Transcript Modelling Fiscal Implications of Education Policies

Part 2
Modeling Fiscal Implications of
Education Policies
Sajitha Bashir
April 25, 2007
Public Finance Analysis and Management
Course, World Bank
Structure of the
Presentation
 Why model fiscal implications of education
policy?
 Structure of models
 Choice of Scenarios
 Examples – DRC and Benin
 Limitations
What will NOT be
covered…
• How to build a model
• Building a model is a technical exercise;
takes time and care. But it is a tool which
can be done by a technician
• For the PER author, what is important is to
understand how to use this tool
Why Undertake Fiscal
Modeling?
 PER analysis should reveal areas where public
resources are
 not aligned with government objectives;
 not used efficiently;
 do not promote equity
 Government Education Plan/Strategy sets out
objectives and strategies
Usually no costs, especially costs
No implementation schedule
Usefulness of Fiscal
Modeling in PER
 Identifies fiscal impact of measures recommended to
improve efficiency
 Which measures create more fiscal space?
 Which measures are under control of policy maker?
 Assess realism and feasibility of proposed plan, its
objectives, strategies and implementation
 Fiscal sustainability; Managerial feasibility
Has impact on policy discussions, especially with Ministry of
Finance
Different Modeling Approaches
Aggregate Fiscal Discipline
Simulation Model
1.
1.
Set Sectoral Objectives and
Strategies (PER/sectoral
analysis/sectoral plan)
2.
Estimate costs
3.
Check macro/budget
implications
4.
Estimate domestic resource
gap – compare with external
financing
5.
Iterations – come up with
realistic resource gap
2.
3.
4.
5.
Sectoral Expenditure Envelope
set by MOF (3-5 years);
usually as part of MTEF
Within Sector: determine
priorities, objectives, strategies
(PER; sectoral analysis)
Cost Strategies
Is it consistent with resource
availability?
Iterations – alternative
strategies; suggest savings
Simulation Model
Purpose
 Evaluate tradeoffs required to arrive at fiscally sustainable
and technically sound educational strategy consistent with
government objectives for coverage, quality, equity
Method
 Develop different scenarios with varying assumptions
Results
 Evolution of expenditures by type
 Evolution of education system (pupil numbers, staff,
schools, classes…)
Structure of Model
 Spreadsheet – all quantifiable variables of
education system are linked to each other
 Five categories of elements
Base year data
Objectives
Assumptions about macro environment
Policy parameters
Results
Simple or Complex
Models?
 Model whole education sector?
Usually desirable to see sub-sectoral trade-offs
 Level of complexity should be determined
by purpose of exercise and results of
sectoral analysis
If focus is on primary, more detailed strategies
at primary level
Building Scenarios
• Input data are fixed
• However many other things can vary:
– Assumptions regarding macro environment
– Policy Objectives
– Policy parameters
Limit Number of Scenarios
 Macro Assumptions x Objectives x Policy
Parameters = potentially scores of
scenarios
 Choose 3 –5 scenarios!
Judgment is required – base selections on
PER/sectoral analysis
What are the critical decisions confronting the
government?
Macro Assumptions
 Economic growth
Determines public receipts, public expenditures
 Demographic growth
Determines growth of child population entering
primary school
Usually invariant across scenarios
Sector Objectives
 Pre-primary
 Population coverage
 Primary
 Entry and completion rates (usually 100%)
 Secondary and higher
 Transition rates
Years by which objectives are to be achieved can
also vary
Key Policy Parameters (1)
 Internal efficiency
 Repetition and drop out rates
 Service delivery targets (access/quality)
 School availability (proximity to habitation) and size
 Teacher pay (by category of teacher)
 Pupil-teacher ratio
 Ratio of teachers to non-teaching staff
 Use of multigrade teaching
 Spending on non-salary items
 Year for attainment of target is also variable
Key Policy Parameters (2)
 Construction
 Type of construction (community?)
 Financing
 % of enrolment in private sector (residual determines
maximum for public financing)
 Set public financing as ratio of domestic resources
 Household financing in public sector (by category of
expenditure and sub-sector) – reasonable in relation to
household income?
 External financing (by category of expenditure and subsector) - realistic?
Illustration – Democratic Republic of Congo
 Challenges: limited public resources; high
dependence on private financing; low coverage
even at primary level but rapid growth at other
levels; inefficiency in public spending
 Policy issues: expansion of post primary levels;
abolition of fees; raising teacher salaries
 EFA plan sets ambitious objectives and
strategies which are not costed
DRC- Common Assumptions of Scenarios
2004
2015 – all scenarios
6 yrs – 2%
6-11 yrs – 2%
12-17 yrs – 2%
population – 1.5%
6%
20%
9%
7% (from 2005)
22% (from 2006)
14% (from 2015)
Population growth
Macro economic framework
GDP growth rate
Public expenditure/GDP ratio
Public receipts/GDP ratio
Percentage enrolled in private institutions
Pre-primary
98% 10% in scenario 1
98% in all other scenarios
Primary
10% 10%
Secondary
11% 20%
Higher
19% 25%
Key Policy Choices Reflected
in Scenarios
• Universal pre-school?
