UUnniitteedd RReeppuubblliicc ooff TTaannzzaanniiaa

PPUUBBLLIICC EEXXPPEENNDDIITTUURREE TTRRAACCKKIINNGG SSUURRVVEEYY FFOORR PPRRIIMMAARRYY AANNDD

SSEECCOONNDDAARRYY EEDDUUCCAATTIIOONN IINN MMAAIINNLLAANNDD TTAANNZZAANNIIAA

FFIINNAALL RREEPPOORRTT,, 88 FFEEBBRRUUAARRYY 22001100

JJEENNSS CCLLAAUUSSSSEENN AANNDD MMUUSSSSAA JJ.. AASSSSAADD

ACKNOWLEDGMENTS

This Public Expenditure Tracking Survey has been conducted on behalf of the Government of the

Republic of Tanzania. The study was implemented under the guidance of the Permanent Secretary to

the Ministry of Education and Vocational Training, Dr. Hamisi O.Dihenga.

A subcommittee was delegated the task to serve as the reference group and oversight committee for

the process by providing professional input during the various phases of the survey. The members of

the subcommittee were Mr. Cyprian Iraba, Mr. Martin Mwanukuzi , Mr. Jumanne Sagini and Ms.

Hadija Maggid from the Ministry of Education and Vocational Training – Mr. Dionis Ndamgoba and

Mr. Pius Mligo from the Ministry of Community Development, Gender and Children – Mr. Lawrence

Madeghe from the Prime Minister's Office, Regional Administration and Local Government – Mr.

Charles Mwamwaja and Mrs. Raheli Ntiga from the Ministry of Finance and Economic Affairs – Mr.

Israel Mwakapalala from the National Bureau of Statistics – the late Mr. Pascal Mdemu from TENMET

– Mrs. Tanya Zebroff from the Department for International Cooperation and Development – Mr.

Anders Frankenberg from the Embassy of Sweden – Mr. Corey Huntington and Ms. Beatrice Omari

from the Canadian International Development Agency.

Another significant provider of input to this survey was Prime Consult International, a company

commissioned by the Ministry of Education and Vocational Training to undertake data collection.

Under the professional management and guidance of their project manager and seven supervisors,

35 enumerators ventured out to the regions, councils and schools enabling us to generate school

level data required for undertaking the analysis.

The list of names could have been extended significantly by including all the professionals that have

assisted us in gaining access to data and information in the Ministry of Education and Vocational

Training, the Ministry of Finance and Economic Affairs including the Accountant General's Office and

regional sub-treasuries, the Prime Minister's Office, the President's Office, the National Bureau of

Statistics, Regional Commissioners and Administrative Secretaries, and District Education Officers

and District Treasury Officers in the 9 regions and 29 councils visited during the study.

Also participants from various organisations and institutions, both government, civil society and

development partner institutions should be acknowledged for their valuable comments and guidance

at stakeholder meetings organised by the Ministry of Education and Vocational Training at various

stages in implementing the survey.

Last but not least, the head teachers and other staff at the schools subject for the survey including

members of the School Management Committees should be acknowledged for their efforts and

assistance during testing of survey tools and implementation of the sample survey.

For all the above assistance, professional guidance and input we wish to extent our sincere

appreciations. Notwithstanding the above, the data collection and compilation remain the

responsibility of the lead consultants and the outcome of the survey with its analyses, conclusions

and recommendations presented in this report are only those of the authors.

Dr. Mussa Assad Jens Claussen

Dar es Salaam, Tanzania Oslo, Norway

8 February 2010 8 February 2010

TABLE OF CONTENTS

List of abbreviations ............................................................................................................................. i

1 Executive summary with main findings and recommendations .................................................. 1

1.1 Objective and scope of survey ............................................................................................ 1

1.2 Approach and methodology ............................................................................................... 2

1.3 Main findings ..................................................................................................................... 3

1.3.1 Primary schools .............................................................................................................. 3

1.3.2 Secondary schools.......................................................................................................... 8

1.4 Conclusions and recommendations .................................................................................. 12

2 Introduction ............................................................................................................................. 17

3 Objectives and scope of survey ................................................................................................. 18

4 Previous studies on resource flows for education ..................................................................... 19

5 Approach and methodology ..................................................................................................... 21

5.1 Main sources of information ............................................................................................ 21

5.2 Approach ......................................................................................................................... 24

5.2.1 Testing and verification phase ...................................................................................... 25

5.2.2 Data collection phase ................................................................................................... 26

5.2.3 Data compilation and analysis ...................................................................................... 27

5.2.4 Reporting and presentation phase ............................................................................... 28

5.3 Sampling .......................................................................................................................... 28

5.3.1 Type of sampling .......................................................................................................... 30

5.3.2 Selection criteria .......................................................................................................... 30

5.3.3 Size of sample .............................................................................................................. 32

5.3.4 Selection of regions and districts .................................................................................. 33

5.3.5 Selection of primary schools ........................................................................................ 34

5.3.6 Selection of secondary schools ..................................................................................... 35

5.4 Some issues related to data and data quality.................................................................... 36

5.4.1 Secondary school data ................................................................................................. 37

5.4.2 Primary school data ..................................................................................................... 39

5.5 Organisation of the survey ............................................................................................... 40

5.6 Quality assurance ............................................................................................................. 41

6 Overview of primary and secondary education ......................................................................... 43

6.1 Primary and secondary education in Tanzania .................................................................. 43

6.2 National objectives for education ..................................................................................... 43

6.3 Primary and secondary education developments ............................................................. 44

6.4 Budget allocations and expenditure for the education sector ........................................... 45

6.5 Government budget alloaction and execution procedures ............................................... 46

6.5.1 Transfers to ministries, councils and schools ................................................................ 47

6.5.2 Other transfers directly to Schools ............................................................................... 48

6.5.3 The grant allocation system for education .................................................................... 49

6.6 Private funding for education ........................................................................................... 51

7 Survey results primary education.............................................................................................. 52

7.1 Data from national records and sample for primary education ......................................... 52

7.2 Flow of resources for primary education .......................................................................... 53

7.2.1 Central government allocations ................................................................................... 53

7.2.2 Education grants .......................................................................................................... 54

7.2.3 Resources for primary schools...................................................................................... 58

7.3 Distribution of resources .................................................................................................. 65

7.3.1 Allocation and transfers of grants to councils ............................................................... 65

7.3.2 Level of expenditure between councils ........................................................................ 66

7.3.3 Allocation of teachers .................................................................................................. 69

7.3.4 Education grants for non-wage inputs .......................................................................... 73

7.4 Private contributions to primary schools .......................................................................... 77

7.5 School level expenditure .................................................................................................. 79

7.6 School level resources and performance .......................................................................... 81

7.7 Distribution and performance related to gender .............................................................. 85

7.7.1 Enrolment of girls......................................................................................................... 85

7.7.2 Teachers ...................................................................................................................... 86

8 Survey results secondary education .......................................................................................... 88

8.1 National records and survey data for secondary education .............................................. 88

8.2 Resources for secondary education .................................................................................. 89

8.3 School level receipts and expenditures ............................................................................. 90

8.3.1 Main sources of revenue for schools ............................................................................ 90

8.3.2 Allocation and cost of teachers .................................................................................... 92

8.3.3 Government grants ...................................................................................................... 96

8.3.4 School level expenditures .......................................................................................... 100

8.4 school level resources and performance ........................................................................ 101

8.5 Secondary education and gender ................................................................................... 104

Annex I - Terms of reference ........................................................................................................... 106

Annex II - Primary School sample .................................................................................................... 118

Annex III - Secondary School sample ............................................................................................... 119

: Page i

LIST OF ABBREVIATIONS

ADEM Agency for the Development of Educational Management

BEDC Basic Education Development Committee

BEMP Basic Education Master Plan

BEST Basic Education Statistics in Tanzania

BOT Bank of Tanzania

CBO Community Based Organisation

CIDA Canadian International Development Agency

CSEE Certificate for Secondary Education Examination

CSO Civil Society Organisation

CSP Country Strategy Paper

DBSPE District Based Support to Primary Education

DED District Executive Director

DEO District Education Officer

DfID Department for International Development

DHS Demographic and Health Survey

DP Development Partners

DRDP District Rural Development Programme

EBG Education Block Grant

ECD Early Childhood Development

EFA Education for All

EMIS Education Management Information System

ESDP Education and Training Sector Development Programme

ESR Education for Self Reliance

ETP Education and Training Policy

FDC Folk Development College

FMIS Financial Management Information System

GBS General Budget Support

GDP Gross Domestic Product

GER Gross Enrolment Rate

GNP Gross National Product

GoT Government of Tanzania

GPA Gross Point Average

GPI Gender Parity Index

: Page ii

HBS Household Budget Survey

ICT Information and Communication Technology

IFMS Integrated Financial Management System

IMF International Monetary Fund

INSET In-service Training

LGA Local Government Authority

LGRP Local Government Reform Programme

M&E Monitoring and Evaluation

MCD Management Capacity Development

MDAs Ministries, Departments and Agencies

MDGs Millennium Development Goals

MKUKUTA Mpango wa Kukuza Uchumi na Kupunguza Umasikini Tanzania (National Strategy

for Growth and Reduction of Poverty)

MoCDGC Ministry of Community Development, Gender and Children

MoEVT Ministry of Education and Vocational Training

MoFEA Ministry of Finance and Economic Affairs

MTEF Medium-Term Expenditure Framework

NBS National Bureau of Statistics

NECTA National Examinations Council of Tanzania

NER Net Enrolment Rate

NFE Non-Formal Education

NGO Non-Governmental Organisation

NSGRP National Strategy for Growth and Reduction of Poverty

P/T ratio Pupil-Teacher Ratio

PAF Performance Assessment Framework

PE Personnel Emoluments

PEDP Primary Education Development Plan

PEFA Public Expenditure and Financial Accountability Assessment

PEFR Public Expenditure Financial Review

PER Public Expenditure Review

PETS Public Expenditure Tracking Survey

PFM Public Financial Management

PFMRP Public Finance Management Reform Programme

PMO Prime Minister’s Office

: Page iii

PMO-RALG Prime Minister’s Office – Regional Administration and Local Government

PO President’s Office

PO-PSM President’s Office – Public Service Management

PO-RALG President’s Office – Regional Administration and Local Government

PRS Poverty Reduction Strategy

PRSP Poverty Reduction Strategy Paper

PS Permanent Secretary

PSLE Primary School Leaving Examination

PSRP Public Service Reform Programme

RAS Regional Administrative Secretary

RC Regional Commissioner

REO Regional Education Officer

SAP Strategic Action Plan

SBS Sector Budget Support

SEDP Secondary Education Development Programme

SEMP Secondary Education Master Plan

SIDA Swedish International Development Agency

SMT School Management Team

SWAP Sector-Wide Approach to Programming

TA Technical Assistance

TASAF Tanzania Social Action Fund

TC Teachers College

TDMS Teacher Development Management Strategy

TEA Tanzania Education Authority

TEMP Teacher Education Master Plan

TEN/MET Tanzania Education Network/Mtandao wa Elimu Tanzania

TGNP Tanzania Gender Networking Programme

TIE Tanzania Institute of Education

ToR Terms of Reference

TSC Teachers’ Service Commission

TSD Teacher Service Department

TSh Tanzanian Shilling

TSM Takwimu za Shule za Mzingi (data collection form Primary Schools)

TSS Takwimu za Shule za Sekondari (data collection form Secondary Schools)

: Page iv

TTU Teachers Trade Union

TVET Technical and Vocational Education and Training

UDsM University of Dar es Salaam

UNDP United Nations Development Programme

UNESCO United Nations Educational, Scientific and Cultural Organization

UNICEF United Nations Children’s Fund

UPE Universal Primary Education

URT United Republic of Tanzania

USD United States Dollar

VET Vocational Education and Training

VETA Vocational Education and Training Authority

VTC Vocational Training Centre

WEC Ward Education Coordinator

: Page 1

1 EXECUTIVE SUMMARY WITH MAIN FINDINGS AND RECOMMENDATIONS

1.1 OBJECTIVE AND SCOPE OF THE SURVEY

Tanzania’s expansion in primary and secondary education over the past years has been impressive

by all standards. However, the question is whether the expansion in education infrastructure and

enrolment has been matched with a commensurate increase in resource allocation, that the

resources have reached out to service delivery providers, and in particular schools, and to what

extent it has been implemented without undue sacrifices on quality of education as measured by

students’ performance. To provide some answers to the above, the Government of Tanzania has

commissioned a Public Expenditure Tracking Survey (PETS) for Primary and Secondary Education in

mainland Tanzania.

This report presents the findings of the survey. The survey covers resource allocation and use for

government primary and secondary schools in mainland Tanzania. Through the survey a

comprehensive set of data was collected from a sample of regions, councils and schools that gives a

fair representation of all regions, councils and schools of mainland Tanzania. The survey data were

combined with national records and many of the analyses performed are based on data for all

regions, councils and schools. It covers all resource flows and use including central level grants,

regional and council level contributions as well as contributions from parents and others. It covers

data and analysis of both wage and non-wage spending. Accordingly, it is the most comprehensive

education sector tracking survey in Tanzania commissioned to date.

The survey was commissioned by the Ministry of Education and Vocational Training (MoEVT) jointly

with the Ministry of Finance and Economic Affairs (MoFEA), the Prime Minister’s Office – Regional

Administration and Local Government (PMO-RALG), the Ministry of Community Development,

Gender & Children (MoCDGC), the National Bureau of Statistics (NBS), representatives of Civil

Society Organisations (CSO) and Development Partners (DP), all represented in a subcommittee

chaired by MoEVT overseeing and guiding the implementation of the survey.

The survey was designed to address several issues related to primary and secondary education in

mainland Tanzania (ref. terms of reference in Annex I), among others:

Flow of public funds; Do funds allocated through state and Local Government Authority

(LGA) budgets reach the schools as intended, if not is this due to delays, leakages and/or

other factors?

