4
The well-being of poor people is the point of Report and this chapter try to give the term
making services work. The value of public “empowerment” a precise and concrete
policy and expenditure is largely determined interpretation. Specifically, the chapter dis-
chapter by the value the poor attach to it. When cusses the potential for poor people to
publicly provided and funded housing is left influence services by:
vacant,174 when food supplies are not eaten,
when free but empty public health clinics
• Increasing their individual purchasing
power.
are bypassed in favor of expensive private
care,175 this money is wasted. • Increasing their collective power over
Improving services means making the providers by organizing in groups.
interests of poor people matter more to • Increasing their “capacity to aspire”177:
providers. Engaging poor clients in an allowing them to take advantage of the
active role—as purchasers, as monitors, and first two by increasing the information
as co-producers (the “short route”)—can needed to develop their personal sense of
improve performance tremendously. capability and entitlement.178
How can public policy help poor people
acquire better services through this route?
By expanding the influence of their own
When will strengthening
choices. By having the income of providers the client-producer link
depend more on the demands of poor matter most?
clients. By increasing the purchasing power In the framework of chapter 3, improving
of poor people. And by providing better client power—the short route of service deliv-
information and a more competitive envi- ery—can overcome various weaknesses of the
ronment to improve the functioning of ser- long route (figure 4.1), even when services
vices. Where such choice is not feasible, remain the responsibility of government. The
governments can expand consumer power clearest case is monitoring providers. Clients
by establishing procedures to make sure are usually in a better position to see what is
complaints are acted on. going on than most supervisors in govern-
Sad to say, governments and donors fre- ment hierarchies—who provide the compact
quently neglect the possible role of poor and management. When the policymaker-
clients in sustaining better services—or provider link is weak because of scarce or diffi-
treat that role merely as an instrument for cult-to-manage supervisory staff, clients may
achieving a technically determined out- be the only ones who regularly interact with
come. Neither governments nor donors are providers. As discussed several times in this
accustomed to asking the poor for advice. Report, improvements in basic education have
Recent initiatives have begun to redress this often depended on participation by parents.
through a variety of ways to increase partic- Although parents cannot monitor all aspects
ipation by communities and civil society. of education, they can monitor attendance by
But the potential for improvement has not teachers and even illiterate parents can tell if
yet been adequately tapped.176 their children are learning to read and write.
In short, the key is to enhance the power Citizens as clients can also make up for
of poor clients in service provision. This shortcomings in the voice or politics relation-
64
09_WDR_Ch04.qxd 8/14/03 8:09 AM Page 65
ship. If governments cannot or will not try to Figure 4.1 Client power in the service delivery framework
determine and act on the desires of the pub-
The state
lic, or if the desires of poor people are system-
atically ignored, there may be few options for
poor communities but to develop mecha- of accounta
ro u te bili
nisms for getting services some other way. ng ty
Lo
The greater the differences among
clients—their heterogeneity—the more that S h o rt r o u t e
direct client power is likely to have an advan- Citizens/clients Providers
tage relative to the “long route.” The greater Nonpoor Poor Frontline Organizations
Client power
the individual differences in preferences for
the type and quality of services provided, the
greater the importance of discretion on the Services
part of providers and the more difficult it is
to monitor the use of this discretion centrally. extend these services in the first place. There is
Sometimes preferences differ geographically, no reason to believe they are all self-corrective
so different levels of government may reflect through replicating aspects of the free market.
this variation. But for many services, the het- Similarly, some settlements constitute
erogeneity of preferences applies all the way communities with sufficiently congruent
down to the individual. Take courtesy and interests among members, egalitarian norms
comfort (caring) relative to technical skill to protect the poor, mutual trust, and the
(curing) in health delivery—or farmers with ability to mobilize information and to act
constraints on their time and other workers collectively—that is, they have social capi-
in the same community with different con- tal.179 But some clearly do not. How many vil-
straints. Certainly people differ in the lages and urban neighborhoods are there in
amounts of water and electricity they want, the developing world? Hundreds of thou-
given their other needs. Government struc- sands? Millions? And how many kinds of
tures may not be flexible enough to accom- social structures are represented? Ensuring
modate this variety. And where local prefer- that poor people have a say in this variety of
ences vary systematically between the poor circumstances demands that policies be
and others, honoring poor people’s prefer- examined and designed with a great deal of
ences over those of the better-off can be a local knowledge and an understanding of
challenge. local conflicts and inequalities. Pretending
For some collective action problems, gov- otherwise will almost certainly do real harm.