• School feeding ?
• Change some service delivery parameters
(staffing norms etc)?
• Trade-off between rapid quantitative
expansion and quality improvement in postprimary
Scenario
Scenario
1
Policy Variables
Expanded Access and
High Quality at all levels
Description






Scenario
2
Limited Access to Preschool

Universal completion of primary education by 2015
universal access to pre-primary education by 2015
continuation of the existing high transition rates between
primary and secondary, and secondary and higher education.
. provision of canteens in all primary schools and food aid to 30
% of students
substantial quality improvement at all levels is envisaged
through high investment and recurrent expenditures,
rehabilitation of infrastructure, elimination of double shifting
maximum class size norms are enforced, in order to reduce
overcrowding in classes that exceed the maximum norm
access to pre-primary education is limited to current level (14
percent of class 1 entrants, according to MICS 2001)
 all other parameters same as Scenario 1
Scenario
Multigrade teaching and
 multigrade teaching in primary
3
rationalisation of staffing
 revision of staffing norms and reducing disparities between
norms
Kinshasa and other provinces
 all other assumptions same as in Scenario 2
Scenario
Enrolment management at
 transition rate between lower secondary (tronc commun) and
4
post-primary levels
higher secondary is reduced
 transition rate between higher secondary and higher education is
reduced
 all other assumptions same as in Scenario 3
Note: All scenarios assume universal admission to class 1 by 2007 and universal completion of primary education by
2015.
DRC- Cost Saving Measures of Scenarios
3 and 4
2004
Primary
Average class size
Pupil-teacher ratio
Multigrade in small schools
Maximum size of multigrade
classes
Rate of substitute teachers a
Ratio Classes/(Directeurs
and Directeurs-adjoints
without teaching charge)
Ratio Classes/(Workers and
guards)
Secondary
Average class size
Pupil-teacher ratio
Teacher/non-teacher ratio
Scenario 1
Scenario 2
Scenario 3
Scenario 4
35
34
No
36.2
35.3
No
36.2
35.3
No
36.5
34.8
Yes
36.5
34.8
Yes
N.A.
N.A
N.A
35
35
2.2%
2.4%
2.4%
5.0%
5.0%
8.3
13
13
29
29
65
148
148
35
35
26
16
5.2
36.7
22.1
6.5
36.7
22.1
6.5
39.6
23.8
4.8
39.9
24.5
4.0
Higher
Pupil-teacher ratio
20
20
20
20
20
Teacher/non-teacher ratio
0.7
0.7
0.7
1.0
1.0
a
Note : The rate is equivalent to the ratio of substitute teachers to teachers teaching in class. Substitute
teachers comprise extra teachers (surnuméraire) and the replacement teachers for women teachers on
maternity leave (instituteurs de relève)
DRC- Impact on Education Indicators
2001-02 or
status quo
2015 (unless otherwise indicated)
Scenarios 1-3
Coverage Indicators
Primary GER
64%
106%
Primary NER
42%
90%
Secondary GER
23%
44%
Higher education students/100000
358
834
population
Size of education system
Number of students
Primary
5 464 261
13 330 800
Secondary
1 615 443
4 619 108
Higher
197 285
599 755
Number of teachers
Primary
145 225
339 391 (344 863)
Secondary
94 618
168 198 (155 844)
Higher
7 899
22 521
Number of graduates leaving the education system in millions (2009-2022)
Primary
1.73
4.18
Lower secondary
2.44
1.13
Higher secondary
1.05
3.40
Undergraduates
0.09
0.14
Post graduates
0.20
1.02
Satisfaction of labor market needs (2009-2022)
Lower secondary
932%
105%
Higher secondary
179%
382%
Undergraduates
204%
214%
Post graduates
729%
2295%
Scenario 4
106%
90%
33%
226
13 330 800
3 478 275
162 110
344 863
113 742
6 078
5.11
6.40
2.82
0.29
0.20
432%
391%
371%
403%
Scenarios
cen. 1
cen. 2
cen. 3
cen. 4
DRC- Expenditure Requirements
(FC and 2001US $ )
Cumulative Expenditures on Education 2005-2014
Education Expenditure in 2015
Total (all
In public sector Investment in External Total
Public
Domestic
sources, public
institutions public sector Aid
Exp
Expenditure
public
+ private)
institutions
(including expenditure
external
aid)
FC
US$
F C US $ % of public
US $ % of
% of total
% of state
(bill)
(bill)
(bill) (bill) sector exp.
