Transcript Public Spending

World Development Report 2004
Making Services Work
for Poor People
Qualidade do Gasto Publico no Brasil
June 26-27, 2003
Ritva Reinikka
Development Research Group, The World Bank
Messages
• Services are failing poor people.
• But they can work. How?
• By empowering poor people to
– Monitor and discipline service providers
– Raise their voice in policymaking
• By strengthening incentives for service
providers to serve the poor
MDGs—Global Aggregates
Eradicate poverty and hunger
Eradicate Poverty & Hunger
Achieve Universal primary education
Universal Primary Education
People living on less than $1 a day
30
Net primary enrollment
100
29
100
2 3 .4
95
20
90
1 4 .5
10
85
84
81
0
80
1990
1998
2015
1990
1999
2015
MDGs—Global Aggregates
Promote gender equality and empower women
Promote Gender Equality
Reduce child mortality
Reduce Child Mortality
Under-5 mortality rate
(per 1,000 live births)
Ratio of girls to boys in primary and
Ratio of girls to boys
secondary school (%)
in primary and secondary school (%)
100
100
100
86
80
89
78
83
80
60
60
40
40
29
20
20
0
0
1990
1999
2005
1990
1999
2015
Outcomes are worse for poor people
Cumulative deaths per 1000 births
Bolivia 1997
Armenia 2000
200
200
150
150
100
100
50
50
0
0
0
1
2
3
4
Years since birth
Poorest
quintile
Richest
quintile
0
5
CAR 1994-95
200
150
150
100
100
50
50
0
0
1
2
3
4
Years since birth
Source: Analysis of Demographic and Health Survey data
5
Cambodia 2000
200
0
1
2
3
4
Years since birth
5
0
1
2
3
4
Years since birth
5
Outcomes are worse for poor people
Percent aged 15 to 19 completing each grade or higher
India 1998-99
Egypt 2000
100
80
60
40
20
0
100
80
60
40
20
0
Richest
quintile
Poorest
quintile
1 2 3 4 5 6 7 8 9
1 2 3 4 5 6 7 8 9
Grade
Grade
Niger 1998
Peru 2000
100
80
60
40
20
0
100
80
60
40
20
0
1 2 3 4 5 6 7 8 9
1 2 3 4 5 6 7 8 9
Grade
Grade
Source: Analysis of Demographic and Health Survey data
Growth is not enough
Percent living on $1/day
Primary completion rate (%)
Under-5 mortality rate
Target
Target
Target
2015 growth
alone
2015 growth
alone
2015 growth
alone
East Asia
14
4
100
100
19
26
Europe and
Central Asia
1
1
100
100
15
26
Latin
America
8
8
100
95
17
30
Middle East
and North
Africa
1
1
100
96
25
41
South Asia
22
15
100
99
43
69
Africa
24
35
100
56
59
151
Increasing public spending is not enough
100
Primary completion rate 1999
(% deviation from rate predicted by GDP per capita)
80
Malawi
Azerbaijan
China
60
40
Zambia
Nepal
Eritrea
20
0
-20
Zimbabwe
India
Mauritius
Czech Rep.
Albania
Lao PDR
Hungary
Argentina
Romania
Kenya
Mexico
Botswana
Poland
Korea, Rep.
Philippines
Slovak Rep.
TogoJamaica
Chile
TTobago
Peru Turkey Nicaragua Gambia
Venezuela
Costa Rica Malaysia Macedonia, FYR
Belarus
Uganda
Ethiopia
Sudan
Thailand
El Salvador
Kuwait
Oman
Paraguay
Mozambique
Brazil
Senegal
-40
Congo, Rep.
Burkina Faso
Morocco
Mauritania
Benin
Mali
-60
-80
Regression line shown:
Coefficient = .184
T-statistic = 1.91
Guinea
Chad
-100
-120
-100
-80
-60
-40
-20
0
20
40
60
Per child public spending on education 1990s average
(% deviation from expenditures predicted by GDP per capita)
80
100
120
Increasing public spending is not enough
150
Under-5 mortalty rate 2000
(% deviation from rate predicted by GDP per capita)
South Africa
Angola
100
Cote d'Ivoire
Saudi Arabia
Namibia
Brazil
Cameroon
Papua New Guinea
Turkmenistan Zambia
50
Zimbabwe
0
Lesotho
Mozambique
Malawi
Macedonia, FYR
Tajikistan
Cambodia
Indonesia
Uzbekistan
-50
Mongolia
Moldova
Korea, Rep.
Malaysia
Slovenia
China
Georgia
Syrian Arab Republic
Croatia
Poland
Slovak Republic
Albania
Bulgaria
-100
Czech Republic
Vietnam
Sri Lanka
-150
-150
-100
-50
0
50
Per capita public spending on health 1990s average
(% deviation from expenditures predicted by GDP per capita)
100
150
Similar changes in public spending can be
associated with vastly different changes in
outcomes
Public spending on prim ary
education per child
Prim ary com pletion rate
10
60
Ethiopia
8
50
6
40
4
30
2
Malawi
20
Malawi
Ethiopia
0
10
1980s
1990
1990s
1999
1980s
1990
1990s
1999
and vastly different changes in spending can
be associated with similar changes in
outcomes.
Public spending on prim ary
education per child
Prim ary com pletion rate
Thailand
100
350
300
250
90
Thailand
80
200
Venezuela
Venezuela
70
150
60
100
50
50
1980s
1990
1990s
1999
1980s
1990
1990s
1999
How are services failing poor
people?
