faster and more equitable MDG progress. For donors, there is one overriding challenge: increasing the efficiency of aid. For developing countries, the key challenges are

  • maintaining efficiency as expenditures scale up rapidly;
  • reducing leakage to ensure that resources are used for intended purposes;
  • doing more to reach the poor; and
  • investing smartly to achieve larger impacts on human development outcomes through complementary investments in water, sanitation, roads, and housing.

Countries Making MDG Progress

Comprehensive data on country and regional progress toward each of the MDGs can be found in World Development Indicators 2006. This chapter identifies some of the countries making exceptionally fast progress toward the MDGs, and the success factors involved, as well as countries where MDG outcomes are worsening.

MDG 1: Nutrition

Malnutrition is an indicator and perpetuator of income poverty. The nonincome poverty target under MDG 1 is to halve, between 1990 and 2015, the proportion of people suffering from hunger, as measured by the percentage of children under five who are underweight. Only 34 of 143 countries are believed to be on track to meet this goal. Almost all of South Asia and much of Africa are off track, and in a number of African countries nutrition outcomes are worsening, reflecting the nexus between HIV and undernutrition.

Two countries making notable progress are The Gambia in Africa and Bangladesh in South Asia. While starting from—and still at—a very high level of child malnutrition, Bangladesh is the only country in South Asia that may achieve the nutrition goal. One factor is the country’s successful scale-up of community-based nutrition programs that work with mothers to improve feeding practices and promote infant and child growth. Tested in the early 1990s by the local nongovernmental organization (NGO), the Bangladesh Rural Advancement Committee (BRAC), the programs have received strong government backing and coordinated donor support since 1995 and are now mainstreamed into Bangladesh’s national health, nutrition, and population sector program. In the next phase nutrition programs will operate with a strong focus on results: disbursements are linked to performance, and good performance is rewarded with bonus funds from donors, many of which operate in Bangladesh through a sectorwide approach (SWAP) funding pool. Nutrition is also one of the six pillars of the Bangladesh Poverty Reduction Strategy Paper (PRSP) (Pelletier, Shekar, Du, and Kostermans 2005).

The strong performance of The Gambia over the last five years can be attributed to high-level political support (the National Nutrition Council and National Nutrition Agency are chaired by the country’s vice president); to development of institutional capacity at the national, divisional, and local levels for monitoring and addressing malnutrition; and to effective and well-targeted interventions. Programs focus on the window of opportunity between pregnancy and the first two years of life, when nutritional interventions have maximum impact on infants’ brain development. The program educates mothers, provides micronutrient supplements, and promotes infant and child growth through improved feeding practices.

MDG 2: Universal Primary School Completion

Estimates of the primary completion rate in developing countries show steady progress in most regions. Although about one-third of all developing countries are considered off track, and another 18 percent have inadequate data, the number of countries that have achieved universal primary completion increased from 37 in 2000 to 50 in 2004, and recent data suggest that the average pace of progress is increasing slightly. In 1990–7, the mean rate of improvement in primary completion was about 1.5 percent per year; since 1998, it has been 2.1 percent per year. Analysis by the secretariat of the EFA FTI, the global partnership to promote primary education progress, shows that the 17 countries that have joined the FTI are registering even faster improvement—about 3 percent per year. In two FTI countries, Niger and Guinea, for example, primary completion rates have been increasing three times faster than before 2002–3. Countries making the fastest progress—such as FTI countries Ethiopia and Mozambique and others such as Cambodia, Benin, and Rwanda—are exceeding the rates of improvement achieved by today’s industrialized countries at a similar point in their history.

Not all of this progress can be attributed to the FTI. But countries become eligible to join FTI by having a “credible” education sector plan agreed with donors, and donor commitment to harmonization and the mobilization of additional grant funding to reward policy and outcome progress are at the core of the FTI’s mutual accountability framework (box 2.1). Niger provides an example of the interaction between policy progress and increased and more flexible donor support (box 2.2).

While the quantitative progress is encouraging, universal primary completion is only a meaningful goal if it signals childrens’ command of a globally relevant level of skills and knowledge. Countries and donors need to give more attention to measuring learning outcomes, testing teachers for content mastery before they are hired, and making sure teachers have the materials and professional development support they need to be effective in the classroom. The FTI must stay focused on promoting completion with quality.

MDG 3: Gender Equality

The first MDG to fall due was a measure of progress toward the goal of gender equality and empowering women—the elimination of disparities in primary and secondary education by 2005. Although it was not met, girls’ enrollments are growing faster than boys’ in all regions, and the prospects for reaching the primary level target by 2015 are good. In South Asia and Sub-Saharan Africa, however, the primary completion rate for girls is still more than 15 percent lower than that for boys (World Bank 2004a). In 2015, it is projected that 21 of 133 countries—12 of them in Sub-Saharan Africa—will still have girls’ to boys’ primary enrollment ratios below 0.9 (Grown, Gupta, and Kes 2005).

Achieving the target of gender equality at other educational levels will be more difficult. The challenge of getting and keeping girls in secondary school is particularly severe. In South Asia, only 47 percent of girls go to secondary school, and in Sub-Saharan Africa, only 30 percent (Grown, Gupta, and Kes 2005). In developing regions as a whole, 80 girls for every 100 boys are enrolled in tertiary education (United Nations Statistics Division). The widest gaps are in Sub-Saharan Africa, where only 68 girls for every 100 boys are enrolled in university or other tertiary-level education, followed by Southern Asia, with 71 girls for every 100 boys. As for literacy, there was some progress in the 1990s in reducing the gender gap, but at the current rate, southern Asia, western Asia, and northern and Sub-Saharan Africa will not achieve the MDG target of parity by 2015 (United Nations Statistics Division).

BOX 2.1Education for All Fast-Track Initiative

Three years after its launch, the EFA FTI appears to have reached cruising altitude. It has an agreed operating framework, an established system of rotating donor leadership, a functioning secretariat, and donor collaboration mechanisms credited with producing tangible gains in harmonization at both the country and global level. Since early 2005, seven countries whose donor partners endorsed their education sector plans have joined the initiative (Djibouti, Kenya, Lesotho, Madagascar, Moldova, Tajikistan, and Timor-Leste), bringing the total to 20.

In its third year of operation, the $445 million EFA FTI Catalytic Fund (in 2003-7) has disbursed $75 million to nine countries that meet the criterion of having too few donors to close the financing gap in their agreed education plans. Five new donors made contributions to the Catalytic Fund, joining four earlier donors on its governing board. Total aid for education in the FTI countries over the past three years is estimated at $350 million a year, most of it—under the initial concept of a virtual fund—flowing through existing donor channels.

A new support facility—the Education Program Development Fund—was launched in 2005, with $30 million in contributions from five donor countries for 2005-7. The fund finances upstream and downstream technical support for countries with weak capacity to develop or implement sound sector plans—a core requirement for joining the initiative. Key beneficiaries are expected to be fragile states, and 25 countries are already benefiting from its support.

At the December 2005 EFA FTI meeting in Beijing, it was agreed that the initiative will expand to as many as 40 additional countries over the next two years and that the Catalytic Fund may need to shift to longer-term financing (beyond the current three-year limit). In addition, the steering committee was expanded and specific actions on harmonization were requested of all members.

Although the initiative will face challenges during 2006, with a change in the secretariat and a continuing need for additional financing, it has clearly emerged as a key vehicle for financing, technical support, and donor harmonization in education.

Source: EFA FTI Secretariat.

Progress in increasing women’s share of nonagricultural wage employment and holding parliamentary seats is notable. Women’s share of the labor force has risen in almost all regions (United Nations Statistics Division), but women are still at a disadvantage in labor markets. Their participation is often restricted by onerous time burdens that result from a very unequal division of tasks in the household and limited infrastructure for child care and other household duties (Blackden and Wooden 2006). The earnings gap between men and women is shrinking, but in developing countries women still earn on average about 30 percent less than men (World Bank 2001). Unemployment rates are consistently higher for women workers (ILO). Occupational segregation is pervasive; women frequently are confined to traditional female occupations that pay low wages (World Bank 2001). Given the current trend in gender equality in political participation, it may be difficult to reach the MDG 3 target of having 30 percent or more of national legislative seats held by women by 2015. By January 2005, only 17 countries had met this target, and globally, the proportion was only 15.9 percent, up from 13.5 percent in 2000 (United Nations Statistics Division).