ernments may not be located at the correct And some services, particularly for health
level to solve them, no matter how willing and modern water and sanitation, need tech-
they are to pursue the interests of the poor. nical inputs to be successful. Patients—as
The boundaries of the political jurisdiction individuals or health boards—are good
may not correspond to the boundaries of the judges of courtesy and attendance. But they
problem. So schools are often the most are much less able to judge clinical quality or
appropriate unit for management and opera- the appropriate mix of curative and preven-
tion. Sanitation services need community tive services. And some health problems have
pressure to ensure that everyone uses fixed- effects that spill over community boundaries.
point defecation, but they are often organized Large pest-control initiatives and other forms
around communities that are larger or of infectious disease control may seem a low
smaller than villages, depending on the den- priority for any one group of citizens, yet will
sity of population. A more active role for be effective only when all participate. Ulti-
communities is needed in such cases. mately, some wide-scale government inter-
It is important to avoid romanticizing vention is necessary. Still, emphasizing the
either form of increasing client power—nei- power of clients is a welcome tonic for the
ther choice nor participation is sufficient for top-down, technocratic orientation that has
all services. Market failures and concerns for characterized much development thinking
equity lead societies to want to improve or until now.
09_WDR_Ch04.qxd 8/14/03 8:09 AM Page 66
The scarcity of commodities due to low there remain problems, such as ensuring the
pricing may lead to other commonly quality of the newly established schools, but
reported problems—illicit sale of materials these are secondary to getting girls to school.
and the demand for under-the-counter pay- The revenues that providers raise from
ments. Indeed, “free” public services are often charges at the point of collection are often the
very expensive. Many countries have serious reason some services can continue at all.
diversions of pharmaceuticals from the pub- Much of the success of the Bamako Initiative
lic stock into private markets, where they in West Africa (see spotlight) stems from the
instantly become expensive. In general, ser- supply of pharmaceuticals made possible by
vices that most directly resemble market charging users for them. Bamako Initiative
goods have a greater problem of diversion villages usually have drugs, other villages usu-
and implicit privatization. ally don’t. Sustainability in piped water sys-
In Eastern Europe the health systems are tems is almost always equivalent to financial
often ranked among the most corrupt of
public services (box 4.2). Under-the-table
payments and pharmaceutical sales to open BOX 4.1 The private sector is preferred in Andhra
markets are the main elements in this assess- Pradesh, India
ment. If directives against such practices can-
not be enforced, countervailing pressure is A study of consumer and producer attitudes was conducted in six districts in the southern
needed (see box 3.6). Formalizing fees and Indian state of Andhra Pradesh.The study included 72 in-depth interviews and 24 focus
groups.
putting purchasing power in the hands of
poorer clients is one possible source of such Private Public
pressure.184 ATTITUDES OF DOCTORS
Exemptions from fees can have perverse
“They speak well, inquire about our health.” “Does not talk to me, does not bother (about
effects by reducing this purchasing power. In my feelings or the details of my problems).”
Benin a measure to raise female school “Ask about everything from A to Z.”
“Don’t tell us what the problem is, first check,
enrollment—waiving fees for girls—led “Look after everyone equally.”
give us medicines and ask us to go.”
teachers to favor the enrollment of boys and “They take money . . . so give powerful
“They are supposed to give us Rs. 1000 and 15
medicine . . . treat better.”
to raise informal fees for girls.185 Of course, kg of rice for family planning operations; they
give us Rs. 500 and 10 kg rice and make us run
the problem could have been solved by abol- around for the rest.”
ishing fees for everyone (if the teachers could “Anyhow they will get their money so they
continue to be paid) or by closer monitoring don’t pay much attention.”
and enforcement by education officials. But CONVENIENCE
in a system that has problems paying teachers
“Treat us quickly. . . .” “Do not attend to us immediately.”
and weak administrative capacity, bolstering
the ability of girls to pay with vouchers seems “We spend money but get cured faster.” “Have to stand in line for everything.”
more likely to succeed. “I know Mr. Reddy. He is a government doctor “Doctor is there from 9 a.m. to 4 p.m.—when
but I go to him in the evening.” we need to go to work.”