• Public spending usually benefits the rich,
not the poor
Expenditure incidence
Health
All health spending
Education
Primary level
All education spending
Guinea
1994
Guinea
1994
Kosovo
2000
Kosovo
2000
Armenia
1999
Armenia
1999
Nicaragua
1998
Nicaragua
1998
Ecuador
1998
Ecuador
1998
Cambodia
1996/97
Cambodia
1996/97
India
1995/96
India
1995/96
Brazil
1997
(NE&SE)
Brazil
1997
(NE&SE)
Cote
d'Ivoire
1995
Cote
d'Ivoire
1995
Morocco
1998/99
Morocco
1998/99
Bangladesh
2000
Bangladesh
2000
Mexico
1996
Mexico
1996
Bulgaria
1995
Bulgaria
1995
Kenya
1992
Kenya
1992
Costa Rica
1992
Costa Rica
1992
Romania
1994
Romania
1994
0
20
Percent
Source: Compiled from various sources
40
0
20
Percent
40
0
20
Percent
40
Primary level
Percent to
richest quintile
Percent to
poorest quintile
0
20
Percent
40
How are services failing poor
people?
• Public spending usually benefits the rich,
not the poor
• Money fails to reach frontline service
providers
– In Uganda, only 13% of non-wage recurrent
spending on primary education reached primary
schools
How are services failing poor
people?
• Public spending benefits the rich more than
the poor
• Money fails to reach frontline service
providers
• Service quality is low for poor people
Examples of low service quality
• Bangladesh: Absenteeism rates for doctors in
primary health care centers: 74%.
• Indonesia: Absenteeism in primary schools: 18%;
in primary health clinics: 42%.
• Zimbabwe: 13% of respondents gave as a reason
for not delivering babies in public facilities that
“nurses hit mothers during delivery”.
• Guinea: 70% of government drugs disappeared.
But services can work
• Infant mortality and malnutrition reduced in
Ceará, Brazil
• Citywide services in Johannesburg, South Africa
reformed
• Cash transfers to households in Mexico increased
enrollment, lowered illness cases
• Citizen report cards in Bangalore, India
• More money reached primary schools in Uganda
A framework of
relationships of accountability
Client
Provider
A framework of
relationships of accountability
Policymaker
Client
Provider
A framework of
relationships of accountability
Policymaker
Client
Provider
Mexico’s PRONASOL, 1989-94
• Large social assistance program (1.2% of
GDP)
• Water, sanitation, electricity and education
construction to poor communities
• Limited poverty impact
– Reduced poverty by 3%
– If better targeted, could have reduced it by 64%
A verag e P er C ap ita P R ON A S OL E xp en d itu re A cco rd in g to P arty in M u n icip al G o vern m en t
450
400
Real pesos of 1995 per capita
350
300
PRI
PAN
PRD
O THER
250
200
150
100
50
0
1989
1990
1991
1992
year
1993
1994
A framework
Policymaker
Client
Provider
Policymaker-Provider:
Contracting NGOs in Cambodia
• Contracting out (CO): NGO can hire and fire,
transfer staff, set wages, procure drugs, etc.
• Contracting in (CI): NGO manages district,
cannot hire and fire (but can transfer staff), $0.25
per capita budget supplement
• Control/Comparison (CC): Services run by
government
12 districts randomly assigned to CC, CI or CO
Utilization of Facilities by Poor
People Sick in last month, %
35
30
25
20
Baseline
Follow-up
15
10
5
0
Control
CI
CO
A framework
Policymaker
Client
Provider
FSSAP Bangladesh
• Criteria:
– Attendance in school
– Passing grade
– Unmarried
• Girls to receive scholarship deposited to
account set up in her name
• School to receive support based on # of
girls
Decentralization
National
Policy
Local Policy
Client
Provider
Client-Provider:
EDUCO Program in El Salvador
• Parents’ associations (ACEs)
– Hire and fire teachers
– Visit schools on regular basis
– Contract with Ministry of Education to deliver
primary education
EDUCO promoted parental involvement… …which boosted student performance
(% increase in test scores per visit)
5
Percent increase
Number of visits
5
4
3
2
5.71
6
5.65
6
1.56
4
3.78
3
2
1
1
0
Traditional
schools
Educo schools
0
Math
Language
EDUCO Effect: School days
missed due to teacher absence
0.00
Avg # of days
missed: 1.34
-0.50
All yrs
-1.00
-1.50
-2.00
-2.50
-3.00
-3.01
-3.50
What not to do
• Leave it to the private sector
• Simply increase public spending
• Apply technocratic solutions
What is to be done?
• Expand information
• Tailor service delivery arrangements to
service characteristics and country
circumstances
Eight sizes fit all?
Difficult to
monitor
Easy to monitor
Eight sizes fit all?
Difficult to
monitor
Heterogeneous
clients
Homogeneous
clients
Easy to monitor
Eight sizes fit all?
Difficult to
monitor
Heterogeneous
clients
Pro-rich politics
Pro-poor politics
Homogeneous
clients
Pro-rich politics
Pro-poor politics
Easy to monitor
Eight sizes fit all?
Difficult to
monitor
Heterogeneous
clients
Easy to monitor
Pro-rich politics
Pro-poor politics
Homogeneous
clients
Pro-rich politics
Pro-poor politics
Government
provision or
contracting
Eight sizes fit all?
Difficult to
monitor
Heterogeneous
clients
Easy to monitor
Pro-rich politics
Pro-poor politics
Homogeneous
clients
Pro-rich politics
Pro-poor politics Centralgovernment
provision
Government
provision or
contracting
Eight sizes fit all?
Difficult to
monitor
Heterogeneous
clients
Pro-rich politics
Easy to monitor
Demand-side
subsidies, copayments by
households
Pro-poor politics
Homogeneous
clients
Pro-rich politics
Pro-poor politics Centralgovernment
provision
Government
provision or
contracting