Notwithstanding the overall slow progress, there are many examples of innovative programs that are helping countries progress toward gender parity, particularly in education, but also in access to economic resources and political participation.

Education. In Bangladesh the emphasis that government and other actors have placed on girls’ education since the 1990s has changed the public discourse and the pattern of education. This emphasis, combined with large-scale stipend programs to reward girls for going to school and sustained expansion of schooling supply, which included attention to latrines, wells, and female teachers, has closed the gender gap in both primary and secondary education. By 2002 girls’ enrollment in primary school was 100 percent, and girls’ attendance and performance on achievement tests surpassed those of boys.

BOX 2.2Jump-starting progress on primary completion in Niger

Five years ago in Niger, only one of every five children completed primary school, one of the lowest completion rates in the world. The government struggled to expand schooling, devoting 80 percent of its education budget to primary schooling, but it received little external support and could not afford to hire more than 250 new teachers per year. But by 2005, Niger was one of the world’s best performers in terms of progress in bringing children through its primary education system. What changed?

A key factor was a politically difficult reform of teacher salary policy: the government froze the recruitment of civil service teachers and promoted system expansion by hiring new “contract teachers” on shorter-term, renewable contracts at a lower salary level, on a par with the average teacher salary across low-income countries. Since then, teacher hiring has jumped from 250 to 2,500-3,000 per year. The 16,000 new contract teachers have more than doubled the teaching stock and made the unit costs of primary education more fiscally sustainable. School enrollments have doubled—from 530,000 to 1.1 million children—representing a 16.6 percent yearly increase. Children in rural areas have been the biggest beneficiaries; in these areas enrollments increased from 38 percent to 51 percent between 2002 and 2005, closing the gap with the national average.

The government’s courageous reform has increased the primary completion rate from 20 percent in 2000 to 36 percent in 2005 and reduced geographic disparities. Official development assistance (ODA) to Niger has tripled, from about $10 million to $39 million per year, since it joined the EFA FTI in 2002, and donors have made concrete progress in merging missions and using common performance indicators. The government’s education sector plan, agreed with donors under the FTI process, is tackling key issues, such as lagging girls’ completion rates, by introducing targeted stipends and trying to improve school functioning with local school management committees and a new information tool: performance “monitoring sheets” that compare schools’ resources and results are posted in each school.

Source: EFA FTI Secretariat.

Mauritania has also made impressive progress—increasing the primary enrollment ratio for girls from 39 percent in 1990 to 85 percent in 2001—by expanding the supply of schooling, recruiting female teachers, and offering girls scholarships and school meals. But at both the primary and secondary levels, boys still perform better in end-of-cycle exams, and the repetition rate for girls is higher.

Access to economic resources. In Argentina, the female share of nonagricultural wage employment rose from 36 percent in 1990 to 48 percent in 2003. Much of the change can be attributed to vocational and technical training programs targeted to women, such as FORMUJER, and to the 2003 Heads of Households program, which paid female heads of households with children under 18 years old for community work.

In Mali, the introduction of small-scale, multifunctional diesel engines that can provide electricity, pump water, and mechanize grain milling and other tasks has improved women’s economic situation, reduced their work burden, and promoted development and poverty reduction in communities. Between 1999 and 2004, 400 diesel engines were installed, reaching 80,000 women in 10 districts. In the first five years of the project, the engines saved women 1 to 3.3 hours of labor per day; the girls-to-boys ratio in grade 5 improved; the proportion of girls entering secondary school rose from 31 percent to 38 percent; and visits to local clinics for injuries caused by burdensome work were reduced. Women’s associations own, manage, and maintain the engines and sell energy services. Women have increased their income from an average of $68 to $122 per year, and the number of women earning at least 150,000 CFAF (West African Francs) increased by a factor of 10 (Modi and others 2005).

Political participation. A few years ago Morocco had the lowest rate of female representation in the Arab world: 0.6 percent. Today 35 female parliamentarians make up 11 percent of the parliament. Five years of research and advocacy by civil society resulted in an implicit 20 percent quota system, applied voluntarily by political parties at the last parliamentary elections.

Another impressive case is Rwanda, currently at the top of the world ranking for women’s political leadership; there, women comprise 49 percent of the National Assembly. Rwanda’s new constitution reserves 24 assembly seats for women. But 15 additional women were elected to nonreserved seats, bringing the total to 39 women in the Lower House (of 80 members). In the Senate a constitutional quota of 30 percent has also been exceeded.

MDG 4: Child Mortality

More than 10.5 million children under the age of five die each year from preventable and treatable causes—4 million during the first month of life and 3 million during the first week. Simple, known, and low-cost treatments for childhood respiratory and diarrheal diseases could keep an estimated two-thirds of these children alive. The MDG for child mortality calls on countries to put these interventions to work to achieve a two-thirds decline from the 1990 baseline by 2015, a reduction of 4.3 percent a year.

A recent study shows that this target rate of improvement is very ambitious compared with the average long-term country experience in improving child survival, and indeed, most low- and middle-income countries today are not making enough progress to reach the goal (Eifert and Gelb 2005). Some 117 (79 percent) of the 148 developing countries for which data are available are estimated to be off track on MDG 4. While the majority of these countries are making some progress, seven countries have seen no improvement since 1990 (box 2.3), and in 15 countries the share of children who die before age five has increased. The last group includes conflict-affected countries, such as Cambodia, Central African Republic, Côte d’Ivoire, Iraq, and Rwanda, and countries hit hard by HIV/AIDS, such as Botswana, Kenya, Lesotho, South Africa, and Swaziland.

BOX 2.3China’s slow progress on child mortality

China, which had one of the world’s most spectacular records of sustained child mortality improvement before 1990, is now considered at risk of not reaching the child mortality MDG. Between 1960 and 1980, China reduced child mortality from 225 to 64 per 1,000 births, a rate of -6.3 percent per year and well above the MDG target rate. Since 1990, however, progress on under-five mortality has slowed substantially—despite a huge increase in economic growth and strong improvements in nutrition, including among children—factors that should greatly improve child survival. Although the most recent estimate of China’s child mortality progress (-3.3 percent per year) is better than progress in many other countries, it is far below the progress of top performers, including most other East Asian countries.

National data also point to widening gaps in under-five survival between richer and poorer provinces, and between boys and girls. China is one of just seven countries in the world where a girl’s risk of dying by age five is higher than a boy’s, and it has the largest gender gap by far. New research suggests that infant and under-five mortality rates among girls in China may actually be rising. While improvements in male infant mortality continue to drive China’s overall child survival progress, female infant mortality is increasing by nearly half a percent per year.

Source: World Bank East Asia Region 2005.

The overall picture is highly troubling. But there are some encouraging factors. First, some countries are sharply increasing child survival (table 2.1), and there are lessons to be learned from their experience. Second, post-2000 data provide some evidence of accelerating rates of progress. Third, led by the Global Alliance on Vaccines and Immunizations (GAVI) (box 2.4), immunization coverage has expanded sharply in many countries over the past few years, producing, for example, a 91 percent drop in measles cases in 19 Sub-Saharan African countries since 2000 and pointing to possible further impacts on child survival rates in the coming years (Otten and others 2005).

TABLE 2.1Sharp increases in child survival for some countries
Under-five mortality rate per 1,000 births
Annual percent change
19901995200020041990-2004
Low-income
Vietnam53443023–5.9
Timor-Leste17214510280–5.4
Bhutan16613310080–5.2
Mongolia108876552–5.2
Lao PDR16313110183–4.8
Sub-Saharan Africa
Eritrea1471229782–4.2
Comoros1201008270–3.8
Cape Verde60504236–3.6
Mozambique235212178152–3.1
Guinea240208175155–3.1
Middle-income
Czech Republic131054–7.7
Egypt, Arab Rep. of104714936–7.5
Peru80604229–7.2
Macedonia, FYR38261814–7.1
Syrian Arab Republic44312216–7.0
Source: World Bank 2006.
Source: World Bank 2006.