Equity in distribution; Is there a correlation between allocation/releases and other factors

such as school level, district and regional characteristics?

Private/community/parent contributions; What is the level of community/parent

contributions to primary and secondary schools?

Quality of education; What is the level of dropouts, completion, exam passed and other

tests as compared to resource intensity and other characteristics of the district/school?

This report presents findings in relation to all the above issues. The survey has focused on resource

flows and use at the national, regional, district and school levels; how the amount of resources

: Page 2

allocated is reaching the schools and what they have been used for. It has also included school level

contributions from other sources including contributions from parents.

The survey covered allocations and expenditure for primary and secondary education for the fiscal

year 2007/2008 (FY2008). To enable analysis of links to education sector performance data, data at

national, district and school levels were collected for the school year 2008 and used as an

approximation to fiscal year performance.

1.2 APPROACH AND METHODOLOGY

The survey was implemented through several phases.

1. The Inception Phase was devoted to design of the overall approach and methodology of the

survey. The inception phase included collection and analysis of data from national records to

perform stratified sampling.

2. In the Testing and Verification Phase survey tools were tested in two regions to see if they

reflected accurately what type of data could be made available and how the information is

recorded and presented at different administrative levels and schools.

3. The Data Collection Phase included entry of data into a prescribed format to minimise

errors. It was followed by several test runs to assess data consistency for quality assurance

purposes.

4. The Data Compilation and Analysis Phase produced four databases; one with regional data,

one with district data and one each with primary and secondary school data. These

databases were generated by consolidating data from national records with data from the

sample of regions, councils and schools.

5. The Reporting and Presentation Phase from which this report constitutes the final output.

The sampling was done as a three step approach;

1. Selection of 7 out of the 27 regions in mainland Tanzania based on management and

administrative considerations. The selection was made by the PETS subcommittee.

2. Selection of 26 councils within the 7 regions according to specific selection criteria

developed during the inception phase to ensure that the sample gave a fair representation

of all districts in mainland Tanzania along several dimensions.

3. Within these councils a proportional sample of 283 primary and 75 secondary schools were

selected based on various criteria as described below.

Results of the analysis during the inception phase indicated that the main factor explaining the

variation for primary schools is to be found at council level i.e. the major differences are between

councils and their approach to funding of primary schools. The rural/urban dimension is closely

correlated to allocation per student and the rating of councils according to financial management

performance criteria. This was taken into account when sampling councils.

For secondary schools the main fund manager was the school itself. In FY2008 transfers were made

from Regional sub-treasuries although MoEVT also made direct procurement on behalf of secondary

schools. It suggested that a representative number of secondary schools was to be selected as a

representative sample in the region since in this case selection of councils was less of an issue.

: Page 3

The assignment was implemented by two international consultants serving as team leaders namely

Jens Claussen, Nordic Consulting Group and Mussa J. Assad, University of Dar es Salaam contracted

by the Department for International Development (DfID). These two consultants had the overall

management responsibility for the implementation of this PETS.

MoEVT contracted a firm, Prime Consult International Ltd. (PCI), Tanzania, to undertake data

collection from the sample of regions, councils and schools using the survey tools designed by the

PETS team leaders. PCI mobilised and managed a total of 35 enumerators supervised by seven

survey team leaders under the overall management of a 'survey manager'.

1.3 MAIN FINDINGS

1.3.1 PRIMARY SCHOOLS

1.3.1.1 FLOW OF FUNDS FOR PRIMARY EDUCATION

According to official data, primary education was allocated approximately 544 bln. TSh through the

state budget in FY2008. With a total of 8.3 million students enrolled in 2008 government primary

schools the total central government allocation for education per primary student was 65,646 TSh

(52.5 USD).

The actual amount transferred and received (central government execution) by councils who

manage primary schools, was 473 bln. TSh equivalent to 87% of the state budget allocation. This

amounts to 57,417 TSh per primary student (45.2 USD). The grants are provided as an Education

Block Grant to pay for teacher salaries and operating costs including a capitation grant to schools, a

special capitation grant indented for schools and development grant. The grants received were used

by councils for funding of primary education, adult education and education administration. In some

councils a minor share was also spent on secondary education.

In 60 councils, expenditure for education exceeded the total education grants received from the

Central Government i.e. the education grants were fully utilised for education sector purposes. In 28

councils the total education grants received were not fully spent. This constituted 1.2% of the total

education grants transferred to all councils. In 44 councils part of the education grants received were

spent on other sectors/purposes than education. This constituted 6.1% of the total education grants

transferred to all the councils. Adjusted for the above, the net amount of total education grants

transferred to councils that was used for education was approximately 52,541 TSh per student (41.4

USD).

The total council expenditure for education was 59,697 TSh per student in Government primary

schools (47.0 USD). The difference between the grants received per student and actual expenditure

per student is made up of other resources like councils' own revenues, contributions from regional

administration and line ministries as well as use of other sector grants received by the councils from

the Central Government.

Actual expenditure on primary education by the councils was approximately 90.1% of total

education expenditure by the councils. This was equivalent to 53,788 TSh per student (42.4 USD).

: Page 4

Of the amounts recorded by councils as primary education expenditure, 84% was for salaries of

school employees. However, not all teachers on the payroll were actually working at the school

which represents a net loss of 5,779 TSh per student (4.6 USD) or 47.0 bln TSh (37.8 million USD) in

total for all schools in mainland Tanzania.

Capitation Grant and Development Grant transferred to primary schools account for 6,436 TSh per

student (5.1 USD). However, the amount actually received by the schools was 6,046 TSh per student

(4.5 USD) i.e. a deviation between transfers and actual receipts equivalent to 6.2 bln TSh (4.9 million

USD) in total for all schools in mainland Tanzania.

The above shows that of the central government wage and non-wage grants allocated for primary

education, 87% was transferred to councils of which 80% was used for education sector purposes

with the balance either remaining unspent in council accounts or spent by the councils on other

purposes than education. When adjusting for the net losses in teachers absence and transfers of

grants from councils to schools, 71% of the initial budget allocation was received by primary schools.

1.3.1.2 DISTRIBUTION OF RESOURCES

The amount of education grants transferred from the central government varies considerably

between councils with allocations between 13 000 and 188 000 TSh per student enrolled. The main

factors determining level of transfers to a council is the number of teachers a council employs and

retains, and its capacity to execute development grants.

Some councils with high Pupil-Teacher ratios (P/T) show a lower level of budget execution i.e. the

councils that from the outset have lower levels of resources for education also display a lower

'capacity' in implementing the education budget initially allocated. They are rural councils among 34

councils identified with lower overall expenditure per student than initially allocated through the

budget.

While some councils received less than 50% of the amount budgeted, some received more than 30%

of the amount budgeted i.e. during the fiscal year a reallocation of the grants took place between

councils. This reallocation does not level out the initial unequal distribution of total resource on per

capita basis, nor per primary school student. The reallocation is a result of councils' ability to utilise

grants allocated i.e. those with low capacity execute less than initially allocated, those with high

capacity execute more than initially allocated. The main factor determining level of executing grants

initially allocated is the ability to employ and retain teachers in the positions allocated.

Education Block Grant allocated from the Central Government to the councils is a mix of

discretionary and formula based grant elements. Allocations for salaries constitute the main element

of the Education Block Grant and is determined by the number of teachers already employed and

the number of new positions allocated across districts. The estimated number of new teachers

entering the market is distributed across councils based on their prior year P/T ratio in an attempt to

reduce the P/T ratio for councils with a ratio above the national target.

When comparing the planned allocation of new teachers with the actual change in numbers of

teachers employed by the councils, the allocation procedure appears to have had limited impact i.e.

councils with high P/T ratio are not able to employ the new teachers in the positions allocated.

Instead many of these positions are shifted to urban councils and schools with lower P/T ratio. As

: Page 5

long as councils with high P/T ratios are not able to employ and/or retain the teachers allocated,

both in terms of numbers and qualifications, there will be limited prospects for changes in total

resource allocations between councils and schools.

There are two main grants transferred to schools from the councils; the Capitation Grant and the

Capital Development Grant. These grants constitute 9% of total school resource inputs and 99.8% of

all government non-wage contributions to the schools.. The former is to support non-wage inputs to

schools, the latter for school infrastructure improvements.

Average capitation grant per student received by the schools was 4,189 TSh per student (3.3 USD).

This is significantly below the 8,000 - 11,000 TSh indicative levels presented in the budget guidelines

to councils for FY2008.

Capitation grants per student vary among schools in the councils with higher levels for rural schools.

In some councils there are significant allocations for some individual schools while others receive

very low levels of capitation grants per student. Only 4% of the schools received capitation grants

per student above 10,000 TSh.

Attempts to equalise non-wage resources through formula based allocations to councils have limited

impact on the actual transfer of capitation grants to schools. Even if the budget guidelines suggest

that councils should allocate a minimum threshold per student enrolled to schools, this is not

implemented by the councils and actual transfers to schools vary significantly.

Several schools also face significant delays in receiving the grants and many receive the first grant

allocation several months into the next school year. While schools in Dar es Salaam on average

recorded their capitation grant transfers already at the beginning of the fiscal year, other urban and

rural schools on average recorded their first transfers at the beginning of the school year. Some

schools did not receive their FY2008 capitation grant before the end of the fiscal year (beginning of

the next school year).

The capital development grant constitutes on average 21% of cash contributions to schools from

councils. It is allocated to schools based on rehabilitation and investment needs. In many councils

some schools receive development grant one year with schools changing each year i.e. 18% of the

schools received capital development grant in FY2008 while the remaining 82% did not receive the

grant. For the schools receiving the grant, the amount can be significant since it is to cover expenses

for desks and chairs as well as construction of classrooms, latrines and other infrastructure.

Some councils procure equipment and/or pay contractors directly for construction of classrooms

and other infrastructure. These contributions are predominantly provided to schools that do not

receive capital development grant in cash. The in kind contributions from councils however

constitute only a minor share of total contributions to schools' capital expenditure. The combined

cash and in kind contributions were transferred to 24% of the schools in the sample.

1.3.1.3 PRIVATE CONTRIBUTIONS TO PRIMARY SCHOOLS

The data on parents cash and in kind contributions are stemming from school records. In terms of in

kind contributions from parents these are estimates based on head teacher and school management

committee inputs. It is fair to assume that these figures are underestimated since many schools do

: Page 6

not keep full records of such contributions, and in particular for contributions managed by the

parents themselves. In addition, schools also receive contributions from NGOs and other private

sector organisations and entities.

In total the private contributions to schools including parents' contributions constituted 3.7% of total

(private and council) cash and in kind contributions to schools and 28.4% of total non-wage

resources. Cash and in kind contributions from parents constituted 1.0% of the total contributions to

schools and 7.7% of non-wage resources.

In 30% of the schools parents made contributions in cash to pay for extra classes, for employment of

additional teachers in addition to teachers on the government payroll and to pay for various

teaching materials and equipment (like desks, chairs, etc.) to supplement contributions in cash and

in kind from the councils.

In 12% of the schools parents also contributed with in kind contributions i.e. the same type of inputs

procured by the school with cash contributions, but in this case it was the parents that organised

collection of funds and procured the inputs.

Urban schools receive more contributions in cash rather than in kind as compared to rural schools

where school management committees/parents to a larger extent manage the cash contributions

themselves and instead procure the inputs. Urban schools receive a significantly higher value of NGO

and other private contributions than rural schools.

1.3.1.4 SCHOOL LEVEL EXPENDITURE

Cash contributions managed by schools are used for funding of 67% of school inputs. The other 33%

are contributions in kind from the councils, parents and others.

The main cost item at school level is construction costs which include construction of new

classrooms, administration buildings, staff houses, latrines etc. and account for 30% of school level

expenditures. These costs are partly funded by cash contributions to schools (51%) or paid by

councils or other contributors directly to contractors (49%). The other main items are textbooks and

other teaching materials which constitute 29% of total school level expenditures mainly paid for by

cash contributions to schools (92%).

Administration and other expenses account for 9% of total school level expenditures. In some cases

parts of these expenditures were for allowances to teachers performing extra classes. The number

of extra classes constitutes 19% of the total number of classes taught in a week. Urban schools

spend less on teachers performing extra classes but on average have higher number of teachers per

student on government payroll than rural schools.

The average expenditure per student on textbooks and teaching materials varies significantly

between schools and most prominently for rural schools. Among the rural schools 4% spent less

than 500 TSh on text books and teaching materials per student and 13% of the schools spent less

than 1000 TSh per student. This stands in contrast to urban schools in which 94% of the schools

spent more than 1000 TSh per student.

: Page 7

There are schools with very limited resources for non-wage inputs and many of them have

simultaneously few and less qualified teachers than others. These are to be found in rural areas of

'rural' councils.

1.3.1.5 SCHOOL LEVEL RESOURCES AND PERFORMANCE

The main indicator for assessing school performance has been the number of students performing

and passing Primary School Leaving Examination (PSLE).

Urban councils (including Dar es Salaam) have better performing schools (as measured by their

average PSLE rank), they have on average lower P/T ratio and spend more on recurrent expenditure

for schools (wage and non-wage inputs). Deployment of teachers is correlated with school

performance when analysing council averages which also serve to explain that better performing

schools have higher recurrent expenditures per student.

Lower P/T ratio (and thus higher spending per student) generally gives a better PSLE pass result for

schools in a council. Spending on non-wage inputs appears to influence school performance

measured by PSLE ranking of schools in a council, but there are significant variations, first and

foremost among rural schools. However, the correlation with one indicator like P/T ratio or non-

wage spending does not alone serve to explain the school performance measured by PSLE ranking.

Council Poverty rates are correlated to school performance when assessing averages for urban

versus rural schools. However, communities within the rural councils are highly diversified in terms

of poverty incidence and so are the school performance, P/T ratios and overall resource inputs.