“Can delay payment by 5–10 days. He is OK “I have not been there, but seeing the
Increase and sustain supply with that, he stays in the village itself.” surroundings . . . I don’t feel like going.”
Greater purchasing power may simply COST
increase supply and overcome bottlenecks “Recent expenses came to Rs. 500 for 3 days . . . “While coming out, compounders ask us for
due to supply problems. In Bangladesh the had to shell out money immediately.” 10–20 Rs.”
Female Secondary School program awards “We have to be prepared to pay, you never “Anyhow, we have to buy medicines from
scholarships to girls if they attend school reg- know how much it is going to cost you.” outside.”
ularly and gives secondary schools a grant ADVANTAGES
based on the number of girls they enroll. Sec- “Even if I have to take a loan I will go to “Malaria treatment—they come, examine
ondary school enrollment in Bangladesh is private place, they treat well.” blood, give tablets.”
increasing, and faster for girls than for boys. It “For family planning operations.”
also led to the establishment, at private “Polio drops.”
expense, of new schools. Desires for single- “In case I do not get cured in private hospital,
sex schools and separate toilet facilities for but it is very rare.”
girls were mysteriously accommodated when Source: Probe Qualitative Research Team (2002).
girls’ attendance meant more money. True,
09_WDR_Ch04.qxd 8/14/03 8:10 AM Page 68
resent under-the-table payments, but there bers than from individuals going it alone.
may be no incentive to complain if, say, a So client power expressed outside market
doctor is using public facilities and materi- transactions will almost always be expressed
als at the same time. Clients know that the through collective action.
service is still cheaper than if they had to go Strengthening participation along the
to the market, and so do not complain.193 client-provider link can fix problems in the
There is a deeper constraint: even when long route of government provision. So com-
there is an opportunity to redress com- munity groups that take on complaining,
plaints, monitoring and follow-through are monitoring, and other means of making sure
public goods—the benefits accrue to the things work properly would be expected at
entire group while the costs are borne by a some point to become institutionalized
few. This is true for communities as well as within government (most likely local govern-
individuals, but groups of people generally ment), or possibly to be supplanted by gov-
find it easier to elicit support from mem- ernment as it improves. After all, collective
11). One hypothesis is that this slows the delivered to all people is one of the many
development of local government capacity. open questions on the agenda.
But the opposite argument has also been
made—that such groups are a catalyst for Client power in eight sizes
developing local government capacity. In To sum up, increasing client power through
northeast Brazil, social investment funds improved choice or direct participation will
led villagers to organize and petition higher be important when people differ—are het-
levels of government to, for example, guar- erogeneous in their preferences—or when
antee a teacher to staff a school built by the either of the two legs of the long route to
community.207 accountability is problematic. In terms of
the decision tree (figure 4.2) that deter-
Sustainability. Participatory water pro- mines which of the eight types of solutions
jects, underway since the 1930s and 1940s, is appropriate, client power matters at all
have often improved water supply—at least three decision points.
for awhile. But at some point water pumps
and other pieces of expensive equipment Decision 1: Are politics pro-poor? Reliance
break down. Covering the capital cost on client power should vary with the capac-
(which is expensive) and obtaining the ity and orientation of government. Also
technical help (also expensive) have always with the question of which level of govern-
been the bottleneck for water projects in ment is problematic. When governments
poor areas. When a new infusion of capital (central, local, or both) are pro-poor, they
is necessary at short notice, the community may choose to enlist client groups as moni-
must look either to donors or to regular tors or solicit their opinions regularly in
sources of funds, such as taxes or other gen- sizes 3 and 4. Sizes 5 through 8, however,
eral revenue. Eight or nine years after the require ways to avoid the problems of gov-
original investment, are the donors still ernment. All four will involve getting infor-
around? Do they have the same priorities mation to clients on their entitlements to
they originally had? Can they respond and the performance of services.
quickly to small individual requests? Often When levels of government differ in
not. These demands will have to be met by their commitment to poor people, the role
local government,208 and projects have been and sponsorship of user groups differ as
evolving to work through them. well. If central government is a better cham-
Such projects may have been a great deal pion of poor people, they may fund com-
better than relying on inadequate govern- munities (if preferences vary between
ment structures. The argument for them is them) or cash transfers or vouchers (if pref-
strongest where the current government erences vary within them) in cases 7 and 8.