BOX 2.4Global Alliance for Vaccines and Immunization

The Global Alliance for Vaccines and Immunization (GAVI) links private and public sector partners to save children’s lives and protect health through the widespread use of vaccines. By the end of 2005 it had made financial commitments totaling $1.4 billion over five years to 72 of the 75 eligible low-income countries. Immunization coverage in these countries has increased from 57 percent in the 1990s to an estimated 65 percent in 2003. GAVI’s strong record has helped attract new funds. It is working with partners to develop the new International Financing Facility for Immunization; commitments in 2005 will yield an additional $3 billion in disbursements to governments before 2015.

While the largest declines in child mortality since 1990 have occurred in the middle-income countries, the experience of many countries demonstrates that appropriate combinations of policies and technologies can produce big gains in child survival, whether the starting level of child deaths is very high, as in Guinea, or very low, as in the Czech Republic.

MDG 5: Maternal Health

Cross-country data on maternal mortality rates are not readily available, so the best indicator is the share of women with access to trained birth attendants at delivery. The latest data, presented in table 2.2, show strong progress in East Asia and the Pacific since 1990 and good coverage in Latin America and the Caribbean. But in Sub-Saharan Africa there has been almost no improvement, and in South Asia the coverage levels remain low, despite significant improvement. Within South Asia the improvement is heavily driven by Bangladesh, which has achieved a very large and sustained increase in women’s access to skilled attendants at birth.

TABLE 2.2Skilled attendants at delivery, by region, 1990 to 2003
RegionPercent of births covered by the data19902003Percentage change

1990-2003
Sub-Saharan Africa6140413
South Asia97273842
East Asia and Pacific80457668
Latin America and the Caribbean70748616
Total developing countries75415738
Sources: Data from Demographic and Health Surveys (DHS), Multiple Indicator Cluster Surveys (MICS), and comparable surveys, weighted by number of births.
Sources: Data from Demographic and Health Surveys (DHS), Multiple Indicator Cluster Surveys (MICS), and comparable surveys, weighted by number of births.

Trends in coverage rates, for key maternal and child interventions, are important for tracking, whether effective interventions are being implemented. The “Countdown to 2015” working group established by the new global Partnership for Maternal, Newborn, and Child Health (PMNCH) is tracking coverage rates for key nutrition, vaccination, and other child health interventions in 60 priority countries where the under-five mortality rate exceeds 90. These data, measured more frequently and reliably than the MDG outcomes, provide a better indication of whether interventions are sufficient and effective. The Countdown working group is also following countries’ adoption of key policies that provide the basis for effective child health programs (box 2.5).

BOX 2.5A new global partnership for the health of mothers, newborns, and children

The new global Partnership for Maternal, Newborn, and Child Health brings together three existing partnerships to strengthen and accelerate the response to MDGs 4 and 5 and to provide a unifying framework for action. With more than 80 organizations and partners, it increases resources, supports national planning processes, and promotes donor convergence at the country level. It also provides leadership in interactions with other relevant players, including the Global Alliance for Vaccines and Immunization; the Global Fund to Fight AIDS, Tuberculosis, and Malaria; and the Roll Back Malaria Partnership.

MDG 6: AIDS, Tuberculosis, and Malaria

Both the number of people living with HIV (40.3 million) and deaths from AIDS continued to increase in 2005. Sub-Saharan Africa remained the most affected region, with 63 percent of all people living with HIV. While no region has yet achieved a declining rate of new infections, recent data suggest that a few countries have begun to do so. These countries provide the first indication that determined government action can halt the spread of AIDS. New evidence suggests that prevention programs initiated some time ago are finally bearing fruit in Zimbabwe, the first documented decline in southern Africa: HIV prevalence appears to have fallen from 26 percent in 2002 to 21 percent in 2004. Haiti’s epidemic, one of the oldest, could also be turning a corner: HIV prevalence among pregnant women there fell from 6.2 percent in 1993 to 3.1 percent in 2003–4; the most pronounced decline was in urban areas.

Increased AIDS mortality may explain much of these downward trends. But evidence that decreased HIV prevalence also reflects real successes of HIV prevention programs comes from recent comparisons of prevention activities across regions in Tanzania and across the Uganda-Kenyan border (box 2.6).

Central to successful programs are interventions of sufficient intensity, as highlighted by Zimbabwe. Despite an unfavorable macro-economic and political context, the country has apparently maintained one of the highest levels of condom use in Africa (86 percent among men, and 82 percent among women for 2000–5). Sales data suggest that high condom use started in the mid-1990s. With other key behavioral changes, including sharp reductions in the number of sexual partners, condom use appears to have lowered the overall prevalence rate (UNAIDS 2005). This finding echoes an earlier analysis that examined condom effectiveness in a general population with HIV prevalence in rural Rakai, Uganda: regular condom use significantly reduces the incidence of HIV (Ahmed 2001).

BOX 2.6HIV prevention works when it is intensive and sustained

That HIV prevention can work is illustrated by comparing the Mbeya region in southwest Tanzania to the Rukwa region in the west. Prevalence rates among pregnant women aged 15-24 increased in both regions from 1988 to 1994-5. They then started to decline in Mbeya—from 20.5 percent in 1994 to 14.6 percent in 2000—but continued to increase in Rukwa. A key factor appears to have been comprehensive prevention programs launched in Mbeya beginning in 1988 with German technical support. These programs led to a steady increase in the use of condoms, treatment of sexually transmitted infections, and observed behavioral changes. Rukwa, with limited resources and no external support, made little effort at prevention—and there has been no documented decline in HIV prevalence there.

Surveillance site data from western Kenya and eastern Uganda also suggests that policies matter. In the early 1990s, HIV prevalence was similar along the border: all 11 sentinel sites in the zone had rates above 10 percent. Since then, however, trends have diverged. By 2000 prevalence rates at the Kenyan sites exceeded 10 percent and varied from 10 percent to 35 percent, but in eastern Uganda, all six sites were below 10 percent, and five of the six were in the range of 3–6 percent. Many analysts believe high-level political commitment in Uganda has made the difference: Uganda’s National AIDS Commission was established in 1986 within the Office of the President; Kenya took the same step only in 1999.

Uganda has also made extensive use of NGOs as well as government health services for comprehensive prevention programs, including voluntary counseling and testing (VCT). While there is some recent evidence that the broad-based approach to HIV prevention and care launched in Kenya after 1999 is also beginning to produce results, its later start appears to show in the different trends between the otherwise similar border populations of western Kenya and eastern Uganda through 2000.

Sources: Jordan-Harder and others 2004; Moore and Hogg 2004.

BOX 2.7Global Fund to Fight AIDS, Tuberculosis, and Malaria

In 2005 Round 5 of the Global Fund to Fight AIDS, Tuberculosis, and Malaria approved new financing for 20 countries totaling $382 million, the lowest amount so far and substantially below the 2004 Round 4 approvals of more than $1 billion for 52 countries. Round 5 proposals included HIV/AIDS ($128 million); tuberculosis ($146 million); malaria ($67 million); and, for the first time, health-system-strengthening activities in two countries ($41 million).

Total approved financing for the Global Fund since Round 1 in 2002 has been $4.9 billion; $1.9 billion was disbursed by December 2005. Global Fund financing has provided funding for the distribution of 7.7 million insecticide-treated bed nets to combat malaria, for the treatment of 1 million tuberculosis cases under DOTS (recommended tuberculosis treatment regime), and funding for nearly 400,000 people with HIV for antiretroviral treatment (by December 2005).

The number of people on antiretroviral treatment more than doubled from 400,000 in late 2003 to about 1 million by end-2005. Coverage now exceeds 80 percent in Argentina, Brazil, Chile, and Cuba. As a result of the scale-up in treatment, between 250,000 and 300,000 deaths were averted in 2005, but the full effects will only be seen in later years.

The World Health Organization estimates 350 million to 500 million clinical episodes of malaria per year. Evidence is accumulating that ITNs can reduce malaria deaths. The United Nations Children’s Fund’s Accelerated Child Survival and Development Initiative, started in 11 countries in West and Central Africa in 2002, and household surveys in 2003 confirmed significant increases in the use of ITNs. Although it is too early to see impacts on the under-five mortality rate, the program has focused on the hardest-to-reach districts and has proved that progress is possible against the odds. A similar program in Ghana sharply increased the use of ITNs from less than 5 percent to more than 75 percent.