The rural/urban dimension within a council is correlated with overall school level performance and

also P/T levels. The most remote schools measured by its distance to council headquarters have less

resource inputs in terms of teachers and lower quality of facilities. Many of them spend more money

on non-wage items per student than other schools but it does not compensate for the lower P/T

ratio and quality of facilities when assessing performance rates measured by PSLE pass rates. The

above observations serve to support a recommendation that teacher allocations to rural and

'remote' schools is an issue that needs to be addressed if school performance is to be improved for

these schools.

Council level data on girls' enrolment suggest that there is a fairly equal enrolment of boys and girls

however there are rural councils with lower share of girls enrolled. There is no gender bias measured

by enrolment in terms of P/T ratios, quality of facilities and allocation of resources per student, and

school performance is not correlated with enrolment of girls.

However, the data from the sample of schools reveal that there are significant variations between

schools in a council, in particular among schools in rural councils. In rural councils there are schools

with enrolment rates for girls as low as 25% of total enrolment while at the other end schools with

rates as high as 75%.

The rural schools in rural councils have the highest P/T ratios, the lowest scores on school facilities

and performance rates and it is among these schools the lowest enrolment rates of girls are to be

found. It means that a specific focus on these schools will also have impact not only on primary

education in general but on girls' education in particular.

: Page 8

Cost per teacher is not linked to gender distribution, i.e. there is no evidence in the data to suggest

that the level of enumeration is linked to gender distribution of teachers. There is no evidence to

suggest that other inputs as well as school performance are linked to distribution of teachers by

gender when assessing gender distribution among teachers across schools in all councils. However,

the gender distribution is an issue related to location of schools.

Urban schools on average have a higher number of female teachers to the total number of teachers.

There is a significant gender bias towards urban schools in our sample (77% of teachers are female

teachers) compared to the average of rural schools (36% of teachers are female teachers).

The analysis shows that location of schools in rural councils determines ability to employ teachers,

and in particular female teachers. In total this has impact on school performance as measured by

lower PSLE pass rates despite that they do not necessarily have less resources for non-wage inputs.

1.3.2 SECONDARY SCHOOLS

1.3.2.1 FLOW OF FUNDS FOR SECONDARY EDUCATION

The State Budget allocation for secondary education under the budget heads of MoEVT was TSh

170.2 bln in FY2008. With 1,128,711 students enrolled in secondary schools in 2008 the equivalent

per student allocation was TSh 150,792 (118.7 USD).

The actual execution of State Budget funds for secondary education was 156.5 bln TSh (91.9% of the

amount allocated through the budget) equivalent to 138,610 TSh per student (109.1 USD).

The total amount spent for secondary education according to MoFEA records was TSh 159.1 bln

equivalent to TSh 140,983 per student enrolled (111.0 USD). The higher level of execution than

Central Government releases is due to revenues generated by secondary schools (fees) and other

contributions including contributions from councils. The non-wage spending constituted 50.3% of

total expenditures; equivalent to 70,909 TSh per student (55.8 USD).

The main source of funding excluding personnel on government payroll for secondary schools is

MoEVT who contributed with 75.1% of all cash receipts for the schools.

The government contributions include 14 different grants to cover recurrent costs (excl salaries) and

for school improvement of physical facilities and equipment.

Rural secondary schools received a higher amount of cash contributions per student than urban

schools. The variation between schools are however significant, in particular for community schools

of which many have been established in the last few years.

1.3.2.2 CONTRIBUTIONS FROM PARENTS

Parent contributions in the form of various fees and other cash contributions accounted for 18.7%

with the balance made up of smaller contributions from councils, NGO's and other private

donations.

Contributions from parents are on average the same for parents with students in community and

government schools, however higher for rural than urban schools.

: Page 9

As many as 95% of the schools (including some government schools) in our sample received a fee

subsidy to compensate for fee exemptions granted to parents with low ability to pay. The data show

that for many of these schools parents still give major cash contributions to the school as voluntary

or non-regulated contributions despite being exempted from regulated school fees.

1.3.2.3 ALLOCATION AND COST OF TEACHERS

The allocation of teachers in secondary schools as measured by pupil/teacher ratios varies between

urban and rural schools. On average there are 29 students to a secondary school teacher paid for

from any source. However, when limited to teachers paid for by the government, the P/T ratio is 41.

Urban schools have more teachers in total and teachers on government payroll per student than

rural schools. Schools in rural areas have a higher proportion of teachers that are paid for by the

schools own cash receipts however despite this they have less teachers per student. Community

schools have far less teachers per student than government schools, both overall and in terms of

teachers on government payroll.

Among the secondary schools, and in particular community schools there are some with high level of

grants per student allocated, high levels of parent contributions and thus resources to employ

teachers from own funds (hence a 'low' P/T ratio). However, there are also schools with low grant

allocations per student, lower levels of parent contributions and thus limited resources to employ

extra teachers above what is paid from the government payroll.

Losses in school inputs stemming from teacher absence were recorded in 56% of the schools in the

sample with an overall average of 13% of the teachers in secondary schools absent a significant part

of the school year. Most of these teachers were on government payroll. There are highest incidences

of absence by teachers in rural schools (16%) and in government schools (18%).

Teachers in rural areas have generally lower qualifications than their urban counterparts. Our

sample shows that approximately 30% of the teachers in rural and community schools are only

licensed teachers with Form VI exam as the highest level of education.

The government schools have the highest share of graduate teachers and very few government

schools employ teachers with only Form IV exam, i.e. in terms of teacher qualifications urban and

government schools are able to attract the best qualified teachers.

1.3.2.4 ALLOCATION AND TRANSFER OF CAPITATION GRANTS

In 2008 the Learning Grant and the School Fee Subsidy in total constituted the Capitation Grant for

secondary schools with TSh 26,630 per student. The capitation grants are allocated as a subsidy to

schools to encourage enrolment of students regardless of their ability to pay for school fees and

teaching materials, i.e. minimise the impact of household income for access to secondary education.

For 10.7% of the schools, all students were considered for exemptions and thus received the full

amount of capitation grant compared to students enrolled.

95% of the schools in our sample received capitation grants to compensate for exemptions granted

to parents with low ability to pay. The total amount of capitation grants released was 16% below the

initial allocation that in particular affected the urban government schools in Dar es Salaam.

: Page 10

However, the majority of schools received the capitation grants as initially allocated, for some there

was a reallocation in favour of rural and community schools.

It would be expected that schools with lower cash contributions from parents would receive a higher

allocation of capitation grant per student while for schools with higher cash contributions from

parents the capitation grant per student would be lower. According to data in our sample this

appears not to be the case. In most cases, and in particular for community schools, parents make

significant cash contributions while the schools at the same time receive School Fee Subsidy for most

of the students.

1.3.2.5 ALLOCATION AND TRANSFER OF DEVELOPMENT GRANTS

The Development Grant is transferred to schools for funding of constructions and rehabilitation of

classrooms, teacher houses, libraries, school laboratories, administration buildings, hostels, etc. with

allocations made on the basis of a fixed unit cost per type of investment. The allocations of the

development grants to individual schools are based on needs assessments i.e. those schools that are

considered to be in need of a certain facility.

The average development grant received by the schools per student was TSh 72,699 during FY2008.

However, there are significant variations between schools based on the extent to which they have

been prioritised for an investment in a specific facility.

At the school level, much more is received per student as development grant per student than

capitation grants. The amounts from the Central Government transferred through the sub-treasury

exceeded the figures initially allocated; i.e. additional sources of funding for investment and

rehabilitation of facilities were provided by others (among others councils and regional

administrations).

The major share of the development grants was transferred to rural community schools. For the

government schools a large amount initially allocated was not released however part of the amount

was reallocated for other schools. The total level of transfers to schools was 93.2% of the budget

allocated.

1.3.2.6 SCHOOL LEVEL EXPENDITURES

The major cost item for all community schools is expenditure for infrastructure (new classrooms,

staff houses, library, laboratory, hostels for boarding schools, etc.). The costs however vary

significantly between schools in respect of infrastructure investments since not all were awarded

resources in the form of development (infrastructure) grants in FY2008 and there were different

amounts for different types of infrastructure.

Another significant cost item is payment to teachers (both on government payroll and other

contracted teachers). On average rural schools, and in particular rural community schools, allocate a

significant share of their costs to employ extra teachers over and above the teachers employed on

government payroll. It also includes additional payments to teachers on the government payroll for

extra classes.

: Page 11

School meals are a significant cost item for the average rural government schools. Urban and

government schools allocate more to textbooks and teaching tools while other teaching and

examination materials are more or less equally distributed.

Comparing the expenditure data with grants and other contributions received it is evident that parts

of the development grants have not been fully executed. Part of the amount remains as balances in

school accounts and part of the grants has also been advanced by the schools for non-investment

purposes.

1.3.2.7 SCHOOL INPUTS AND SCHOOL PERFORMANCE

For 2008, Form II pass rates were on average 74% for all schools in the sample and Form IV

examination results displayed an aggregate pass rate of 84% on average for the schools.

There are notable variations in performance between rural and urban secondary schools. While

schools in Dar es Salaam had an average pass rate of 91%, other urban schools had a pass rate of

69% and rural schools a pass rate of 72%.

The difference in examination performance is significant between government and community

schools. While government owned schools have a 90% pass rate on average, community owned

schools have an average pass rate of 71%.

For Form IV results the same pattern is observed. With an overall average pass rate of 84%, the

average for government schools is 96% while community schools have a pass rate of 82%.

Urban government schools are performing better as measured by Form II and IV pass rates while

rural community schools have lower pass rates. It is in rural community schools that a major share of

the expansion in enrolment has taken place during the last years.

Our data suggest that non-wage spending per student is correlated with the performance of a school

(measured by pass rates). It is also linked to the Gross Point Average (GPA).. The more total non-

wage spending per student the higher the average O-Level GPA for a school.

The magnitude of the input from contract teachers is correlated with schools' non-wage resources

from which they are paid. It means that more non-wage finance for schools allow them to contract

more teachers and retain them for a longer period of time..

However, pass rates are not correlated with overall P/T ratios, but with P/T ratios of teachers on

government payroll. Firstly, this is because the utilisation of non-wage resources by community

schools on contract teachers are on account of other none-wage inputs. Secondly, these teachers

have generally lower qualifications than teachers on government payroll.

In total it means contract teachers yield a lower return as measured by performance due to lower

levels of qualifications than teachers on government payroll and lower resource input for none-wage

spending since these resources are also used for paying additional teacher salaries. It suggests,

similar to primary schools, that there is a need to address teacher allocation disparities by providing

for more teachers to rural schools with equal qualifications as their urban counterparts. To raise

performance levels for community schools, and in particular rural community schools they should be

prioritised when allocating more teachers on government payroll and/or promote employment of

contracted teachers with higher qualifications through regulations and/or incentives.

: Page 12

1.3.2.8 SECONDARY EDUCATION AND GENDER

In our sample 42% of the students enrolled in secondary schools are girls with a lower share for rural

schools. Some schools have lower than 20% share of girls (rural schools) i.e. there is a significant

challenge for enrolling more girls for secondary education in rural schools.

The gender distribution of teachers is notably different for secondary schools compared to primary

schools with an overall average of 26% female teachers. This low share of female teachers concerns

first and foremost rural and community schools and some has only male teachers serving.

P/T ratios and resources per student are linked to school location and so is gender distribution in

terms of female students and teachers. Anecdotal evidence suggests that employment of female

teachers has impact on girls’ enrolment in schools.

1.4 CONCLUSIONS AND RECOMMENDATIONS

The most significant findings of this survey are related to allocation and transfer of resources to

schools and their link to school performance. The following are intended as inputs to a discussion on

how to address the main challenges observed from this study; efficiency losses due to observed

weakness in transfers of resources to school levels and the significant disparity in resources and

performance between schools.

The findings can be summarised into some key conclusions and recommendations, many which

apply simultaneously to both primary and secondary education especially since secondary education

has also been devolved to councils after the FY2008. Related to the main issues raised in the terms

of reference for the survey the following conclusions can be drawn:

Flow of public funds;

o For secondary schools non-wage grants reach the schools broadly as intended. For

primary schools however, the amount of grants received is determined by council

management practises and decisions. They vary significantly between councils which

serve to explain why some councils do not fully utilise the grants received or apply

part of them for a different purpose than education..

o The difference in council management practises and capacities also impact on timing

of grant releases, monitoring of teacher attendance and the extent to which grants

transferred to the schools actually are received by them.

o The major share of public funds spent for the benefit of schools are for the

enumeration of teachers. The extent to which a council and/or school actually

employ the teachers for the staff positions allocated is depending on its location.

Schools in rural communities are often not able to employ the staff positions

allocated which also adversely affects its performance as measured by student pass

rates.

: Page 13

Equity in distribution;

o The grant allocation system for primary education system leaving decisions at the

discretion of councils in terms of how much schools should receive, is the main

cause for the significant unequal distribution of resources between primary schools.

o As mentioned above, the ability to attract and retain teachers is also a major

challenge for rural councils and in particular for their schools in rural communities

which is the main factor explaining the inequality in resource input between councils

and schools.

Private/community/parent contributions;

o While private contributions in general are not a significant resource for primary

education, it is so for community secondary schools.

o A large share of the private contributions to secondary schools, and in particular

parent contributions, are used for employment of teachers. However, the

contributions are not sufficient to reduce the P/T ratios to the level of government

secondary schools and teachers employed have generally lower qualifications than

teachers on government payroll.

'Quality' of education;

o Resource intensity determines school performance as measured by pass rates, first

and foremost the number and qualifications of teachers a school has employed

relative to the number of students.

o As mentioned above, schools in rural communities have a challenge in employing

and retaining teachers, and in particular female teachers. This has impact on

enrolment of girls in the schools in rural communities.

Following the above, there are in particular two main issues that need to be considered; the

education grant system and employment of teachers;

1. The central government non-wage grants for secondary education were broadly transferred

as budgeted to secondary schools in accordance with the grant allocation procedures

managed by MoEVT, this despite being a complex grant system of 14 different grants. In

contrast the education grant system for primary education display some major challenges.