system, especially the local government, is If local government is better, they can pro-
weak, with few prospects for changing any vide or contract for these services. When no
time soon. level of government is pro-poor, then
This should, however, be a tactic that donors, if they are inclined to be involved at
supports a longer-term strategy of develop- all, might choose to fund community
ing governmental capacity—strategic incre- groups or organizations within civil society,
mentalism, discussed in chapter 3. Caution being careful not to undermine the devel-
is required when there appears to be a opment of government capacities.
tradeoff between improving services in the
short run and undermining delivery capac- Decision 2: Does heterogeneity matter?
ity in the longer run. And the political con- Sizes 3, 4, 7, and 8 directly involve clients.
sequences of participatory projects should When preferences differ by location then
be the subject of careful evaluations. All this decentralization to local government or to
complicates bringing these interventions to community groups (depending on the
scale. It may be possible to replicate com- capacity and pro-poor orientation of the
munity efforts in many places, but whether former) makes sense. If they differ by indi-
this is the best way to make sure services are vidual, then purchasing power and compe-
09_WDR_Ch04.qxd 8/14/03 8:14 AM Page 75
tition for individual business are preferable. work service users improve services either
Providing information to clients is critical by choice in purchasing or by active partic-
for translating their choices into better ser- ipation.
vices. It is only in size 1, where government is
perfectly capable of providing services
Decision 3: Is monitoring easy or hard? directly, that client participation is optional.
When monitoring is easier for clients than Possibly size 3 as well if government can
for governments (at any level) then client accommodate varying needs of clients. For
input may be required for sizes 2, 4, 6, and all other cases, the client needs to be placed
8. Parents of children, patients, and net- more firmly at the center of service delivery.
10_pgs 76-77_Ch04Spotlight.qxd 8/14/03 3:16 PM Page 76
Scaling up incrementally
strengthening the power of communities analysis of the main constraints in the three The Bamako approach was implemented
over service providers. Policymakers bal- countries led to emphasis on service delivery gradually, with the support of UNICEF,
anced this power with sustained central strategies focusing on the poor.211 Priorities WHO, and the World Bank, building on a
involvement in subsidizing and regulat- included:212 variety of pilot projects.213 Since the early
ing services—and in guiding community 1980s, it was progressively scaled up in the
management. • Implementing community-owned revolv- three countries—from 44 health facilities in
ing funds for drugs with local retention Benin to 400 in 2002, from 18 in Guinea to
The initiative improved the access, avail-
and management of all financial proceeds. 367, and from 1 in Mali to 559. This raised
ability, affordability, and use of health
services. Over the more than 10 years of • Revitalizing existing health centers, the population with access to services
implementation in these three countries, expanding the network, and providing within 5 kilometers to 86 percent in Benin,
community-owned services restored access monthly outreach services to villages 60 percent in Guinea, and 40 percent in
to primary and secondary health services within 15 kilometers of facilities. Mali, covering more than 20 million peo-
for more than 20 million people. They • Stepping up social mobilization and ple. Importantly, a legal framework was
raised and sustained immunization cover- community-based communication. developed to support the contractual rela-
age. They increased the use of services tionship with communities, the cost-shar-
among children and women in the poorest
• Pricing the most effective interventions
ing arrangements, the availability of essen-
below private sector prices, through sub-
fifth of the populace. And they led to a tial drugs, and community participation
sidies from the government and donors
sharper decline in mortality in rural areas policies. Community associations and
and through internal cross-subsidies
than in urban areas. management committees were registered as
within the system. Local criteria were
Despite the various targeting mecha- legal entities with ability to receive public
established for exemptions (table 1).
nisms, affordability remains a problem for funds.