Togo launched a potentially important new model of intervention in December 2004. The country conducted a major national campaign of ITN distribution combined with measles and polio immunization and deworming. Preliminary results indicate that the campaign increased possession of ITNs from 6 percent to 62 percent, averaged across all households. An estimated 98 percent of households with a child under five years of age now have at least one ITN, and 95 percent of households received it from the distribution campaign. The combined vaccination/bed net distribution appears to be a promising strategy for increasing cost-effectiveness.

MDG 7: Environmental Sustainability (Water and Sanitation Targets)

Although infrastructure in many regions has expanded rapidly in absolute terms, population growth has increased even faster. Only two regions—East Asia and Latin America and the Caribbean—are on track to meet the MDG targets for water and sanitation, which are crucial for progress on the health MDGs, as well as being important in their own right. Europe and Central Asia has gone backward—from nearly universal coverage in 1990 to eroded service and unsafe water quality as a result of protracted maintenance failures. Sub-Saharan Africa has had the slowest progress: only an estimated 64 percent of the population has access to safe water and 37 percent to improved sanitation.

The poorest quintiles in every region have the least access to water and sanitation, but the biggest divide is geography. Even in the lowestincome countries, 83 percent of urban dwellers have access to water, compared with only 55 percent of rural dwellers. In lower-middle-income and upper-middle-income countries, more than 93 percent of the urban population has water access and more than 84 percent to sanitation, whereas coverage in rural areas is at least 20 percentage points lower.

Taking advantage of economies of scale and scope in water, sanitation, and housing provision in urban areas is key, as is using small-scale local providers in rural areas. As noted in chapter 1, Africa has been at the forefront in developing new ways to provide sustainable services to very poor and isolated populations. It has established community-managed village water supplies and private cleaning services for latrines. Since the government loosened its monopoly on cesspit cleaning, households in Dar es Salaam, Tanzania, have been able to choose among providers on the basis of price and performance.

Turning Resources into MDG Results

Global Monitoring Report 2005 focused on the constraint that human resource availability poses for many low-income countries attempting to accelerate MDG progress. It concluded that leveraging scarce skills (through contract teachers or community health workers, for example) was helping some countries overcome the provider constraint and allowing them to scale up service delivery. However, managing migration—especially of health manpower—dealing with the impact of HIV/AIDS on provider numbers, and expanding training systems all remain difficult challenges.

This report focuses on financing. The latest available data begin to offer an empirical basis for analyzing key questions on the financing constraints to MDG progress. How much funding is being mobilized? How efficiently is it being used? And what are the biggest challenges to improving results for investment?

External Financing

The latest data show that ODA commitments for education and health have increased substantially since the MDGs were adopted in 2000 (figure 2.1); 2004 saw the sharpest rise yet. Total assistance for health in 2004 is estimated at $11.4 billion and that for education, $9.5 billion. Within education, support for primary education has had the largest increase, and commitments to low-income countries have risen much more than aid to middle-income countries. In 2000–4, commitments for primary education in low-income countries rose 175 percent in constant dollars, compared with 87 percent in education ODA overall. Fifty percent of education support in low-income countries is now for primary education, compared with about 33 percent in the late 1990s. The realignment toward the MDG agenda has been clear.

FIGURE 2.1Development assistance for education and health

Source: OECD DAC for education and health commitments; HIV/AIDS estimated commitments from Lewis (2005a).

Note: Values for HIV/AIDS for the years 2001 and 2003 are interpolated.

ODA for health is increasing greatly as well. Since 2000 new commitments have grown 83 percent in constant dollars. The share for primary care has declined, however, from about 28 percent of total health assistance in 1999 to 15 percent in 2004. The fastest-growing segment of health assistance is for HIV/AIDS programs, up from about 14 percent of development assistance for health in 2000 to more than 30 percent in 2004. There are also important institutional changes: new private financing sources, notably the Gates Foundation, have become major players. Gates Foundation support for health has averaged about $500 million a year over the past five years. And the launch of some 70 global health partnerships over the past 10 years has also changed the landscape: in 2004 they delivered more than 20 percent of total health assistance (Gottret and Schieber 2006).

Developing-Country Spending

Fiscal data remain incomplete for 2003. Most major countries are missing from International Monetary Fund (IMF) data, the only source of standardized cross-country data on (central) government education and health spending, both in relation to GDP and as a share of the overall budget. Data available for 2003 cover only 21 of 79 middle-income countries and 27 of 57 low-income countries. For this small sample, however, the data do show some significant upward trends.1

While average spending on education across the sample has not increased, in the subset of low-income countries for which comparable data from 2000 and 2003 are available, spending on education as a share of GDP has increased somewhat since 2000, in some cases significantly. Public spending on health has not, however (figure 2.2). But public spending on health, especially in low-income countries, is generally less than half of total national health spending, most of which is private, out-of-pocket spending on private health services and pharmaceuticals. For the small sample of middle-income countries covered by these data, spending on both sectors increased as a share of GDP, especially for education.

FIGURE 2.2Developing-country spending on education and health

Source: IMF.

Note: Unweighted averages based on corresponding available data for 2000 and 2003.

As a share of the central government budget, education and health expenditures increased slightly between 2000 and 2003 in every region except the Middle East and North Africa, and for education, in Sub-Saharan Africa (figure 2.3). Countries registering particularly high increases in the budget shares for education and health include Angola, Bhutan, Botswana, Burkina Faso, Chile, Guyana, Kenya, Kyrgyz Republic, Madagascar, Malawi, Moldova, Peru, Senegal, Sierra Leone, Vietnam, and Zambia. Nevertheless, all but a handful of countries fall short of the Fast-Track Initiative target of 20 percent of the budget for education: shares average around 15 percent. For spending on health, only three countries in the entire sample—Chile, São Tomé and Principe, and Turkmenistan—even approach the target of 15 percent of the budget set by African ministers in 2002. While the general trend is up, a few countries have seen large drops in spending shares for education and health: Cameroon, The Gambia, and (to a lesser extent) Mozambique.

FIGURE 2.3Share of total government spending for education and health, by region

Source: IMF.

Note: Unweighted averages based on corresponding available data for 2000 and 2003.

The incompleteness of the data makes all observations tentative. It also poses a real issue for efforts to link resources to results in monitoring MDG progress, at least in the human development sectors. The current data platform is wholly inadequate.

Both the IMF and the World Bank expend significant resources collecting and analyzing government revenue and expenditure data. The IMF focuses mainly on fiscal aggregates; the Bank analyzes subsectoral and sub-national spending patterns and spending effectiveness in the context of its public expenditure reviews and sector studies. Yet no systematic, cross-country database unites these data. The only available series is produced by the IMF, and it has significant country gaps and no data on subsectoral or sub-national social spending. It is impossible to say today, for example, whether increased ODA for primary education in Sub-Saharan Africa has been reflected in any increase in government spending on primary education across that region. Meaningful efforts to determine whether increased resources are producing results depend on such data.

The Development Assistance Committee (DAC) of the Organisation for Economic Cooperation and Development (OECD) has made good progress over the past two years in upgrading the quality and timeliness of cross-country data on ODA, in part in recognition that these data are crucial for monitoring progress in implementing the Monterrey accords and donor commitments related to the MDGs. A parallel effort by the Bretton Woods institutions to ensure common standards and classifications in collecting government expenditure data and to unify these in a single, standardized cross-country database for use by both institutions—as well as others—is sorely needed in a world where not just results, but the efficient use of resources to produce results, is a core policy interest.

Spending Money Effectively—Five Key Challenges

Maintaining Efficiency as Spending Scales Up

Evidence from OECD countries shows that major increases in public expenditures in sectors such as education often translate mainly into higher unit costs rather than increases in output. In the United States, for example, researchers have documented that a tripling of real education spending per student since 1960 has all been absorbed by higher teacher salaries and lower class sizes and has had no measurable impact on either student numbers or average student learning levels (Hanushek 2003). Given the large increase in spending needed to scale up education, health, and water and sanitation coverage in many developing countries and documented institutional weaknesses, it is important to track how well increased spending is managed, including its impact on input prices and unit costs.

A recent study called for more attention to this issue in the MDG context and analyzed developing-country experience over the 1980s and 1990s in three sectors: primary education, childhood immunization, and road maintenance (Roberts 2005). While unit costs for education rose in six of the eight countries studied, those for road maintenance were constant or declining, and those for childhood immunization were stable in one country (Bangladesh), despite a massive expansion of program coverage; in another case, Nepal, data were inconclusive.