These challenges are first and foremost related to the diverse management practises

between councils in their allocation and utilisation of these grants.. This creates significant

disparities in resource allocation for primary education between councils and between

schools within a council. The major challenges are related to the following;

o The grant allocation system consists of a complex mix of formula and discretionary

grants which creates a challenge for many councils in managing their execution of

their education budgets. The budget guidelines did not clarify fully to what extent

grants transferred were earmarked for a specific purpose and/or the actual amount

to be transferred to schools with councils making different interpretations in terms

of how the grants were to be allocated and used.

: Page 14

o Some councils were not able to utilise the grants in full due to weaknesses in

administrative and financial management capacity, some used part of the education

grants for a different purpose or sector and in some councils the transfers of

capitation and development grants to schools did not reach the schools with the full

amount with losses observed in transfers from the council bank accounts to school

bank accounts.

o With additional grants for secondary education devolved to councils the above will

create an additional major challenge that likely will lead to a similar disparity over

time for secondary schools as observed for primary schools.

2. The disparity in terms of resource allocation, performance and gender (both students and

teachers) between schools can first and foremost be attributed to the ability to employ and

retain teachers (primary and secondary), in particular teachers with adequate qualifications

(secondary).

o The current system for allocation of new teachers to schools with high P/T ratio does

not level out the disparity both in terms of numbers and qualifications of teachers.

o Schools in rural communities face a specific challenge in their ability to employ the

teachers they were allocated through the budget, and in particular female teachers.

o The ability to employ teachers with adequate qualifications is a major issue and

particularly for the secondary community schools in rural areas.

o Teacher absence is a major issue not only by resulting in an efficiency loss in

financial terms but impact on school performance.

To address the above issues the following should be considered;

1. The system for allocation of grants to councils needs to be revisited. The system is a complex

mix of formula and discretionary general and earmarked sector grants. It makes it difficult

for a council to determine the level of grants it is entitled to and what additional resources

to expect from achievements in particular sectors; i.e. the system will most likely not have

much merit unless the councils fully understand what amounts they are entitled to, for what

purpose, how much is to be transferred to schools and how much can be used by the

councils themselves. There is accordingly a scope to review and simplify the system if it is to

have real influence on council budget execution.

2. The above requires that the procedures that need to be followed in utilising the grants

received are clearly spelled out in the budget guidelines and other regulations, not least

clearly specify the amount that should be transferred to schools.

3. This study also shows that adjustments to the system and procedures for grant allocations is

one aspect to consider but not sufficient to improve observed weaknesses and change the

disparity in resource allocation between councils and schools. It needs to be combined with

targeted assistance and supervision for councils with low levels of execution.

4. It requires more monitoring of council budget execution to ensure that regulations are

followed in particular for councils that utilise grants for a different purpose than intended

and/or where there are significant delays in release of grants to schools.

5. Using the data on quarterly budget execution by councils compiled by PMO-RALG combined

with education statistics from MoEVT can serve as one source of information for closer

: Page 15

monitoring of compliance and to identify councils which display low levels of budget

execution which then requires specific follow up and supervision.

6. Councils where there are major deviations between transfers to schools and school receipts

also needs more attention. This is not only an education sector issue but also an issue

related to fiscal management in general. These councils display several qualifications in the

reports of the Controller and Auditor General and should be subject to specific follow up and

supervision to identify if it is related to weakness in financial management capacity, non-

compliance with financial and other regulations and/or system of bank transfers.

7. Both Capitation and Development Grants to schools represent small shares of total

education expenditures but nevertheless significant shares of non-wage school level

resources. To ensure that grants, in particular grants to cover recurrent expenditures,

actually reach schools, the grants could be transferred directly to schools rather than via the

councils e.g. direct transfers from MoEVT to secondary schools in FY2008 proved to reach

out to schools while primary schools were depending on the diversified allocation decisions

and/or management practises between councils.

8. One option similar to many other countries could be to introduce 'block grant' for non-wage

inputs directly to schools to ensure that they actually receive the amount they are entitled

to. One important element for ensuring that spending decisions reflect needs at the school

level is an effective decision making arrangement and oversight function. This oversight

function appears already to be established in most schools of mainland Tanzania.

9. To change the significant disparity in P/T and performance among schools the specific

incentives (allowances) for teachers serving at schools located in rural communities should

be introduced. These incentives could be directed at teachers to take up positions allocated

through increased PE ceilings but today not executed since councils and schools are not able

to employ them or retain teachers employed. The incentive would require a two step

approach similar to other countries practising diversified remuneration packages for the

same positions. The first step would be to define rural schools for which the incentive will

apply. The second to ensure that these incentives are applied by councils to equalise P/T

ratios among the schools in the council.

10. Since gender disparity is a major issue and in particular for secondary schools and schools in

rural communities, additional efforts will be needed first and foremost in the education and

recruitment of female teachers and particularly for secondary schools which likely will also

have impact on the level of girls' enrolment.

11. Monitoring of teacher attendance needs to be strengthened. Firstly, more information is

needed to assess reasons for teachers not attending which could be accommodated by a

specific labour survey on teacher attendance and mobility including newly graduates.

Secondly, some of the remedial measures entail capturing and summarization of teacher

attendance at school and council levels using a uniform format of a record. Effective

oversight will also entail formalisation of teacher absence at all times regardless of reasons

for absence and more regular monitoring of councils based on periodic records submitted by

schools.

This study has also shown that there are some challenges and opportunities related to data quality

and their application for sector monitoring and resource allocations;

: Page 16

1. First and foremost this is related to updating school registries to have a full account of all

government schools in mainland Tanzania. These registries should then be used as point of

departure for consolidating data acquired by different institutions for monitoring purposes

and in allocation of grants to schools.

2. Secondly, implementation of procedures for verification of data collected related to number

of teachers and students in schools would contribute to quality assurance.

3. Thirdly, it is related to comprehensiveness in accounting and reporting on sector

expenditures. The latter is partly related to implementation of the Integrated Financial

Management System (IFMIS) at council level and partly related to the amount and details of

information consolidated into the central database.

4. Payroll data maintained at PO-PSM/MoFEA should be enhanced to include information not

only on function, salary level, date of employment and pay station, but also at which

government institution (school) an employee is serving. This will significantly improve the

ability to monitor teacher attendance when consolidating payroll data with data from the

schools and councils.

The above has been elaborated further in a section on data quality in the report.

: Page 17

2 INTRODUCTION

This Public Expenditure Tracking Survey (PETS) for primary and secondary education in Tanzania has

been commissioned by the Ministry of Education and Vocational Training (MoEVT) jointly with the

Ministry of Finance and Economic Affairs (MoFEA), the Prime Minister’s Office – Regional

Administration and Local Government (PMO-RALG), the Ministry of Community Development,

Gender & Children (MoCDGC), the National Bureau of Statistics (NBS), representatives of Civil

Society Organisations (CSO) and Development Partners (DP), all represented at a subcommittee

chaired by MoEVT overseeing and guiding the process.

The objectives of the study were to map and assess:

1. Flow of public funds to primary and secondary education.

2. Private contributions to schools.

3. Equity in distribution of public funds for primary and secondary education.

4. Link between school performance and allocation.

This study is the most comprehensive education sector tracking study in Tanzania commissioned to

date. Previous Public Expenditure Tracking Surveys have been limited in scope by focussing on

specific education sector grants and/or limited samples.

This survey covered both primary and secondary education, and included both wage and non-wage

expenditures. It c collected data from a sample of councils and schools which in size gives a fair

representation of all councils and schools of mainland Tanzania and combined survey data with

national records. Accordingly many of the analysis performed are based on data for all regions,

councils and schools of mainland Tanzania and the report present results from analysis using both

national records and survey data.

The PETS has been designed to present national level aggregates and as such all results are

presented without any specific reference to a region, council or school but rather along council and

school level characteristics (such as rural/urban, revenue per capita, enrolment, school density,

pupil/teacher ratios etc.), i.e. results have deliberately not been presented in a way that allows the

reader to trace results to a specific region, council and school. This approach has added value to the

data collection process since all respondents were made aware that information provided by them

would be treated with discretion and not exposed to the general public. It reduced the level of

suspicion by many respondents of being followed by an inspection should some of the information

collected suggest that fund management is not strictly in compliance with regulations.

In the following sections we present the scope of the study (chapter 3). In chapter 4 a presentation

of previous PETS surveys in education are briefly presented. In chapter 5 we present the approach

and methodology including sampling of regions, councils and schools subject for the survey. In

chapter 6 we present an overview of the sub-sectors before presenting the results from the survey

for primary and secondary education respectively (chapter 7 and 8). A summary with the

recommendations are to be found in chapter 1 of the report.

: Page 18

3 OBJECTIVES AND SCOPE OF SURVEY

The objective and scope of the PETS have been outlined in the Terms of Reference (ref. Annex I).

After the inception phase with initial analysis of data and discussions with the Sub-committee

overseeing the implementation of the PETS, it was decided that the study would focus on the

following four main issues;

1. Flow of public funds responding to questions such as;

o Who spends the public funds allocated through the central and regional budgets?

o What are they spent on i.e. how much spending benefits school level service

delivery directly and indirectly?

o How much of funds allocated reaches the 'school' and how do they spend the funds?

2. Private contributions to schools; i.e. what is the level of parent/community/other

contributions to schools?

3. Equity in distribution; i.e. is there a correlation between allocation per student and other

school and district level characteristics?

4. Link between school performance and allocation; i.e. is there correlation between allocation

per student and school level performance indicators?

The survey has included analysis of wage and non-wage related spending which means analysis of

teacher salaries, all recurrent and capital expenditures including spending for expanding existing

schools by construction of new classrooms and buildings. It has included both primary and

secondary education expenditures.

Analysis of linkages between expenditure and student/school level performance has been based on

data on number of student passing exams, students' attendance and participation to the extent such

data was available. Quality of some inputs and services measured by qualification of teachers,

quality of physical facilities e.g. distribution of various equipment and infrastructure per school, has

also been analysed based on data collected from the sample of schools subject for this survey.

It was beyond the scope of this survey to measure teacher and/or parents/students perceptions of

access to and/or the quality of services and facilities. If to be included it would have required a

technical assessment of facilities and information from parents, students and community

representatives which would have required substantially more time and resources allocated to

include information from.

The study focus on government and community schools receiving public funding. Private schools

have not been included in the study.

The fiscal data for the survey have covered the fiscal year 2007/2008 (FY2008). The school year runs

from January through December (the calendar year). When analysing correlation between school

performance data and revenue/expenditure data, performance indicators for the school year 2008

have been used as an approximation.

: Page 19

4 PREVIOUS STUDIES ON RESOURCE FLOWS FOR EDUCATION

There have been several attempts to analyse the link between education sector spending and

outcomes most notably in the Public Expenditure Reviews (PER) and Public Expenditure and

Financial Accountability Assessments (PEFA). Furthermore, several efforts have been made to track

public expenditure in education and assess the extent to which earmarked transfers for specific

education sector expenditures are managed as intended and resulting in intended outcomes

(PriceWaterhouse Coopers, 1999; REPOA and ESRF, 2001; Björkman and Madestam, 2003; REPOA,

2004).

The 1999 PETS covered two sectors (education and health) and employed a limited sample of only

three districts in three regions. These were Kondoa (Dodoma), Kiteto (Arusha) and Hai (Kilimanjaro).

The sample was not representative and therefore findings had little scope for aggregation. The study

sought to track funds and resources in kind from the Central Government as well as resources

collected at the district and frontline facility level. The financial years tracked were 1996/97,

1997/98 and 1998/99.

The study provided useful insights – it reported that only 43 percent of transfers reached the

schools. However, perhaps the size of the sample limited its relevance at the national level. This

study features very little in discussions, subsequent studies as well as in the literature suggesting

that it was not well publicised and provided limited input to decision making.

The second PETS in Tanzania was jointly executed in 2001 by two Tanzanian foundations; Research

on Poverty Alleviation (REPOA) and the Economic and Social Research Foundation (ESRF) – two well

established research organizations with reputation on policy research. The study is referred to as a

‘pro-poor’ PETS because it paid close attention to the pro-poor policy priorities of the Government

of Tanzania (GoT) that commissioned the survey. The study covered five districts (Regions in

brackets) – Babati (Arusha), Kisarawe (Pwani), Mtwara Urban (Mtwara), Dodoma Rural (Dodoma)

and Kigoma Urban (Kigoma). Its sector coverage was more extensive and included primary

education, primary health care, water and rural roads. It covered the financial years 1999/2000 up to

and first half of the financial year 2000/2001.

Resembling the earlier study this attempt was also short on analysis although it made interesting

observations on so called 'leakages' (less than 50% of transfers reaching schools) and excessive

expenditure on district level discretionary expense items. Again, like its predecessor this study was

not made widely available and referenced in analytical work in sector level Public Expenditure

Reviews (PER). Consequently, subsequent works do not build on this effort.

Two consultants (Björkman and Madestam) conducted, on behalf of the World Bank a PETS ‘pilot’ in

June 2003 covering the fiscal year 2002/03. It was a very narrow study in coverage but also in scope

as it focused only on the Primary Education Development Programme (PEDP) covering the capitation

grant, textbooks and the development grant. The study covered 15 primary schools in six districts in

only two regions – Kibaha and Bagamoyo (Pwani Region) and Masasi, Mtwara Urban, Mtwara Rural,

and Tandahimba (Mtwara Region).

This study attempted to relate resource flows to enrolment and test score data which previous

studies never did. Some findings of this pilot study contrasted with those of the previous two PETS,

especially on 'leakages' – it reported only a 5% leakage with 95% of transfers reaching schools.

: Page 20

REPOA was commissioned by the PER Working Group to undertake what was a fairly extensive PETS

in 2004 whose main objective was stated as ‘.. to establish more exact knowledge on the actual

amount of resources disbursed from central level that reaches the schools in PEDP'. It was based on

a sample of 210 schools across 21 districts, in 7 regions. The regions were sampled according to their

ranking on the Human Development Index while the councils were sampled according to their

proximity to the regional headquarter, and schools within districts were sampled according to

proximity to council headquarter.