(bill) GDP
education
budget
expenditure
4 793 13.0
4 287 11.5
38%
4.9
17.5%
73%
35 %
3 760 10.2
3 188 8.6
30%
3.1
14.9%
71%
28 %
3 670 9.9
3 097 8.4
31%
3.1
14.5%
70%
26 %
3 126 8.5
2 626 7.1
31%
2.7
11.2%
70%
19 %
DRC - Preliminary Conclusions
 Universal pre-school is not feasible
 Staffing rationalization/use of multigrade teaching
yields considerable savings
 Reducing transition rates in post primary
education is still required
Scenario 4 is most acceptable:
 Examine relative unit costs and composition of
expenditures to further assess suitability
Other trade-offs are possible…
 Eliminate school feeding - expensive even when
targeted to 30 % of pupils
What is its objective ? (increase attendance?
improve student attentiveness?) Are resources
better used elsewhere – e.g., to raise teachers’
salaries?
 Raise pupil-teacher ratio
 Stagger construction
 Raise private financing share in higher
education (but equity trade-off)
Benin – Issues
 Primary GER – 97 % but high disparities
between regions, gender and social groups
 Quality very low – less than 10 percent of 3rd
graders could read with comprehension
 Primary completion rate – 46 %
 Repetition rate – 36 % in final primary grade
 Less than 2 % of domestic education budget on
books and teacher training
 Very rapid growth in higher education (mainly
private, but also public)
Benin – Objectives of
Modelling
• Not to evaluate policy choices for a plan
• Simulation model, used in context of PER,
was used to identify the main issues to be
addressed by education policy
Large Differences in Salaries of
Primary Teachers
Salary per teacher in $
(% of per cap GDP)
Primary teachers
% of total teachers
1 860
(4.9)
100
Permanent
3 011
(7.9)
56.1
Contractual
750
(2.0)
19.9
Community
300
(0.8)
23.9
Francophone Africa
(6.3)
Anglophone Africa
(3.6)
Distribution of schools by pupilteacher ratio in grade 1
Distribution of Class Size in Grade 1, 2002-03
1000
800
600
400
200
185
165
145
125
85
105
65
45
5
25
0
Class size
Objectives for Primary Education
Goal
Target
Impact
Primary
completion
100 % (1.76
Classrooms/
teachers
Quality
Pupil-teacher
ratio : 40
Internal
efficiency
Repetition rate : Reduces
10
rooms/teachers
millions pupils by
2015)
Teachers
Expenditure requirements – policy
trade-offs
 Number of teachers will need to double
 Using civil servant teachers at current salary
levels: expenditure needs will multiply by 4
 Using contractual/community teachers:
expenditures will multiply by 3
 A new statute for teachers?
 Post primary transition rate will also need to be
reduced
Example: Algeria
• Simulation model used in context of public expenditure
analysis
• Government Plan proposed massive expansion of higher
education
– Focus on infrastructure building; continue existing policy of free
student food and accommodation
• Modelling identified large impact on recurrent budget of
cost of student accommodation/food
– No hard budget constraint – oil windfall !
• Helped to identify policy focus: should government be
focusing on managing student dormitories or quality
improvement and institutional reform of higher education?
Limitations (1)
• Powerful tool for decision-making
However, the most important first step is to
formulate the policy choices
Hence, not a substitute for analysis and decision
Limitations (2)
• Model links only variables that can be quantified
• Outcomes related to quality improvement (eg
learning achievement) cannot be modelled
– How do you model 1 million “better educated” students
versus 1 million “less educated”students?
• We model the inputs associated with better quality,
hence focus is on costs , rather than outcomes
Major Reforms to Improve Quality – Difficult
to Model
 Example: Benin – French as language of
instruction from class 1 may be impeding
quality
 Alternatives are:
 use local language (18 in Benin)
 use small groups in classes 1 and 2 (teaching aides?)
 Use radio instruction to reinforce learning
 How do you model costs?
 Use some key cost drivers (eg additional teachers;
teacher support; additional materials)
Limitations (3)
• Many policy parameters are set as policy
objectives (eg reducing repetition rate or drop-out
rate)
• But they are not necessarily under the control of
the policy maker
 We don’t clearly understand how provision of
inputs (schools, teachers, etc) impacts on
repetition (make implicit assumptions)
 Many demand side factors affect parameters
Limitations (4)
 Policy parameters with greatest fiscal
impact often the most difficult to change
(e.g., teacher pay policy, ratios of teachers
to administrative staff)
 Often difficult to link with budget
preparation
Conditions for Effective Use of Simulation
Models
 Sound technical team
 Policy makers and managers who use it as a
tool for decision-making
Are results communicated for decision-making?
 Institutional mechanisms for negotiating
resources (for example, with Ministry of
Finance, donors) AND technical capacity
 Model should be updated regularly
Validation and documentation