The study found that Bangladesh had expanded immunization coverage from virtually zero in the 1980s to 70 percent in the 1990s with no increase in unit costs. Although unit costs did fluctuate annually, with lumpy expenditures on equipment and vehicles, the program gradually penetrated hard-to-reach areas over a 12-year period, and the unit cost remained virtually the same as the average cost. The explanation appears to be the combination of falling real prices for the vaccines delivered2 as the number of global suppliers increased; reductions in wastage; and improvements in staff productivity, which exceeded increases in salaries over the period.

For education, unit costs tended to increase as overall spending rose. Not surprisingly, given that teachers’ salaries represent about 80 percent of education system costs, the prime driver was rising real salaries. Interestingly, however, the increases did not appear to reflect any short-run rigidities in the supply of new teachers, but rather the political necessity of reversing extended periods of real wage decline. In all six countries where unit costs increased throughout the period (Bolivia, Ghana, Mozambique, Uganda, Yemen, and Vietnam), severe previous wage compression leading to concerns about the quality and motivation of teachers was the main rationale: wage increases were considered essential to protect long-term education quality and sustained enrollment growth. Concomitant increases in the per-student costs of school construction, learning materials, and teacher training were observed, but these increases were much smaller.

In some phases of the expansion in all of the countries, unit costs declined as a result of “one-off” efficiency measures such as the introduction of double-shift schools or the reallocation of excess administrative staff at headquarters to classrooms. Rising pupil-teacher ratios in several cases also kept costs stable for some period, as in Uganda after 1996—but with a marked deterioration in quality, as reflected in declines in student test performance. Only in two of the six countries—Ethiopia and Mauritania—was the education system able to maintain rapid expansion with declining marginal costs over a fairly extended period, and in both cases the expansion appeared to be linked to initially high average teachers’ salaries relative to the market for skilled workers. Even in these cases, however, erosion of perceived quality eventually became a political concern and led to wage policy reversals.

A new effort to extend this analysis takes a slightly different approach (Bruns, Mariscal, and Gacougnolle forthcoming). Instead of focusing on the countries with the fastest spending growth, it examines the countries with the best outcome performance—the 10 countries with the fastest improvement in primary completion rates in the period 1997–2003. Primary completion progress signals not only that education systems are rapidly expanding enrollments, but also that they are managing to maintain reasonable internal efficiency and at least minimally adequate quality—otherwise dropout rates will increase.

The picture that emerges is consistent with Roberts’ (2005) education sector analysis. For at least some period, most of the 10 countries achieved rapid enrollment expansion with no increase in real spending per student (figure 2.4). Indeed, between 1990 and 2000, unit costs declined in all the countries except one (The Gambia), as enrollment growth outstripped budget growth. Since 1999, though, several countries have experienced significant real increases in education spending, which has allowed unit costs to rise in four countries and to remain constant in two. In the remaining four countries, however, unit costs continued to decline through 2002. Almost all of the countries have seen real budget growth, but enrollments have continued to rise faster. Interestingly, the country with the largest increase in unit costs between 1999 and 2002 is Ethiopia, which according to the new analysis and Roberts’ study had a long period of unit cost compression.

FIGURE 2.4Education unit costs in best-performing developing countries, 1999–2002

How long unit costs can decline before major negative effects on quality are observed is related to the starting level of system efficiency. In this sample of low-income countries, primary education unit costs around 2000 averaged about 12 percent of per capita GDP but ranged widely—from 35 percent in Niger to 4 percent in Cambodia. This disparity helps to explain why Niger’s policy actions to make the costs of primary schooling more sustainable (box 2.1) have been such a key factor in the country’s recent progress in increasing enrollments and completion.

Conversely, the increases in unit costs in Cambodia since 1999 have clearly been necessary to improve quality. For Côte d’Ivoire, Ethiopia, The Gambia, and Mali, despite the sustained period of unit cost compression in the 1990s, the unit cost in primary education around 2000 was still slightly above the low-income country average. However, cost compression has continued in the past few years in most of these countries—raising the question of long-term sustainability of enrollment and completion rate progress if budgets do not begin to keep pace with system expansion.

Under the EFA FTI and GAVI, data on country expenditures, service delivery, and unit costs are beginning to be tracked more systematically. This tracking should provide a stronger basis for monitoring the efficiency of spending as aid flows and domestic spending increase—and for learning from the countries that manage costs and quality best.

Making Aid More Efficient

The increased volume of ODA for education and health is important. But its impact will be blunted unless it can be used for the core expenditures countries incur to scale up service delivery in education and health. All MDG costing exercises have found the incremental financing needs in health and education to be largely recurrent expenditures, above all recurrent salary costs for health workers and teachers (Devarajan, Miller, and Swanson 2002; Bruns, Mingat, and Rakotomalala 2003; United Nations Millennium Project 2005). Although even the poorest countries need eventually to sustain their education and health systems with domestic resources, many can achieve this goal only gradually, as their economies grow and fiscal capacities deepen.

The first issue is the flexibility of external financing. A high share of today’s ODA for health and education is transferred in forms that cannot be applied to core budgetary outlays. A recent study of 14 countries receiving poverty reduction support credits found that only 20 percent of donor commitments were provided as either general or sectoral budget support (Foster 2004). Fully 50 percent of assistance ran outside the budget, and 30 percent escaped government reporting altogether. If only 20 cents on each dollar of aid can be used to fill core financing gaps, true MDG financing needs might be five times those estimated to date.

With respect to education, the disconnect between countries’ need for flexible budget support and the high share of bilateral aid transferred as technical assistance is large. As figure 2.5 shows, bilateral donors’ reports to the DAC indicate that at least 30 percent of all education funding is spent on consultants, studies, or training. The 2004 DAC data are not complete, but they show a small decline in the average share of donor aid transferred as technical cooperation and suggest that a few donors, such as the United States, may be trying to reduce this share. On a growing volume of aid for education, this represents some modest progress.

FIGURE 2.5Share of bilateral education ODA commitments reported as at least half technical assistance

Source: OECD DAC database.

Technical assistance can provide needed analyses, capacity building, and practical experience. But it is also often badly coordinated among donors and poorly prioritized; rarely do donors compare study proposals to the alternative that, in some countries, 100 days of consultancy support could equal the annual cost of paying 100 teachers or keeping 5,000 children in school. The EFA FTI’s new Education Program Development Fund attempts to tackle this issue by pooling donor resources for planning support and studies and by ensuring that these resources respond to government priorities. But it is too early to judge whether this effort produces a net reduction in technical cooperation as a share of donor support to these countries—or clear increases in coherence and quality.

A second issue is predictability. In scaling up health and education services, developing-country governments must take calculated risks that donor promises of future aid will materialize to help fund long-term recurrent spending obligations. Yet studies of overall aid flows show that commitments are highly volatile and that disbursements are poorly correlated with commitments, generally lower than commitments, and no less volatile. Worse, volatility tends to be procyclical, making it even harder for countries facing other revenue shocks to smooth recurrent expenditures, and volatility rises with aid dependence (Eifert and Gelb 2005; Bulír and Hamann 2002, 2005).

Figure 2.6 shows the substantial annual fluctuations in ODA disbursements for education and health for a sample of low-income countries between 1998 and 2004. Although there is a clear upward trend in almost all of the cases, volatility around the trend line is substantial. The issue is that volatility is particularly crippling in the social sectors because of the high share of spending on personnel—80 percent in education and 60 percent in health. Unstable resources impede new hiring, necessary salary adjustments to retain skilled providers, and regular flows of salary payments to frontline workers, undermining service delivery.

FIGURE 2.6ODA disbursements for education and health

Source: World Bank estimates from OECD DAC database.

A recent econometric analysis of child mortality outcomes from 1995–2000 in 75 developing countries found that both low levels and high volatility of donor funding for health explained the relatively slow progress of some countries in reducing under-five mortality (Bokhari, Gottret, and Gai 2005). More research is needed, but the study suggests an important possible reason that previous studies have typically found little correlation between government health spending and key health outcomes—the constraint on efficient expenditure imposed by aid volatility. More work is needed to develop systems of aid allocation that respond to performance on service delivery yet ensure stable, long-term flows of support (Eifert and Gelb 2005).