The 2004 PETS found that the disbursement system from the central level to councils was complex

and consequently, tracking financial flows from the Accountant General down to frontline facility

levels was a complex task.

The study reported that for the fiscal years 2002 and 2003, inflows of overall capitation grant

reaching school level was in the range of 54% to 64% of the central level disbursement. For

development grants the recorded inflow of development grant at the school level was 84% of the

central level disbursement for 2002 and 2003. This survey also reported significant variations in the

inflow of capitation grant at the school level – varying from TSh 1,600 to TSh 8,700 per pupil per

year. In contrast to the previous studies the 2004 PETS makes reference to the previous two studies

(Björkman and Madestam, 2003 and REPOA and ESRF, 2001).

Like other previous studies the 2004 PETS was limited to some specific grants and covered only

primary education. It also focussed entirely on to what extent a grant intended for a specific use

actually reached the school while it was beyond the scope of the study to assess if the funds not

reaching the schools were actually also spent on primary education activities indirectly benefiting

the schools (training, supervision, investments in new schools, etc.) even if this would not be in

accordance with the intended use of the grant (earmarking not being effective). Unfortunately, the

amount assessed as not reaching schools was interpreted by some decision makers as 'leakage' i.e.

as if the resources not reaching schools were not used for expenditure benefitting schools and/or

other public expenditure.

There are a multitude of other PETS being undertaken at district level throughout the country by civil

society organisations. The public expenditure tracking work they undertake vary in terms of size,

sector, type of data tracked as well as methodologies employed. However, there is little substantive

information that is publicly available from this type of activity.

: Page 21

5 APPROACH AND METHODOLOGY

5.1 MAIN SOURCES OF INFORMATION

There are several sources of information which have been used to compile aggregated and

disaggregated information for primary and secondary education used in this PETS survey. This

information stems from national records and statistics presented for Tanzania mainland in total as

well as by regions, districts and schools. During the inception phase the following were identified as

sources and records which have been used both for sampling and for analysis when combining the

national records with data collected from the sample of regions, districts and schools;

1. Budget figures from MoFEA budget books, by sector, region and district votes/sub-votes. The accounting figures presented in budget documents are stated not to reflect accurate accounting information but only estimates of expenditures and in some cases no expenditure data are presented. Accordingly this source has only been used for analysing budget allocations, not actual expenditure.

2. Data from the regional budget department in MoFEA used for purposes of determining formula based and discretionary grant allocations to districts for different sectors.

3. Budget execution and accounting figures from the Financial Management Information System (FMIS)1 maintained by the Accountant General's office in MoFEA. These data show actual execution and consolidated accounting figures for central and regional levels. It includes some accounts for some sub-warrant holders such as secondary schools but not districts (and primary schools).

4. Data on council level revenues and expenditures by sector (not sub-sector) compiled by PMO-RALG based on annual statement of expenditures submitted by all councils at the close of the fiscal year. These data do not necessarily present final accounts since the data are collected from districts without any verification of the accuracy of the data. However, the data are submitted after they have been approved by the Council and as such represents official financial statements from the Council even if they are presented differently from the final accounts subject for audit. In many reports reviewed during the inception phase of this survey, expenditures on education as presented from accounts of the Councils have often been presented as primary education expenditure while they in fact also include expenditure on adult education, secondary education and general education administration at council level.

5. Aggregate data on primary schools from the MoEVT Education Management Information System (EMIS) which show aggregates of school data by district but not data for individual schools since these are contained by each district.

6. Data on each primary school from MoEVT prepared specifically for this PETS by collecting the detailed EMIS reports from councils included in our sample and compiling records per schools on the number of students and teachers per school2.

1 The system is frequently referred to as the EPICOR system since the system used in Tanzania is based on management software modules initially developed by Epicor Software Corporation.

2 Not all councils report number of teachers.

: Page 22

7. Data on secondary schools from MoEVT EMIS which show aggregates of school data by district but not data for individual schools.

8. MoEVT registry of all secondary schools in Tanzania which shows name, location, ownership and name of head teacher/contact person for each school.

9. MoEVT data on various transfers (recurrent and capital) made to each individual secondary school including the type and amount of transfer made.

10. Data on examination results for each primary and secondary school in Tanzania compiled by NECTA. This source of information has been used to compile a registry of primary schools. However, it does not include new schools where classes have not yet graduated or performed exams.

11. NBS data on districts and regions including population projections, area and classification of urban centres and rural communities.

12. NBS Household Budget Survey data to conduct comparative analysis of education spending by households with PETS survey data among others related to level of private contributions.

The above sources were used for initial analysis during the inception phase, for sampling of councils

and schools and merged with sample survey data for the sample of regions, councils and schools

selected. The data collected through the sample survey included education statistics and financial

information from the regional sub-treasury (transfers to sub-warrant holders/secondary schools),

transfers and revenue/expenditures of Regional Administrative Secretary (RAS), transfers and

revenue/expenditures of councils, and data on all receipts and expenditures by each school.

The above data were used to compile a database on primary and secondary education expenditures

and education performance indicators by regions, districts and schools. These data gave an

opportunity to make a stratified sample according to key variables, but also an opportunity to

compare sampled data with national data i.e. aggregate sampled data to national aggregates.

Figure 1 – Main source of information for Primary Education

The sources of information related to primary education is illustrated in figure 1 above.

Of particular relevance for sampling was the data on Council's performance related to Public

Financial Management (PFM) performance. This PETS study is about PFM i.e. how funds are being

PO-PMS

Teacher payroll

MoFEABudget books

volume I – IIIwith expenditure

estimates

Regional, District

and SchoolQuestionnaire

NBS

Population/area per districtHBS 2007

MoEFA Commissioner

Budget RegionsData on allocations

by District

Login TanzaniaData on receipts

and Expenditures by District

MoEVTEMIS Data on

Primary Education by District

NEC Data on

Primary schoolsPLSE

MoEFA EPICOR

Data on transfers by Region

Data base on Primary Educationby region, district

and school

PO-PMS

Teacher payroll

MoFEABudget books

volume I – IIIwith expenditure

estimates

Regional, District

and SchoolQuestionnaire

NBS

Population/area per districtHBS 2007

MoEFA Commissioner

Budget RegionsData on allocations

by District

Login TanzaniaData on receipts

and Expenditures by District

MoEVTEMIS Data on

Primary Education by District

NEC Data on

Primary schoolsPLSE

MoEFA EPICOR

Data on transfers by Region

Data base on Primary Educationby region, district

and school

: Page 23

managed and translated into service delivery. It is based on the assumption that challenges in PFM

impact transfer of funds and their application. In the Council database, data are presented along

several dimensions of PFM performance by Council3 and since funding for primary education is a

Council finance issue as a result of fiscal decentralisation, the data gave an opportunity to provide a

representative sample of districts according to PFM ratings.

The data from NECTA were used to compile a data base of primary schools. In total there were

13,948 primary schools reporting examination results to NECTA as compared to the total number of

15,673 primary schools registered in EMIS i.e. assuming the difference are schools which do not

have classes up to Standard VII, many which were built in recent years with their new classes yet to

reach Standard VII levels. There are differences between the NECTA data and EMIS data on PSLE

both at national and district levels, most likely due to timing of reporting from different districts and

schools with adjustments made between the two times of reporting, but they are minor (less than

1%).

As illustrated in figure 2 many of the same sources of information apply also to secondary education

although with some notable differences due to management of funding. While primary education to

a large extent is decentralised to councils, resources for secondary education was in 2008 to a large

extent under the responsibility of MoEVT4. The main differences in sources of information relate to

school level data, payroll information and grant allocations for schools.

Figure 2 - Sources of information for Secondary Education

MoEVT maintains records of all secondary schools and availed a secondary school registry showing

name and location of the schools. Furthermore, MoEVT availed information on transfers of grants to

each individual school which was executed by the Regional sub-treasury. Finally, MoEVT also

receives a full copy of the secondary school payroll from the central payroll maintained by the PO-

3 Councils are rated according to qualifications made in audit reports of Council accounts.

4 Councils also made contributions to secondary education in FY2008, first and foremost by

contributions to school and classroom construction and some operational costs.

MoEVT

SecondaryTeacher payroll

MoFEABudget books

volume I – IIIwith expenditure

estimates

Regional, District

and SchoolQuestionnaire

NBS

Population/area per districtHBS 2007

MoEFA Commissioner

Budget RegionsData on

allocations

by District

Login TanzaniaData on receipts

and Expenditures by District

MoEVTEMIS Data on

Secondary Education by District

MoEVTData on

Secondary Schools

MoEFA EPICOR

Data on transfers by Region

Data base on Secondary

Education

by region, district

and school

MoEVT

SecondaryTeacher payroll

MoFEABudget books

volume I – IIIwith expenditure

estimates

Regional, District

and SchoolQuestionnaire

NBS

Population/area per districtHBS 2007

MoEFA Commissioner

Budget RegionsData on

allocations

by District

Login TanzaniaData on receipts

and Expenditures by District

MoEVTEMIS Data on

Secondary Education by District

MoEVTData on

Secondary Schools

MoEFA EPICOR

Data on transfers by Region

Data base on Secondary

Education

by region, district

and school

: Page 24

PSM. In total these data enabled us to generate a database on secondary schools with a greater

extent of information for each school than for primary schools.

5.2 APPROACH

The basic approach has been to collect information from the sources above to build school level,

council level and regional level databases on financial and education sector performance

information. This information has been combined with data from the survey of the sample of

regions, councils, primary and secondary schools selected.

It allowed comparison of revenue and expenditure with budget allocations and to relate these

analyses to various regional, council and school level characteristics. Furthermore, it allowed more

in-depth analysis of correlations between resource efforts as measured by allocations and

expenditures, regional, council and school level characteristics, and the various education sector

performance indicators generated from the various sources.

The PETS was initiated with an Inception Phase to assess the availability of information required to

address the evaluation questions in the Terms of Reference. On the basis of this assessment and

preliminary analysis of data, a detailed approach, methodology and work plan were designed for the

PETS. The analysis of national, regional and council level data was used to select an initial sample of

councils and schools within the councils to be subject for additional data collection using survey

tools (questionnaires). The councils and schools were sampled from seven regions in mainland

Tanzania (one third of the regions) based on the proposed methodology. The final sampling is

presented in sections below.

A visit to one district and two schools during the inception phase gave some insight into what

information could be made available at regional, council and school levels. With this information a

first draft of the main survey tools, four questionnaires, was developed; (i) one for regional sub-

treasuries and Regional Administrative Secretary (RAS), (ii) one for councils, and one each for (iii)

primary and (iv) secondary schools.

The outcome of the inception phase was presented in an Inception Report5 for consideration by the

PETS subcommittee, the main reference group for the PETS consultants overseeing the

implementation of the PETS. The report presented a proposed methodology and approach and

included presentation of initial analysis used for a preliminary sample of councils and schools.

Annexed to the report were a presentation of the various national level data sources mentioned

above as well as a first draft of the main survey tools; the questionnaires. After reaching an

agreement on the approach and methodology with some adjustments made to the proposal in the

Inception Report, the survey was implemented through four additional phases;

1. A testing and verification phase.

2. A data collection phase.

5 Public Expenditure Tracking Survey for Primary and Secondary Education in Tanzania. Inception

Report, J. Claussen and M. Assad, 28 May 2009.

: Page 25

3. A data compilation and analysis phase.

4. A reporting and presentation phase.

5.2.1 TESTING AND VERIFICATION PHASE

The initial draft survey tools were tested by the consultants in two regions6 to see if survey tools

were reflecting accurately what type of data could be made available and how the information is

recorded and presented.

The testing phase proved invaluable for ensuring that the tools accommodated the differences

between regions, councils and schools concerning what sources of revenue finances primary and

secondary education and differences in allocating resources to schools. Furthermore, it showed that

information on resource flows to councils and schools is maintained at different administrative

levels. Since the execution of payments for goods and services benefitting schools are made at

different levels, this information needed to be captured at these levels for each school sampled and

compared with records at the schools on in kind contributions received (such as teacher salaries,

desks, chairs, teaching materials, etc.).

Both in the case of primary and secondary schools it also showed that Central Government transfers

to primary and secondary education is just one source among several sources of funding. In many

cases both RAS and councils supplement Central Government earmarked transfers to schools or

allocations paying for school construction and maintenance. Some councils also contribute to

secondary schools despite that only primary schools were under their authority in 2008 while

secondary schools were under the responsibility of MoEVT. In some cases it was observed that RAS

transferred money to councils to supplement their funding of both primary and secondary schools,

in other cases some RAS transferred money directly to schools or procured goods and services for

the benefit of individual schools. Thus in some regions and councils, and for some schools,

discretionary funds at regional and district levels have supplemented central level grants.

To accommodate these variations in sources of funding and types of transfers many of the same

questions were raised in the different questionnaires to Regional Sub-Treasury, RAS, councils and

schools i.e. to identify who funded a particular input accounted for as primary education, who

actually made the payment and who was the final beneficiary of the input paid for.

Analysis of school records and education sector statistics at central level showed some

inconsistencies when attempts were made to consolidate this information. Furthermore, some

records were incomplete or maintained at a level of aggregation that limited the ability to analyse

variations between schools (especially in the case of primary schools). Accordingly, the survey tools

were also designed to resolve these inconsistencies.

The questionnaires also included questions related to data already contained in national records,

both as concerns councils and schools (like number of schools, number of teachers, number of

6 The testing of the survey tools was conducted in Kilimanjaro and Pwani (Coastal) regions and included testing of the tools with regional sub-treasuries, Regional Administrative Secretary, district education offices and district treasuries of four councils and a total of eight primary and four secondary schools in the councils visited.

: Page 26

students, number of students passing exams, etc.). This was done to obtain detailed school level

information but also as a quality assurance measure related to data contained in national records.