many of the poorest families. But even with • Having communities participate in a
biannual analysis of progress and prob-
limited inclusion of the poorest people,
lems in coverage with health services— Better health outcomes
improvements were significant.210
and in the planning and budgeting of for poor people
services. Over the 12 or so years of implementation
Revitalizing health networks • Tracing and tracking defaulters—and in Benin and Guinea, and more than 7 years
In these three countries, serious disrup- using community representatives to in Mali, health outcomes and health service
tions to the situation of health services had increase demand. use improved significantly. Under-five mor-
occurred during the 1980s as a result of a
severe economic recession and financial
Table 1 Reaching out to benefit the poorest groups
indebtedness. The health budget in Benin
went from $3.31 per capita in 1983 to Disease Geographical Cross-subsidies Exempting the poor
targeting targeting
$2.69 in 1986. In Mali, rural infrastructure
was almost nonexistent, and in Guinea, Focus on the burden of Focus on rural areas. • Higher markup and • Exemptions left to the
health services had almost totally disap- diseases of the poor: Larger subsidies to co-payments on discretion of
malaria, diarrhea, poorer regions diseases with lower communities
peared—except in the capital city, respiratory infections, levels of priority • Exempted categories
Conakry—during the last years of the malnutrition, • High subsidies for include widows,
Sekou Toure regime. The vast majority of reproductive health child health services orphans
poor families in the three countries did • Free immunization
and oral rehydration
not have access to drugs and professional therapy as well as
health services. National immunization promotion activities
coverage was under 15 percent, and less
10_pgs 76-77_Ch04Spotlight.qxd 8/14/03 3:16 PM Page 77
Figure 1 Under-five mortality has been reduced in Mali, Benin, and Figure 2 Improvements in under-five mortality
Guinea, 1980–2002 among the poor in Mali
Deaths per thousand Deaths per thousand births
450 400
Mali poor 1987
400 350 1996
Mali average Guinea average
350 300 2000
Guinea poor
300 Benin poor 250
250 200
Benin average
200 150
150 100
100 50
1972 1976 1980 1984 1988 1992 1996 2000 0
Source: Krippenberg and others 2003. Calculated from Demographic and Health Survey
Poorest Second Middle Fourth Richest
data for Benin 1996 and 2001; Guinea 1992 and 1999; and Mali 1987, 1996, and 2001. fifth fifth
Source: Calculated from Demographic and Health Survey
data 1987 and 1996 (based on births in the last five years
before the survey).
tality declined significantly, even among the fied with the quality of care, although 48 Community financing—a seat
poorest. The poor-rich gap narrowed in the percent were not “fully” satisfied. Health at the table
three countries (figures 1 and 2). In Guinea, care users found the availability of drugs to The community financing of key opera-
the decline was steepest for the rural popu- be high (over 80 percent said drugs were tional costs bought communities a seat at
lation and poorer groups. available) and the overall quality of care to the table. Donors and governments had to
Immunization levels increased in all be good (91 percent). systematically negotiate new activities with
three countries.214 They are very high in Greater access reduced travel costs, and community organizations. Governments in
Benin, close to 80 percent—one of the the availability of drugs reduced the need to all three countries, with the support of
highest rates in Sub-Saharan Africa. Immu- visit distant sources of care. Prices have donors, continued to subsidize health cen-
nization rates are lower in Guinea and Mali, been kept below those of alternative ters, particularly to support revolving drug
largely because of problems of access (fig- sources. In Benin the median household funds in the poorest regions. In Benin and
ure 3). Coverage of other health interven- spending on curative care in a health center Mali today the public subsidy to health ser-
tions also increased. The use of health ser- was $2 in 1989, less than half that at private vices is about the same per capita for rich
vices by children under five in Benin providers ($5) or traditional healers ($7).217 and poor regions. In Guinea, however, pub-
increased from less than 0.1 visit per year to Poor people still saw price as a barrier.218 lic spending has benefited richer groups
more than 1.0. In Mali exclusive breastfeed- And a large proportion of the poor still do most. But all three countries face the chal-
ing and the use of professional services for not use key health services in all three coun- lenge of emphasizing household behavior
antenatal care,215 deliveries, and treatment tries. In Benin and Guinea the health system change and protecting the poorest and
of diarrhea and acute respiratory infections allowed for exemptions, and most health most vulnerable. Establishing mechanisms
increased for all groups, including the centers had revenue that they could have to subsidize and protect the poor remains a
poorest (figure 4).216 used to subsidize the poorest, but almost priority of the current reform process.
In an independent evaluation in 1996 in none did. Management committees typi-
Benin, 75 percent of informants were satis- cally valued investment over redistribution.
75 Guinea 80
1996
Mali 60 2000
50 1987
40
25
20
0
1988 1990 1992 1994 1996 1998 0
Poorest Second Middle Fourth Richest
Sources: World Health Organization, UNICEF, and
fifth fifth
Demographic and Health Survey data.
Source: Analysis of Demographic and Health Survey data.