Finally, the transactions costs of accessing aid are high. Donors have made some inroads on this issue with the 2005 Paris Declaration on Aid Harmonization and the recent establishment of explicit targets for progress. In education, the EFA FTI is explicitly supporting achievement of the Paris targets among education donors and beginning to track progress.

A recent report for the EFA FTI Secretariat examined donor practices in three African countries. Although these countries are not necessarily representative of FTI countries, the data paint an encouraging picture. Education donors in these countries are well ahead of donors in general on the Paris targets. In all three countries 100 percent of donor support is “on plan,” aligned to the FTI-endorsed national education sector strategy, 100 percent of commitments are multi-year, and 100 percent of donors participate in an annual government-led review of sector performance. In Burkina Faso education donors have reached several other Paris targets as well: 100 percent of donors use shared progress indicators, 100 percent conduct joint evaluation studies, and 60 percent have aligned their financing with the country budget cycle (European Commission 2005).

In health, however, the consensus is increasing that the situation has worsened in recent years. The most difficult issue is the large number of global health programs, which have introduced “vertical” streams of assistance for specific health goals into countries’ “horizontal” health systems. These programs get high marks for raising global awareness of major health issues; attracting new funding for health; and getting countries, donor agencies, private donors, and the for-profit sector in health (notably the big pharmaceutical companies) to collaborate on new solutions to global health challenges. The spread of HIV/AIDS treatments, increased immunizations, research on new vaccines, and increased funding for diseases that disproportionately afflict low-income countries (malaria, riverblindness, Chagas disease, and dengue fever) can all be attributed to the advocacy and funding efforts of specific global partnerships.

But these programs’ success factors have a downside. Countries and their donor partners are increasingly concerned that the increase in global health partnerships is also increasing transactions costs, duplicating and fragmenting health delivery services, distorting some countries’ health priorities, and undermining holistic fiscal and sectorwide planning, when significant NGO activities supported by global funders are unknown to public officials.

The planning cycles, coordinating mechanisms, appraisal processes, financing channels, surveillance metrics, procurement requirements, and audit and reporting requirements of global health programs can differ from program to program and from governments’ health systems. Although weak expenditure management in some countries may justify ring-fenced donor funding channels, it does not justify a plethora of channels and procedures. Nor does it justify multiple coordination mechanisms: Angola and the Democratic Republic of Congo have each been required to establish four HIV/AIDS coordinating bodies. Nor does it justify the high cost of multiple reports: hosting missions and writing reports for different health programs is estimated to absorb 50–70 percent of the time of a district medical officer in Tanzania (McKinsey and Company 2005). Multiple parallel procurement requirements have not only increased countries’ administrative costs, but also reduced their bulk purchasing power. Initiatives such as the “three ones” and the Global Task Team on HIV/AIDS aim at improving coordination, but there is clearly room for progress.

Concerns are also being raised about the long-term fiscal sustainability of donor-induced shifts in health spending, particularly in connection with HIV/AIDS programs. Although developing countries have increased domestic funding for these programs since 2000, this increased funding is dwarfed by the 300 percent rise in external support, largely from multilateral sources such as the Global Fund. About 60 percent of total AIDS program spending in developing countries is now externally financed, and this share is much higher in many African countries. In Ethiopia in 2003-4, external funding for HIV/AIDS equaled the overall public health budget. And in both Uganda and Zambia, AIDS funds exceeded public health spending by almost 185 percent (Lewis 2005b).

At the December 2005 High-Level Forum on the Health Millennium Development Goals, health ministers from a range of developing countries agreed with bilateral and multilateral donor representatives that action is urgently needed to address transactions costs and sustainability issues in health. Participants noted that countries with strong unified health sector plans—such as Bangladesh, China, Ghana, the Kyrgyz Republic, Tanzania, and Vietnam—had been more successful than others in using support from global health partnerships without distorting their own health sector priorities. Countries with strong leadership on HIV/AIDS, such as Rwanda, also report success in “pushing back” on global health partnerships and bilateral funders to insist on harmonized indicators, joint monitoring, and pooled funding.

A proposal currently being developed would commit donors and countries to a shared set of principles and establish a small global secretariat to monitor and support adherence to these principles at the country level. The “good practice” principles being discussed are joint commitment by countries and their donor partners to

  • one unified, country-led health sector plan aligned with the PRSP, covering all actors—the national health system, the private sector, and NGOs;
  • reliance on country systems, budget frameworks (such as the medium-term expenditure frameworks), reporting arrangements, and aid coordination mechanisms;
  • support for strengthening health system capacity;
  • a results focus; and
  • a medium-term to long-term (10-year) horizon for aid commitments and greater predictability.

Since the forum, at least three global health partnerships have discussed the best practice principles at their board meetings. Given the limited consensus on the same issues among donors for health just one year earlier, the current climate is promising.

Reducing Leakage

Donor support for increased and more flexible aid hangs crucially on countries’ ability to demonstrate effective use of that support. Yet leakage—or the failure of resources to be used effectively for intended purposes—is an acknowledged issue in developing countries where governance systems and accountability pressures are often weak. Even where the private sector plays a key role in health, education, or water and sanitation service delivery, achieving the MDGs will depend fundamentally on developing countries’ ability to strengthen public sector performance.

Not all leakage is corruption. Poor management can lead to the wrong choice of technologies, suboptimal allocations of resources, and other inefficiencies that benefit no one in particular: badly designed subsidy programs or unbalanced education budgets that leave teachers with no books, new construction inaccessible to people with disabilities, and inappropriate drug purchases.

But health and education systems have features that create special opportunities for private gain at public expense. Asymmetric information between providers (doctors, teachers) and their patients or students give the providers heightened power to extract private payments for public services that should be free. The inherent decentralization of school and health systems makes it hard to monitor performance on the front line and as a result gives rise to absenteeism, theft, and shirking of duties. Large-scale procurement and payment systems, especially in health, create scope for graft and theft, as seen in developed countries as well. Finally, rapid increases in external funding, particularly for HIV/AIDS programs, can pressure governments to disburse large amounts of funding quickly, sometimes through wholly new institutional channels. Recent studies provide growing evidence of three different types of leakage.

Central losses. Very little systematic information exists on “losses at the top” of developing-country education and health systems. Although the scale of such losses cannot be estimated, they are clearly the point at which the most concentrated leakage of government and donor funds occurs.

In health, the hospital sector typically represents 30–50 percent of public health spending and its large-scale, often centralized procurement presents clear opportunities. In Colombia procurement overpayments were estimated at $2 million a year, enough for health insurance coverage for 24,000 people (Di Tella and Savedoff 2000). In Ghana survey respondents estimated that 21 percent of hospital procurement was corrupt and that 18 percent of a contract’s value was typically required in kickbacks to public officials (World Bank 2000). In Venezuela and Costa Rica, over two-thirds of medical staff reported knowledge of stolen materials, equipment, or drugs, and in Uganda researchers estimated the average leakage rate for drugs across public facilities at 73 percent (McPake and others 1999). Parallel issues exist in education—procurement fraud in construction and theft and misprocurement of textbooks—but the market value of school books is lower than for drugs, and construction is often less equipment intensive.

Leakage of funds allocated to front-line facilities. More systematic information is emerging from public expenditure tracking studies (PETS) on the extent to which funds budgeted for schools, clinics, or specific programs fail to reach intended levels. Surveys in more than a dozen countries have consistently found discrepancies, although the scale varies considerably. In Ghana, for example, only 20 percent of budgeted nonwage transfers to health clinics were actually received, and only 51 percent of transfers to primary schools. Clinics in Tanzania received only 59 percent of nonwage transfers. Very commonly, funds are received only with substantial delays.

Front-line service failures. Their aggregate size may or may not be as large as losses at the center, but the most widespread losses and abuses in health and education systems occur on the front lines—providers absent from duty and providers demanding informal payments for services that are legally free. Crosscountry studies show, on average, that one health worker in three is missing during unannounced facility visits, although there is a range. Absentee rates in education average about half that. Provider absence seriously disrupts service delivery, lowers system productivity, and depresses the demand for services. Researchers in India found that facility closures followed no pattern, meaning that a patient’s likelihood of finding a provider was unpredictable (Bannerjee, Deaton, and Duflo 2004). The deterrent to seeking services is thus high.