Several of the questions were also used to identify to what extent resource use like teacher salaries,

per student grants, etc. is based on the actual number of teachers performing in schools and

students attending classes.

The questionnaires also contained information to trace efficiency in actual resource flows. It

included data on time from allocation to actual receipt by beneficiary. It also included data on

amount of transfers recorded by national and local government levels to be compared with data

recorded as received by beneficiary as per school cash books and bank statements. This was done to

assess to what extent the amount allocated and transferred from national and local government

levels was actually received by the intended beneficiary.

The testing process resulted in five major revisions of the questionnaires, among others with the

professional input of MoFEA (who executes the central level transfers), PO-PSM (who maintains the

central payroll), PMO-RALG (who monitors council level transfers and budget execution) and MoEVT

(who has detailed knowledge of how the education system is managed and funded). It is not least

due to the input from these government agencies that the final questionnaire proved to be

reflecting the actual resource flows and management arrangements at the different levels of the

Tanzanian education system and which made the enumeration exercise successful as concerns

quality and accuracy of data.

5.2.2 DATA COLLECTION PHASE

The data collection phase commenced after MoEVT had contracted a firm to mobilise and manage

enumerators. The firm contracted mobilised approximately 50 remunerators and a two day training

workshop was conducted by the PETS consultants on the process of collecting data and information

from statistics and accounts available at regional, council and school levels as well as interview

techniques and processes of completing the questionnaires. The management of the firm had

significant prior experience in designing and conducting similar surveys which proved to be valuable

in facilitating the enumeration exercise.

The main feature of this PETS was to use data evidenced by and checked with available

documentation. It means that data were not only obtained in an interview but also by verifying data

as presented in school cash books and ledgers, school attendance records (schools and districts),

school statistics (districts and regional education offices) and budget execution reports (regional

treasuries). All this documentation was subsequently consolidated with the data from national

records obtained during the inception phase.

The workshop presented samples of the above mentioned documentation as part of the training.

During the workshop, the data entry format was also introduced with a presentation of each of the

four questionnaires.

The workshop was also used to organise the enumerators into seven survey teams, one for each

region with sub-teams for the 27 councils sampled. The teams varied in size pending the number of

councils and schools selected in each region as a result of proportional sampling of councils and

schools (ref. section on sampling below). In total 35 enumerators participated in the data collection

: Page 27

process under the supervision of seven regional survey team managers/coordinators and one overall

manager7.

The data collection phase started in June 2009 and continued through July 2009. During the time

most of the primary schools had closed for mid-term break. Nevertheless, all except a few of the

schools were visited8 with head teacher and other staff present with the assistance of the District

Education Officer (DEO). However, the DEO did not participate in the actual data collection process.

The few schools that could not be visited were substituted with others by a substitution procedure

defined prior to data collection and to ensure that the sampling criteria were met.

The final task for the enumeration firm was the entry of data into the prescribed format provided by

the PETS consultants. The data was submitted to the PETS consultants mid August 2009 which then

initiated a process of data compilation and several test runs to check consistency of data collected.

5.2.3 DATA COMPILATION AND ANALYSIS

Four data bases were generated; one with regional data, one with district data and one each with

primary and secondary school data. These databases were generated by consolidating data from

national records with data from the questionnaires. It meant reconciling and merging regional,

district and school level records with the aggregate regional, council and school level records already

generated during the inception phase i.e. sample data were merged and added to the databases

covering all regions, councils and schools of mainland Tanzania9.

The data compilation task proved to be time consuming since all the different national records are

entered in very different systems, some which could export data in ASCII or Excel formats, others in

PDF formats and some which had to be re-entered manually. It demonstrated that there is a scope

to harmonise these systems in order to consolidate information, to improve sector monitoring and

for improving quality assurance related to national records and statistics. As an illustration, in many

of the national records containing school level data, the schools are only identified by name, not a

unique school registration code. Comparing names however, proved to be challenging and time

consuming since they had been entered differently in each file and sometimes with spelling errors,

sometimes a school name was the same for two schools and then its regional or council location was

the only key to differentiate them.

A number of test runs were conducted for regional, council and school level data to assess

consistency and accuracy of data collected from the different sources. The result of these test runs

required some additional consultations with the enumeration firm as well as with national

institutions to verify and/or adjust data entries.

7 A more detailed presentation of the data collection process is provided in PETS - Enumerators Synthesis Report, August 2009 by Prime Consult International Ltd.

8 Some few substitutions were made due to unavailability of head teachers as described in the report by the enumeration firm.

9 National records for primary schools were only available for the seven regions selected since there is no updated central registry of primary schools.

: Page 28

Following the above, a number of frequency tables was generated (descriptive statistics) which have

was used for analysis to respond to the main issues presented in the terms of reference. The results

of these analyses were presented in a first draft report.

5.2.4 REPORTING AND PRESENTATION PHASE

A first draft report was submitted in September 2009. 10 The report presented the findings from

analysis of a first run of the data collected. At this stage the analysis was presented mainly in the

form of descriptive statistics. It was presented at a stakeholders meeting organised by MoEVT and at

the joint annual education sector review meeting. The feedback and comments from these venues

were included in the work of the consultants and was reflected in this second output from the

reporting and presentation phase; a full Draft Final Report intended to respond to all issues agreed

to during the Inception Phase.

The Government and DPs presented separate comments to the Draft Final Report. Many of the

comments was observations related to findings and some requested more analysis, and in particular

presentation of more elaborate conclusions and recommendations. Some of the comments revealed

a need to clarify presentation of analysis and conclusions which have been done by revising relevant

sections. For this Final Report many of the comments have been incorporated while some have been

addressed in a cover letter in submission of the report since they called for analysis beyond the

scope agreed in the inception phase.

5.3 SAMPLING

There were several sources of information available to get an initial overview of the structure of the

education sector for mainland Tanzania. The NBS regional and district population census data, the

MoFEA budget and accounting data including payroll for civil servants, the LGA expenditure and

financial performance statistics, the PO-PSM payroll data, the NECTA registry of PSLE results and the

MoEVT EMIS database (BEST) of primary and secondary education by regions and districts were all

consolidated into four databases; regional, district, secondary and primary school databases.

During the inception phase some initial analyses were conducted using these databases to test some

hypothesis on school performance related to various characteristics of regions and councils such as

degree of urbanisation, size of schools, student/teacher ratios, regional and district poverty

indicators, and council level financial management performance according to result of audits by the

Controller and Auditor General (CAG).

There were two sets of samples to be selected; a sample of primary schools and a sample of

secondary schools. When selecting samples the key issue was to achieve a representative sample of

schools related to the main evaluation questions of the PETS; Do public resources intended for

schools reach down to school level and are they translated into service delivery i.e. applied for their

intended use?

10 Public Expenditure Tracking Survey for Primary and Secondary Education in Tanzania - First Draft Report, 21 September 2009, by Jens Claussen and Mussa J. Assad

: Page 29

The Terms of Reference suggested a sample of 10 regions, 3 districts in each region and a total

sample of 300 primary and 150 secondary schools within these regional and district strata. The

selection of regions was to represent a balanced selection related to regional poverty indicators.

Selection of districts and schools should take into account a fair representation in urban versus rural

schools. The resulting key indicators for this sampling method are illustrated in table 1 below.

Table 1 – Sampling according to the Terms of Reference11

Total Sample % of total % of schools in

sampled councils

Regions 21 10 48

Districts 133 30 23

Primary schools 15257 300 2 9

Secondary schools 2520 150 6 26

Results from the initial analysis conducted during the inception phase suggested that the above

sampling method would not accurately respond to the main evaluation questions, in particular

related to primary education. As presented in sections below, this is because regional location is

neither the main determinant for poverty levels nor for school level performance and allocation.

For primary schools the main determinant is various district level characteristics, and they vary

significantly within a region (with Dar es Salaam as an exception) i.e. district level performance is not

depending on regional location but its location within a region (degree of urbanisation as measured

by distance to the regional capital). Ideally, for primary schools, the sample of schools should have

been selected from districts regardless of regional location. As explained later however, due to

limitations in time frame and resources it was necessary to limit the geographical outreach of

districts in the survey for logistical reasons i.e. select districts in a limited number of regions.

In terms of secondary schools it is also the proximity to an urban centre that appears to be the main

determinant for some key characteristics related to performance and resource allocation.

Accordingly, the approach to sampling was decided as a three step approach;

1. The first step was selection of regions due to management and administrative

considerations. These regions were decided upon by the PETS subcommittee.

2. The second step was to select districts within these regions. The task for the PETS

consultants was to develop a method for selection of districts which ensured that they gave

a fair representation of all districts in mainland Tanzania.

3. Finally, within these districts primary and secondary schools were selected based on various

criteria as described below.

The survey was to include public schools i.e. not private schools. Approximately 97.3% of all primary

schools in 2008 were government schools12. In the case of secondary schools it included community

schools (established and managed by the community) since they are considered as government

11 Data on number of schools in 2008

12 MoEVT EMIS database

: Page 30

schools (approximately 80.0% of all secondary schools in 2008 were government and community

schools). Since the survey focus on 2008 school year performance and using fiscal year 2007/2008

for analysing resource flows to schools, the school records from which the sample was selected only

included schools that had been established before July 2007 (i.e. in full operation during the fiscal

year 2007/08).

For secondary schools MoEVT maintains a fairly updated registry of schools which also shows

ownership. While it does not contain the year in which the school was established, when comparing

registry records from 2006, 2007 and 2008 with MoEVT data on secondary school transfers and

NECTA data on exams we were able to identify the approximate year and month the school was

established.

Since the survey was to compare also school performance data like examination results, only schools

that had been recorded with such results by NECTA was included i.e. some schools may have been

established in 2006 and 2007, but did not yet have any classes in form IV, and these school were

accordingly not included in the sample. According to NECTA there were 2,647 secondary schools

with 233,848 students sitting for O-level exams of which 1,926 were government and community

schools. It is among the latter the sample of secondary schools were selected although substitutions

had to be made during data collection since in a few cases the MoEVT registry had recorded a school

as government or community owned while it in fact was private.

The same approach was applied for identifying primary schools that would be used for the sampling.

In this case however, the only detailed consolidated record of primary schools is to be found in the

NECTA records of examination results. According to the MoEVT EMIS database there were 15,673

primary schools in 2008 with 1,065,819 students ending Std. VII. Of these 15,257 were government

primary schools with 1,055,219 students ending Std. VII. According to NECTA there were 13,948

schools with 1,017,865 students sitting for the PSLE. The lower number of NECTA schools and

students may be due to errors and omissions in data collection and compilation by NECTA and/or

MoEVT. The sample of the selection of schools was made from the NECTA database of schools.

5.3.1 TYPE OF SAMPLING

The number of primary schools in a district varies from 14 to more than 314 according to EMIS data

and for secondary schools from as few as 3 in one district to as many as 103 (in one of the Dar es

Salaam districts).

Following the inception phase it was decided to use proportional sampling to maintain a reasonable

sample of schools for all districts, i.e. in districts with many schools a larger number was selected for

the sample and/or in districts with lower number of schools fewer schools were selected for the

sample. This was however adjusted with relatively higher number of schools selected in districts

with a small number of schools and lower number of schools in districts with high number of

schools, this to maintain statistical significance for the sample selected in a district.

5.3.2 SELECTION CRITERIA

The scope of the survey was first and foremost to analyse flow of funds to its intended use. As

observed by various previous studies of PFM issues including the PETS 2004 survey, the main

: Page 31

variation in resource allocation and use for primary schools is to be found at council level i.e. the

major differences are between councils' approach to funding/support to primary schools.

This was confirmed by the analysis of average allocation per student by council from the preliminary

databases generated during this inception phase (as illustrated by figure 3 below). Variations

between districts within a region appear often to be more significant than variations in regional

averages i.e. using regional averages hides the fact that education sector performance and

characteristics varies significantly within a region.

Figure 3 – Average expenditure per primary student per council 2007/2008 (in TSh)

This observation was also supported by the recent NBS HBS which suggests that classifications of a

region is not a main determinant of the frequency of poverty but level of urbanisation within the

region is. Urban councils have the lowest level of poverty while peri-urban areas has the highest

level of poverty and lowest level of public services per capita i.e. peri-urban areas of Dar es Salaam

and Mbeya have a higher share of poor households than urban councils of 'remote regions' such as

Kaghera and Singida with lower population densities.

Accordingly, the key was to use criteria for selecting council level strata (rather than regional strata)

and with a reasonable sample of primary schools in the district. Ideally then for primary schools the

selection of councils should have been made among all councils regardless of which region they are

located in. For management purposes however, it was decided to select seven regions (including Dar

es Salaam) but ensure that these regions would include councils that reflected the significant

variation between districts in mainland Tanzania.

For primary education, public funds are managed by councils. Initial analysis suggested that there is

a correlation between PFM performance criteria as measured by different types of audit

qualifications from council audit reports and level of allocation to different service delivery units and

sectors at the council level.

Accordingly, one stratification criterion was to ensure a fair representation of councils according to

PFM performance criteria. The PFM rating proved however to be correlated to the rural/urban

dimension, i.e. urban centres which have higher levels of public revenue per capita have better PFM

ratings than rural districts with lower public revenue resources per capita. This observation can be

-

20 000

40 000

60 000

80 000

100 000

120 000

140 000

160 000

180 000

Regions/Districts

Exp

end

itu

rep

er s

tud

ent

: Page 32

explained by the fact that urban centres have higher level of administrative capacity and enjoy

economy of scale in management of public funds. Subsequently, the stratification criterion used was

proximity to major urban centres. This was used for selecting councils within a region and primary

schools within a council.

For secondary schools the main fund manager is the school itself and transfers are made from

regional sub-treasuries although MoEVT also makes direct procurements on behalf of secondary

schools. It suggested that a representative number of secondary schools were to be selected as a

representative sample in the region since in this case selection of councils was less of an issue.

The rural/urban dimension is closely correlated to allocation per student. Allocation per student is in

turn correlated to the pupil/teacher ratio (P/T) given the fact that teacher salaries are the main cost

component for schools. From several PETS and other PFM studies related to education, it has also

been observed that large schools tend to receive higher non-wage allocations per student than

smaller schools.

Other non-wage allocations are also correlated to the rural/urban dimension similar to the PFM

performance rating mentioned above i.e. urban schools tend to be larger, have lower P/T ratios per

student and thus higher total allocations per student as well as located in councils with higher PFM

ratings than rural schools. Accordingly, the combination of location and size was used for

stratification by first selecting councils by location and secondly selecting schools to achieve a

representative sample in terms of size.

5.3.3 SIZE OF SAMPLE

This PETS survey has included all recurrent and capital allocations as well as private contributions to

primary as well as secondary schools. The study has allowed an in-depth understanding of

differences in transactional practices between councils and school levels for primary schools and

maintains a reasonable sample at regional level for secondary schools.

The sample included all districts in three regions to allow full consolidation of resource flows from

central levels to councils. For the other four regions with more than five districts 50% of the districts

was selected.

Selection of schools was made as a proportional sample. For primary schools a sample of 10% was

selected although in districts with less than 50 schools, the sample was increased to 20% and for

districts with more than 200 schools reduced 5% to maintain acceptable statistical significance

within the resource and time frame allocated to this PETS.

For secondary schools a similar approach was taken, however in this case the major issue was to

select a representative sample along the regional urban/rural dimension since secondary schools

receive public funds predominantly from the MoEVT budget transferred through regional sub-

treasuries. The key sampling criterion was a fair representation of urban versus rural schools.

The above lead to an adjusted sample as illustrated in the table 2 compared to the sample suggested

in the Terms of Reference (ref. table 1).

: Page 33

Table 2– Adjusted sample

Total13 Sample % of total % of schools in

sampled councils

Regions 21 7 33

Districts 133 27 20

Primary schools 13,948 283 2 10

Secondary schools 1,926 75 3 12

5.3.4 SELECTION OF REGIONS AND DISTRICTS

The regional sample selected by the PETS subcommittee were the following:

1. Dar es Salaam - (Unique region)

2. Arusha - Northern Zone

3. Kagera - Lake Zone

4. Mbeya - Southern Highland Zone

5. Lindi - Southern Zone

6. Kigoma - Western Zone

7. Singida - Central Zone

The PETS consultants were charged with the task to ensure that selection of districts and schools

within these regions was representative of the total population of districts and schools in mainland

Tanzania.

In the fiscal year 2007/2008 there were 134 district votes of which Dar es Salaam City Council

featured as a separate vote. However, education grants were allocated directly to the three district

councils within Dar es Salaam and not the City Council.

Some new councils were established during 2007; Mpanda and Njombe Town Councils as well as

Arusha Rural, Bahi, Chamwino, Chato, Longido, Meru, Misenyi, Mkinga, Nanyumbu, Rorya and Siha

District Councils which during 2007/08 was separated from their parent councils and received

separate allocations for all transfers from July 2007. Arumeru and Dodoma District Councils were

fully absorbed into the new councils. This required a special effort in reconciling school year data

(which follows the calendar year) with financial data (which are according to the fiscal year) for the

new councils established.

The net increase in councils from FY2007 to FY2008 was 11. With this change in councils the total

number of councils in mainland Tanzania receiving education grants was in FY2008 133.

The criteria for selection of councils were;

13 The number of primary and secondary schools in 2008 depends on which source of information is

used. The MoEVT EMIS statistics (BEST) give a different figure for government and community

schools than the registry of secondary schools provided by MoEVT and again a different number

than the schools according to the payroll data for teachers and the NECTA records of schools.

: Page 34

All three councils of Dar es Salaam were selected.

For regions with less than six councils, all councils were selected to allow consolidation of

data from councils to regional levels.

In regions with more than five councils, 50% of the councils was selected.

The following table show the councils selected according to the above mentioned sampling

methodology.

Table 3 - Councils selected for the sample

Vote ID Council name Official name Region Type

702001 Arusha Urban Arusha Arusha Town Council

703084 Karatu Karatu Arusha District Council

703006 Monduli Monduli Arusha District Council

703007 Ngorongoro Ngorongoro Arusha District Council

882019 Ilala Ilala Dar es Salaam Municipal Council

882020 Kinondoni Kinondoni Dar es Salaam Municipal Council

882021 Temeke Temeke Dar es Salaam Municipal Council

872002 Bukoba Urban Bukoba Kagera Town Council

873074 Karagwe Karagwe Kagera District Council

873076 Muleba Muleba Kagera District Council

873078 Ngara Ngara Kagera District Council

743022 Kasulu Kasulu Kigoma District Council

743023 Kibondo Kibondo Kigoma District Council

743021 Kigoma Rural Kigoma Kigoma District Council

742005 Kigoma Urban Kigoma TC Kigoma Town Council

763030 Kilwa Kilwa Lindi District Council

763032 Lindi Rural Lindi Lindi District Council

762006 Lindi Urban Lindi Lindi Town Council

763029 Nachingwea Nachingwea Lindi District Council

783037 Chunya Chunya Mbeya District Council

783038 Ileje Ileje Mbeya District Council

782007 Mbeya Urban Mbeya Mbeya Town Council

783042 Rungwe Rungwe Mbeya District Council

843063 Iramba Iramba Singida District Council

843064 Manyoni Manyoni Singida District Council

843062 Singida Rural Singida Singida District Council

842014 Singida Urban Singida Singida Town Council

5.3.5 SELECTION OF PRIMARY SCHOOLS

The resulting proportional sample of primary schools per district according to the procedure

described in the sections above are provided in annex II. The selection of schools was made from the

database on PSLE results provided by NECTA since there is no updated national registry of primary

schools in Tanzania. As previously mentioned the NECTA database does not include schools with less

than Standard VII classes, i.e. it includes 89% of all primary schools when compared to the total

number of primary schools in 2008 as per MoEVT EMIS.

: Page 35

The sample was selected from the NECTA data to ensure that only schools which were fully

operational during the period subject for the survey was included. However, it proved to include also

some private/NGO schools. In these cases they were be substituted with another school from the

total primary school database with a school which met the same selection criteria for a school in the

council. The substitution was done according to a predefined selection procedure.

The sampling of schools was based on a normalised distribution of schools within the councils by

subdividing them into three clusters of small, medium and large schools as measured by number of

students. Since there is no central registry of schools that shows number of teacher and students per

school this was done using PSLE data for primary school exam leavers (assuming a correlation

between number of Standard VII students passing PSLE and size of schools).

13 schools were substituted during the data collection phase (4.6% of the total sample). In six cases

this was because the school proved to be a private school. In four cases the schools were visited but

the head teachers were not available and accordingly records could not be examined. Another three

schools proved to be located at very remote places (Kigoma region) and with the timeframe and

budget allocated out of reach for the enumerators.

5.3.6 SELECTION OF SECONDARY SCHOOLS

In 2008 secondary schools received direct funding from MoEVT and accordingly the impact of council

variables was assumed to be of less significance for secondary schools as it would be for primary

schools. When selecting secondary schools a proportional sample was used with 10% of all schools in

the regions selected but adjusted upwards for urban centres due to a low number of schools (a total

of between 5 and 29 in town/municipal councils). The resulting sample of secondary schools is

presented in annex III.

The database for sampling was provided by MoEVT from the input to the MoEVT EMIS. It does not

clearly distinguish between ownership. Reconciling the data with data on school grants from MoEVT

also indicate differences in number of secondary schools. The school registry of secondary schools

initially received showed name, location, ownership but deviated from both of the above mentioned

records. Most likely this is because they are using different sources, are maintained by different

administrative units and are updated at different intervals. It however created a risk that some of

the schools selected were private secondary schools, or, that some secondary schools actually

established were not included in the records used for sampling. The differences in the number of

secondary schools in these data files were approximately +/- 300 schools which was less than 10% of

all secondary schools in 2008.

The total number of secondary schools sampled was 75 which is less than the initial sample agreed

to during the inception phase. This is because the decisions by the PETS subcommittee to change the

regions after the close of the Inception Phase had impact on the total population of secondary

schools in the regions selected for sampling (the new regions selected had fewer secondary schools

than the regions initially selected). The reduction in the sample, however, did not have impact on

the credibility of the analysis since the same sampling method was used.

As in the case of primary schools, if the sample included a private/NGO school it was to be

substituted with another school from the database of all secondary schools. The substitution was

: Page 36

done according to a predefined selection procedure. The actual substitution was made for only two

schools (2.6% of the total sample). One of the schools had opened after June 2007 although

according to records used for sampling it suggested it had been opened before. In the other case

the school had no records and accordingly no data could be collected14.

5.4 SOME ISSUES RELATED TO DATA AND DATA QUALITY

In undertaking this PETS a number of observations have been during the process of collecting and

compiling national records and in reconciling records from different sources. In the following we will

focus in particular on the MoEVT Education Management Information System (EMIS) which serves as

the basis for monitoring education sector performance and also as input for the MoFEA grant

allocation formulas for allocations to councils (and in 2008 MoEVT allocations to secondary schools).

The first issue that needed to be addressed in our survey was how to get an accurate and credible

account of how many primary and secondary schools were physically established in mainland

Tanzania in 2007 and 2008, who the owners were and where the schools were located. The second

issue was to establish how many teachers and students were attached to these schools in 2008. The

third to establish how much resources they were allocated in the fiscal year 2007/2008.

For the purpose of this study all the above were of importance to be able to make a stratified

sample; firstly because we were provided a sample of regions from which to make a representative

sample for mainland Tanzania. It required information on the distribution of schools and education

resources in sampled regions compared to mainland Tanzania in total. Secondly, it was required to

be able to sample councils within these regions, again on the basis of council and school level

features that could then ensure that the sample was reflecting a fair selection of councils and

schools of mainland Tanzania, and if not, calculate coefficients to adjust results to represent

mainland averages.

In reconciling information from different sources we found deviations and inconsistencies in the

data. These deviations and inconsistencies became even more visible when collecting data from the

sample of regions, councils and schools using the survey tool. The data collected through sampling

allowed us to make adjustments to national data for the purpose of these analyses.

The issue of credible data is not only an issue of importance for monitoring of education sector

developments but also to have a full and accurate account of government assets. It is also of

importance because national education records and statistics are used in decisions on allocation of

resources.

While the errors and inconsistencies were minor compared to the total and accounted for only 2-3%

of the total number of records for primary and secondary schools, they have significant impact for

individual councils and schools in allocation of resources; some schools received far less than their

prorate share of resources (when applying formula based allocations), some did not feature in the

records even if they were eligible for some allocations and some received allocations they were not

14 The head teacher had passed away and the successor could not find any records.

: Page 37

entitled to due to ownership or more than their entitlement because teacher and/or student records

in the data being used for decision making exceed what is the reality in the school.

In the following we will address some of these issues in more detail for primary and secondary

schools. Based on our observations we also try to make some recommendations for further

improvement of data quality.

5.4.1 SECONDARY SCHOOL DATA

In terms of secondary schools, there is a database maintained by MoEVT showing name, address and

ownership status of secondary schools (a 'registry of secondary schools'). During the process of

implementing this survey some observations were made related to these and other records used for

generating statistics on secondary schools;

The school registration code featuring in the secondary school registry was not available in

most of the other databases on secondary schools, i.e. reconciling data sets from one source

to another proved to be a challenging task, especially since the number of schools in one

year varied from one source to another.

The actual year the school was established was not available in most of the school records

used for the purpose of this survey. Comparing different versions of the school files for

secondary schools revealed that the number of secondary schools in a school year reported

by EMIS differed from the number of schools registered in that year according to the school

registry file maintained by MoEVT and again from the records on transfers to schools .

The school registration number in the NECTA database on examination results is different

from the MoEVT registration code. In the NECTA database school ownership was not shown.

Examination results according to NECTA differ from those presented in the EMIS statistics.

In other words the only unique identity of a school for all these datasets was the school name. Since

school names differ for some schools between databases due to different spelling when entering the

data, and some schools have the same names, additional information such as region or council

location is required to identify the school. However, in some cases the same school appeared with

different locations in different databases.

Some schools appear as a government school in some databases while in others the same schools

appears as a community or NGO school. Errors in the classification of ownership have led to

government allocations to private schools of grants eligible only to government/community schools.

All the above suggest that to improve quality assurance and credibility of data, all databases should

use the same school registration code to allow full reconciliation and quality control . The following

may be considered for further improvement of data quality and monitoring.

The school registry should be updated regularly and be used for reconciling information with

all other sources i.e. the point of departure for all databases should be the registry and they

should all apply the same unique school registration code. The registry should include the

year (and date) the school was actually established.

The TSS forms used for collecting data from secondary schools should be checked and

reconciled with the registry to ensure that data for all schools are collected accurately.

: Page 38

Our sample survey revealed deviations between the EMIS database, data in TSS forms and

school information collected through our survey tools from school records. As an approach

to quality control, MoEVT should also consider implementation of sample data audits of

schools to verify data as they are presented in TSS forms. The task of verification of school

level data could be delegated to the Regional Education Officers (REO) and/or DEOs.

Payroll data from PO-PSM records, MoFEA and MoEVT contained information on teachers

allocated by pay station. A pay station was in most cases one school making payments to

teachers at several schools i.e. the data on payroll did not identify at which school the

teacher was actually serving. The latter is an important element to be able to reconcile TSS

data on teachers at a school with teachers actually on the payroll. If the payroll included not

only information such as function, salary level, date of employment and pay station, but also

at which government institution (school) the person is serving it will significantly improve

monitoring of teacher attendance when consolidating payroll data with data from TSS forms.

Getting official records and accounting data on transfers and application of grants to schools proved

also to be a challenge;

The regional sub-treasury data contain records of financial transactions at the school levels

by expenditure category. This would have been the most reliable basis for capture of

regional financial data. However, for FY2008 it proved to be incomplete in that not many

schools had been captured in the FMIS database and in a format required for analytical

purposes.