Many reasons for absence are legitimate, including rural workers’ need to travel to distant towns for paychecks or supplies. Low salaries and payment arrears, which force many competent and committed providers to work second and third jobs to survive, are clearly root causes of much absenteeism. But weak accountability pressures also contribute; absentee workers rarely face sanctions.

Equally pervasive are informal payments. In health, patients can be impelled to pay to be seen by a physician, to be admitted to hospital, to move up in the queue, to get an exemption from official fees, or to ensure better-quality treatment, as well as for basic supplies and services, such as blood supplies, drugs, food, or bed sheets. Informal payments are also seen in education—for tutoring, for graduation, for passing grades, and even for university access. The wedge between public sector pay scales and what providers can command in the market contribute to the demand for informal payments. So does the absolute insufficiency and irregularity of recurrent budget transfers to the facility level in many countries. Without informal payments, service delivery would be impossible in many places.

Informal payments disproportionately hurt the poorest. In about half of the 29 countries with data, the average informal payment for publicly provided health care was more than one-quarter of monthly per capita income—clearly imposing a hardship on low-income families (Lewis 2005b). In such environments, major illnesses pose a deep threat to families, forcing them to sell assets or incur debt to obtain needed medical care (Lewis 2000; Falkingham 2002, 2004; Killingsworth and others 1999). Evidence from Kazakhstan showed poor households spending more than twice their monthly income for health care in acute cases (Lewis 2005b).

Although increased aid could ease the conditions that spawn poor service delivery and corruption on the front lines—the low and irregular pay, the ill-maintained facilities, the shortages of books, supplies, or medicines—it can do so only if system managers are accountable for resource use and if incentives exist for performance.

Corruption at the top levels of health and education systems is harder to root out, and additional resources can simply increase the opportunities. The recent mismanagement of funding for HIV/AIDS programs in Kenya, Uganda, and the Ukraine provides glaring examples (Transparency International 2006).

Weak management systems can be improved, however, and corruption can be curbed; later chapters of this report examine the cross-country experience with strategies to strengthen governance. Health and education systems typically need to strengthen accountability relationships along two axes of the “accountability triangle” presented in World Development Report 2004: (1) strengthening system management—that is, policy makers’ ability to contract with providers—and (2) strengthening “client power” or users’ ability to demand better service from providers (World Bank 2004b). In both areas, many developing countries are making progress in putting transparency and new accountability relationships to work.

The following are some of the most promising strategies for improving system management:

  • Use transparent allocation rules, procedures, and the power of information to reduce leakage. Clear rules and procedures concerning the basic package of services to which people are entitled, fee scales, resources a school or clinic should receive, and effective basic accounting and record keeping can have immediate impacts. Tbilisi Children’s Hospital began posting lists of fees, and informal payments fell. Cash registers were installed in Kenyan hospitals to collect user fees and revenues rose 400 percent in three years, with no change in utilization rates (Vian 2006). Transfers to schools in Uganda went up when the government openly publicized allocations in the press and on each school’s door.
  • Measure results, performance, and impact. Clear measures of output and performance are fundamental for systems improvement. Parents are ill-equipped to evaluate the quality of local schools without data on how well their students are learning compared to students at similar schools. Basic data on education and health system performance is fragmented and two or more years out of date in most developing countries. Both to better manage the resources they have as well as to make the case for more support, countries need robust, realtime information on systemwide outcomes and intermediate results (primary completion rates, immunizations and other services delivered, morbidity and mortality data).
  • Implement a credible and effective audit function. In Brazil and Chile, the capacity of the federal government to conduct random audits of enrollment records and budgets was crucial for the implementation of large-scale financing reforms that pay schools on the basis of attendance, creating powerful new incentives for schools and mayors to get hard-to-reach children into school.
  • Focus on provider quality, deployment, and incentives. Skilled providers are the most expensive resources in health and education systems; recruiting, deploying, equipping, and supervising them carefully are key for the productivity of spending. Average salaries may need adjustment in many contexts, but research also shows that nonsalary inducements (such as housing, training, research opportunities, and public recognition) are as important as salary incentives for providers’ motivation and development. Systems also need the authority to reward performance, and discipline, transfer, or terminate employees who engage in abuses. Technology may help; a program in India that provided digital cameras (with a tamper-proof time and date function) to remote rural primary schools and rewarded teachers who supplied one picture per day of the class in session found teacher absence went down to 22 percent compared with 42 percent in control schools (Duflo and Hanna 2005).

Accountability will be strongest if education and health systems need to answer to stakeholders on the performance of public services. Reforms that are increasing “client power” in different countries include:

  • Involve communities in monitoring and management. An increasing number of countries in all regions are devolving some control over schools and health clinics to local communities. In Rwanda community councils are being given the power to hire and fire health clinic personnel and determine bonus pay. In Brazil, El Salvador, Guatemala, Honduras, India, Mexico, and Nicaragua, among other countries, village education committees and parent-teacher associations have a voice in hiring teachers and managing school accounts. Creating formal oversight bodies at the community level can place strong accountability pressures on local providers.But unleashing this client power often requires significant efforts to inform communities of their responsibilities and to build their skills to shoulder them. In India, two years after village education committees were mandated to receive direct school funding, only 12 percent of rural households surveyed in northern India knew about it, and only 26 percent of the committees had met in the previous six months (Pandey 2005). In Brazil, in the poor rural municipalities where auditors found irregularities in the use of federally transferred funds, the community councils created to supervise the transfers admitted they lacked the skills and power to challenge local mayors (Transparency International 2004).

Efforts to equip communities with better information, such as the “report cards” on schools’ performance being developed in Cambodia and the training journalists in Benin receive to report on local school committee meetings, are important steps to shift what is often an imbalance in status, knowledge, and power between providers and beneficiaries, especially the poor. Letting clients provide feedback on the quality of public services—for example, through “citizen report cards”—is also empowering, as a signal from the public sector that citizens’ views are valued.

Donors are legitimately concerned about the risks of providing flexible aid where public sector performance is weak. But, as noted above, the core financing need in most low-income countries for increasing health and education services remains recurrent personnel costs. Stronger capacity across mainstream ministry functions—planning, budgeting, and expenditure management—is clearly needed. Better quality and more timely administrative data on system performance are also needed, because they provide the platform to track results for money spent. And donors’ support must be guided by credible sector strategies that address key constraints, clarify accountability, and improve incentives for performance.

Doing More to Reach the Poor

The MDGs aim to extend human welfare improvements to all countries. But some of the goals—especially in health—can be achieved through investments that primarily benefit the better-off, while largely bypassing the poor (Gwatkin and others 2000). Last year’s report showed that two-thirds of the countries that had reduced child mortality from 1990 to 2001 saw a widening gap in outcomes for families in the lowest income quintile. And although education progress has been more pro-poor, in one-third of the countries that improved primary enrollments from 1990 to 2002, the poorest have lagged behind.

Since 2002 a new wave of demographic and health survey data has become available, shedding light on countries’ more recent strategies and progress—and capturing experience since the MDGs were adopted. Analysis of the 10 developing countries with survey data for 2003 and 2004 shows the following:

  • Nine of 10 are making rapid progress on child mortality. The exception was Kenya, where child mortality increased between 1999 and 2003.
  • Seven of 10 countries are making strong progress in child immunizations, and in almost all of these the poor are benefiting most.
  • Better quality antenatal care is also reaching the poor.
  • Several countries—including Bangladesh, Kenya, Morocco, and Mozambique (where outcomes for the poor have improved, while the average has not)—are reducing child malnutrition faster for the lowest income quintile than for the population average.
  • Eight of the 10 countries have improved primary completion rates, and in five of these, the poorest quintile has improved most.

Child mortality declined in 9 of the 10 countries between 1998-9 and 2003-4, and in Indonesia, Madagascar, and the Philippines the current pace of decline (more than 4.3 percent a year) is sufficient to reach the MDG (figure 2.7). In Madagascar the annual rate of improvement since the MDG baseline year, 1990 (-2.3 percent per year), puts it off track to meet the MDG in global estimates, but the new data show that progress has accelerated sharply over the last several years. Among Indonesia, Madagascar, and the Philippines, however, only Indonesia has seen the improvement for the poorest quintile keep pace with that for the population average. In Bolivia, Cameroon, and Mozambique, on the other hand, outcomes for the poor are improving faster than the mean, but the overall rate of improvement is not fast enough to reach the MDG. In Bangladesh, Burkina Faso, and Morocco, the poorest quintiles are lagging, but the gaps are not wide.

FIGURE 2.7Annual reductions in child mortality (number of child deaths per 1,000 live births)

Source: World Bank estimates from Demographic and Health Surveys.

Note: The boxed numbers show the number of child deaths per 1,000 live births in the most recent survey.

Immunization coverage is also improving in many of these countries, with impressive progress in reaching families in the poorest quintile. Except in Bangladesh (which already had a very high level of immunizations) and Madagascar (where there is a slight gap), the poorest groups have experienced much greater improvement in access to immunizations than the population as a whole. In three of the countries, however, no improvement was registered. While in the Philippines the average level of immunization coverage is quite high, in Kenya and Indonesia, it is not. It should be recalled, however, that this period was a time of economic crisis in Indonesia, which makes the country’s continued progress in addressing child mortality all the more impressive.

What explains these different patterns? Progress on child mortality reflects complex determinants (such as mothers’ education, household income, and household access to water and sanitation), many of which can be slow to change. Is there a trade-off between faster aggregate progress toward key goals, such as child mortality, and progress for the poor? Given the higher marginal costs of extending basic services to rural areas, where a high share of the poor live, many observers would respond in the affirmative. Although these survey data permit no clear answers, they do provide some encouraging evidence that key health interventions, such as immunizations and access to trained providers for antenatal care, are spreading quickly and increasingly reaching the poorest groups, at least in this sample of countries.

In education the picture is also one of progress; primary completion increased substantially in three countries and more modestly in five others. Two countries, however, (Kenya and Bolivia) experienced declines, especially among the poorest quintile (figure 2.9). When considering these data, it is important to remember that they reflect changes in the education system and participation rates from roughly a decade ago, as they are based on reported schooling attainment for the 15–19-year-old cohort of the population. The five countries with modest annual improvements are, with the exception of Cameroon, countries that have already achieved close to universal primary enrollments and relatively high primary completion, so it is logical that marginal changes should be lower. In half of the countries, the poorest quintile improved more than the average. Mozambique and Burkina Faso were exceptionally progressive in reaching the poorest children over this period.

FIGURE 2.8Delivery of immunizations

Source: World Bank estimates from Demographic and Health Surveys.

Note: The boxed numbers show the immunization coverage (for children 12-24 months) in the most recent survey.

FIGURE 2.9Share of 15–19-year-olds who have completed primary school

Source: World Bank estimates from Demographic and Health Surveys.

Finally, the recent surveys provide clear evidence that awareness of HIV/AIDS has become almost universal in Sub-Saharan Africa, but far less so in other regions. In Bolivia, Indonesia, and Bangladesh, less than 33 percent of women in the lowest-income quintile have “heard of” HIV/AIDS. In Bangladesh there was improvement from 1999 (only 8 percent of low-income women had heard of AIDS) to 2004 (29 percent), but there is clearly a long way to go. And in both Indonesia and Bolivia, awareness did not increase over the period.

Investing Smartly across Sectors

Water supply and sanitation investments have important effects on health, especially child health. Nearby water and school latrines can dramatically change mothers’ ability to care for their children and girls’ school attendance. Roads improve school attendance and use of health facilities. And health and education investments have clear complementary effects: mothers’ education is a strong correlate of child survival, and school attainment is affected by family illness, especially HIV/AIDS. There is no single route to MDG progress.

Research on what infrastructure investment means for human development outcomes shows that effects can be large. In Zimbabwe a carefully controlled study of preschool children found that those living in households that used wood-burning stoves for cooking were more than twice as likely to suffer from acute respiratory infections (ARIs) as children in homes with natural gas or electrification (Mishra 2003). ARIs are one of the leading causes of childhood illness and death in Africa, and it is not clear that any direct health intervention could produce a 50 percent drop in their incidence.

In Coahuila, Mexico, the “piso firme” program has upgraded dirt floors to cement floors in slum housing, benefiting more than 34,000 people since 2000. An impact evaluation found clear health benefits for children under age five: 20 percent less diarrhea and anemia and 12 percent less infection with parasites. The improved health also translated into better brain development (children in the treatment homes scored 8 percent higher on cognitive tests), and older children missed fewer days of school.

Donors and governments concerned about making money work need to give priority to establishing careful baseline studies and robust control groups wherever possible before launching innovative programs. Donors also need to recognize that the knowledge that comes from rigorous impact evaluation is a global public good and needs to be heavily supported, especially in low-income countries. Developing-country policy makers can provide a demand-side stimulus to good impact evaluation by asking, “Where’s the evidence?” when considering new policies and programs.

Priorities for Global Action

Evidence of tangible progress toward the MDGs is greater today than one year ago. Many low-income countries have accelerated progress on primary completion. Childhood immunizations have increased greatly and are reducing needless child deaths. National malaria campaigns are getting visibility and distributing bed nets and treatment on a larger scale than ever before. HIV prevention programs are beginning to work, and the extension of effective drug therapies to AIDS victims in the developing world has been rapid.

Development assistance for the MDGs has increased sharply, and donor support appears to be more targeted than ever to the lowest-income countries with country-owned poverty reduction strategies and commitment to these goals. Countries have made progress in expanding service coverage while managing unit costs. But the world is still far from achieving the human development MDGs; donors and countries must keep working on ways to speed the pace of progress. The analysis in this chapter points to six priority areas for global action:

  • Accelerate harmonization in health. Global health partners need mechanisms for aligning policies and programs and for harmonizing procurement, disbursement, and reporting at the country level, as well as a mechanism for coordination and intermediation at the global level. The education sector through the EFA FTI has made notable progress in developing both global and country coordination mechanisms that are improving donor alignment and lowering transaction costs. A parallel mechanism is urgently needed in health.
  • Increase the flexibility and predictability of ODA for social sectors. Where sector policies are sound, fiduciary conditions are adequate, and the capacity to measure sector outcomes and intermediate indicators is in place, donors should shift assistance as much as possible to budget support. This shift will permit countries to scale up health and education coverage most efficiently and lower the costs of attaining the MDGs. Donors and countries should plan multiyear expenditure programs on a disbursements rather than a commitments basis, and the schedule for donor disbursements should be clear over at least a three-year horizon, extended annually.
  • Improve measurement of results, performance, and impact. Developing countries seeking flexible aid need to demonstrate adequate public expenditure management. But they also need the ability to track education, health, water and sanitation sector performance on a timely and reliable basis, so that donors can have real-time data comparing spending and results. These data include data on key outcomes, as well as intermediate indicators, and accurate financial reporting at all levels of the system. Countries should insist on rigorous evaluation of pilot programs to guide decisions on where to increase spending. Because such evaluations can be expensive and have a high element of global public good, donors need to increase their support for them.
  • Monitor outcomes of the poorest groups. Extending basic health, education, water and sanitation to the poorest segments of the population can be difficult and costly. Donors should support country policy choices that make investments more pro-poor, even at the risk of slowing overall progress on reaching the MDGs. These choices will hinge on countries’ ability to track outcomes by income group, gender, ethnicity, and region. Regular household surveys are essential.
  • Strengthen the accountability of health, education, and water and sanitation systems. Achieving the MDGs depends above all on more effective delivery of basic health, education, water and sanitation services in many countries. Countries will make the most progress if they find ways to strengthen sector management, the incentives for providers, and the voice of clients at the point of service delivery.
  • Develop a systematic cross-country database of public expenditures on social sectors. The OECD DAC has made good progress over the past two years in upgrading the quality and timeliness of crosscountry data on ODA, in part in recognition that these data are crucial for monitoring progress in implementing the Monterrey accords and donor commitments related to the MDGs. Sorely needed is a parallel effort by the Bretton Woods institutions to ensure shared standards and classifications in the collection of government expenditure data and to unify these in a single, standardized cross-country database.

Notes

1.Note that the data for 1990, 1995, and 2000 are not comparable to those presented in Global Monitoring Report 2005 because of a change in the series.
2.The WHO Expanded Program of Immunization (EPI), which Bangladesh followed, vaccinated children against six diseases: diphtheria, measles, pertussis, polio, tetanus, and tuberculosis.

    Other Resources Citing This Publication