The 'Deposit Account' could not generate Expenditure Analysis Reports by Sub-Warrant

Holder and therefore no analysis of official records by schools proved to be possible i.e. not

all sub-treasuries kept accounts per secondary school of transfers to and payments made for

schools.

The 'Recurrent Account' was able to generate the Expenditure Analysis Report by Sub-

Warrant Holder but again only for 5 of the 75 schools in our sample. This small sample was a

weak basis to undertake any rigorous analysis. Even within this small sample it is evident

that regional level financial data records contain only a portion of expenditure incurred by

schools15.

Financial data are not collected, reported and analysed in such a way to allow the focus to

be at the school level for both wage and non-wage expenditures. These two aspects will

require attention if reporting and accountability for resources is to be improved and tracked

to individual schools. Neither the regional sub-treasuries nor the schools have data on

personnel emoluments by school.

However, with secondary school allocations being devolved to councils, the specific observations

made on the issues related to primary also becomes more relevant in terms of secondary.

15 As an example; in 2008 there were 114 government and community schools in Arusha region but the MoFEA IFMS system could only produce detailed expenditure returns for approximately 30 secondary schools (apy-stations). When comparing the amount of expenditure in our sample by different categories with the same amounts for those few schools of the sample in the MoFEA IFMS reports showed only a fraction of the amounts recorded as spent by the school.

: Page 39

5.4.2 PRIMARY SCHOOL DATA

Many of the same issues related to secondary school data also apply to primary school data.

However, in this case it was an additional challenge; there was no national registry of primary

schools. The school registration code, name, location etc. was only available at council levels. The

councils consolidate the TSM forms from schools and submits aggregate data for the councils to

MoEVT. The only consolidated records of primary schools for mainland Tanzania are maintained by

NECTA but this database only contains records of schools with students sitting for PSLE exams.

Schools with classes yet to reach this level were not included, nor did the data distinguish between

government and private schools.

According to the NECTA database there were 13,948 primary schools in mainland Tanzania in 2008.

According to the EMIS data for 2008 it was 15,673 primary schools, i.e. this assumes that 1,725

primary schools were not performing exams (because they were established in recent years and no

classes yet reached standard VII). However part of the deviation may be due to a statistical error in

data collection and compilation.

According to the NECTA database there were 1,017,865 students sitting for PSLE exams in 2008

while there were 1,065,819 students enrolled in Form VII in 2008 according to the EMIS data i.e. it

may be interpreted as a dropout of form VII of 47,954 students (4.4%) during the school year and/or

a statistical error due to data collection.

If there was a national registry of primary schools both the above data sources could be reconciled

with this registry to get assurance that data had been collected from all schools. While this may

prove to be a challenge considering the number of primary schools, such a registry is not a major

task to update once established. If all data sets and records include a unique school registration

code, reconciling data becomes a simple data reconciliation process.

As with secondary schools, the payroll data only contained information on persons paid from a pay

station, in the case of primary school teachers, the councils. There were approximately 155,000

primary school teachers according to EMIS data in 2008. To maintain accurate centralised records of

at which school these teachers are employed will be a major task since a significant number of

teachers change schools (and councils) within the year. Information on changes in positions within a

council is information maintained by the council while changes between councils is registered in the

central payroll records.

For quality assurance of the EMIS record the same approach in conducting sample audits may be

applied to ensure that records on number of teachers are accurate. Furthermore, the records should

distinguish between teachers on government payroll and teachers paid by other sources. As

evidence by the survey data, primary schools (like secondary schools) employ teachers paid by

contributions from, among others, parents.

MoEVT is currently implementing a project to change its data collection and consolidation process.

By decentralising computerised data entry to councils a complete data register of primary schools

will be available at MoEVT with details of each primary (and secondary) school. However, the project

appears to be based on the assumption that all 133 councils will be able to maintain a functioning

computer for data registration.

: Page 40

A similar exercise to decentralise EPICOR (the Financial Management Information System) to all

councils has not yet succeeded. According to statistics from the MoFEA Accountant General's Office

more than 80 councils had a computerised EPICOR module installed. However, during the survey it

was revealed that several of the councils assumed to have a computerised EPICOR system installed

did not used it due to communication problems, unreliable supply of electricity and/or staff

turnovers. Thus the decision to computerise school recording and monitoring systems at council

level may need to be reconsidered. It may require, in an intermediary stage, to actually reconcile

data at the central level rather than council level if the quality and credibility of the data are to be

improved.

5.5 ORGANISATION OF THE SURVEY

The implementation of the assignment was made under the supervision and guidance of a

subcommittee chaired by MoEVT and with participation of other stakeholders.

The assignment was implemented by two international consultants serving as team leaders namely

Jens Claussen, Nordic Consulting Group and Dr. Mussa J. Assad, University of Dar es Salaam

contracted by the Department for International Development (DfID). These two consultants had the

overall management responsibility for the implementation of this PETS.

The third and important organisational element for implementation of the assignment was a firm

contracted by MoEVT, Prime Consult International Ltd. (PMI),Tanzania, who mobilised and managed

a total of 35 enumerators supervised by seven survey team leaders under the overall management

of a 'survey manager'.

: Page 41

Figure 4 – Organisation

In addition to MoEVT there have been several other institutions that also have provided important

inputs to the survey (ref. section on sources of information). These have first and foremost been

MoFEA including the Accountant General and budget departments for budget and accounting data,

PO-PSM for additional payroll data, PMO-RALG for council data, NBS for household and population

data and the sample of regional treasuries and councils selected for survey.

Figure 4 illustrates the organisational arrangement and line of authority.

Since the assignment was not contracted in full to external service providers, it required a process of

close consultations between the subcommittee and team leaders for each step and phase of its

implementation. A PETS subcommittee meeting with the team leaders was convened both at the

entry and the end of each phase.

5.6 QUALITY ASSURANCE

There have been several quality assurance elements integrated in the approach and methodology:

1. Consultations with the subcommittee at the entry and end of each phase was conducted to ensure continued compliance with agreed scope of the PETS.

2. The subcommittee was represented by managers and professionals with substantial knowledge of the primary and secondary education system in Tanzania as well as PFM issues more in general. The inclusion of these members in the actual implementation of various tasks beyond consultations in subcommittee meetings was done to test and verify information collected.

3. The generation of databases from existing sources allowed consolidation of information with data collected during field survey.

: Page 42

4. In the survey tool some questions were included for which the data were already known to enable testing of quality and accuracy of data collected with use of the survey tools.

5. During data compilation and analysis several test runs was conducted to check consistency of data collected.

6. Results of analysis was presented at various stake holder and subcommittee meetings which also served as additional quality assurance related to analyses and findings.

All the above contributed to the process of assuring that the data could be given a fair assessment

and an accurate interpretation for analysis.

: Page 43

6 OVERVIEW OF PRIMARY AND SECONDARY EDUCATION

6.1 PRIMARY AND SECONDARY EDUCATION IN TANZANIA

The Government of Tanzania takes education as one coherent entity, from basic through higher

education. To implement this line of thought it was decided in February 2008 to reorganise

government ministries – one result of which was the detachment of the Higher and Technical

Education Divisions from the then Ministry of Higher Education, Science and Technology to become

part of the Ministry of Education and Vocational Training (MoEVT). Components of Tanzania’s

education sector therefore include pre-primary, primary, secondary, teacher education, adult and

non-formal education, folk education, technical and vocational training, and higher education.

Except for folk education the entire spectrum of education is under the auspices of MoEVT.

The Tanzanian educational system is based on the 7-4-2 system: seven years of primary school,

followed by four years of secondary school leading to Ordinary Level (0-level) exams, followed by

two more years leading to the Advanced Level (A-level) exams.

Kiswahili is used as the medium of instruction in primary education and is also taught as a subject.

English is taught as a subject from Standard III onwards and is the medium of instruction in

secondary schools and subsequent institutions of higher learning. All primary school textbooks,

except English textbooks, are written in Kiswahili. However, English medium and international

primary schools use textbooks that are in English.

In the final year of primary education (Standard VII) students sit for the PSLE. In the second year of

secondary school, there is a national assessment examination that allows those who pass to

continue to study for an additional two years. The Form II examinations also aim at identifying weak

students to whom remedial teaching could be recommended and provided. After two years of

further study, students conduct the Certificate of Secondary Education Exam (CSEE), held in

November and the results are available in March of the following year. After another two years of

study, A-level exams are given. After the final year of secondary school – the thirteenth year –

students can take the Advanced Certificate Examination, which is recognized all over the world.

6.2 NATIONAL OBJECTIVES FOR EDUCATION

Provision of good quality education to all Tanzanians is accorded high priority by the Government of

Tanzania because of its centrality in bringing about the nation’s social and economic development.

The Government of Tanzania is committed to achieving its Education for All (EFA) goals in line with

MDGs. The following are the specific goals of EFA:

√ Ensuring equitable access to quality primary and secondary education for boys and girls, and

universal literacy among women and men.

√ Expanding and improving comprehensive early childhood care and education, especially for the

most vulnerable and disadvantaged children.

√ Ensuring that the learning needs of all young people and adults are met through equitable

access to appropriate learning and life skills programmes.

: Page 44

√ Eliminating gender disparities in primary and secondary education by 2005 and achieving gender

equality in education by 2015, with a focus on ensuring girls full and equal access to and

achievement in basic education of good quality.

√ Improving all aspects of the quality of education and ensuring excellence of all so that

recognised and measurable learning outcomes are achieved by all, especially in literacy,

innumeracy and essential life skills.

√ Ensure that, by 2015, children everywhere, boys and girls alike, will be able to complete a full

course of primary schooling.

6.3 PRIMARY AND SECONDARY EDUCATION DEVELOPMENTS

During the last decade, the Government of Tanzania has made major efforts in strengthening the

outreach, access and quality of its education system. The launching of the Primary Education

Development Plan (PEDP) 2002–2006 and the Secondary Education Development Plan (SEDP) 2004–

2009 were concerted government efforts to revitalise the education system on the mainland which

will continue under the umbrella of the Education Sector Development Program (ESDP).

The results so far have among others been impressive gains in coverage and enrolment in primary

and secondary education as illustrated in table 4. A particular issue is the significant growth in

number of secondary schools which has accommodated a similar growth in Net Enrolment Rate

(NER) for secondary education.

Table 4 – Key indicators of primary and secondary education

2002 2003 2004 2005 2006 2007 2008

Primary enrolment

(in 1000 students) 5,981 6563 7,083 7,541 7,959 8,316 8,410

NER Primary 80.7 88.5 90.5 94.8 96.1 97.3 97.2

Number of Primary schools 12,286 12,815 13,689 14,257 14,70

0 15,264 15,673

Secondary enrolment

(in 1000 students) 323 355 433 524 676 1,021 1,222

NER Secondary 6.0 6.3 5.9 10.1 13.1 20.6 23.5

Number of Secondary schools 1,059 1,083 1,291 1,745 2,289 3,485 3,798

Source: MoEVT EMIS data.

According to EMIS data there were 15,673 primary schools in 2008 of which 15,257 were

government funded. There were 3,798 secondary schools of which 3,039 were funded from the

central government and council budgets.

There has been a substantial increase in the number of primary schools and enrolment rates. In 2008

a total of 8,410,094 students were enrolled in primary schools from a figure of 5,972,077 in year

2002. The Gross Enrolment Rates (GER) for the year 2008 was 112.3% while NER was 97.2%.

However, there are some disparities across regions, such as 91.4% in Tabora as compared to 99% in

Mtwara.

: Page 45

Enrolment in secondary schools has also increased significantly. The total enrolment of Form I to VI

increased from 323,000 in 2002 to 1,222,403 in 2008 i.e. more than six times the enrolment of 2002.

The percentage of girls in Form I to IV was 45% while that of ‘A’ level was 40% of total enrolment in

2008. Secondary education NER (Form 1-6) was 23.5% in 2008 as compared to only 6.0% in 2002.

This increase is attributed to the increased number of community secondary schools constructed.

The increase in the enrolment has called for significant improvements in providing for support

facilities such as school furniture, water and sanitation, teachers’ houses, library and laboratories as

well as the recruitment and deployment of teachers.

6.4 BUDGET ALLOCATIONS AND EXPENDITURE FOR THE EDUCATION SECTOR

As illustrated in Table 5 the proportion of the national budget that is allocated to the education

sector has increased from about 17.4% in 2005/06 to 18.5% in 2006/7 which dropped slightly to

18.3% in 2007/08. In the revised budget estimates of 2008/9 the share of the education sector is

indicated at 19.8%. When analysed at an aggregate level there has been an increase [both in

absolute and relative terms] in budget allocation for the education sector over the three-year period

2005/06 - 2007/08. This is illustrative of the national effort to steer more resources towards the

education sector.

Table 5 – Budget allocations for Education (TSh million)16

Sub sector 2005/06 2006/07 2007/08 2008/09

Primary Education 390,974 491,243 544,220 666,419

Secondary Education 104,483 119,987 174,227 140,196

Vocational Training 8,926 10,654 18,978 8,007

Other Basic Education 36,137 56,208 49,661 69,630

Teacher Education - MoEVT 8,540 10,438 19,257 25,250

Inspectorate - MoEVT 3,712 4,920 4,800 5,900

Chief Education Officer - MoEVT 6,641 5,857 6,651 5,702

Administration - MoEVT 12,733 31,411 14,673 26,867

Culture & Nat. Languages - MoEVT 2,012 0 0 0

UNESCO Commission - MoEVT 0 0 0 381

Public Serv. Comm.(Teachers Service) 2,499 3,581 4,281 3,700

Folk Development 2,360 2,568 3,132 5,386

Total Basic Education 542,879 680,659 790,219 957,438

Total Higher Education 158,245 231,356 317,218 472,934

Total Education 701,124 912,015 1,107,437 1,430,372

Total Govt. Budget 4,035,100 4,850,600 6,066,800 7,215,631

Education to Total Govt. Budget 17.4% 18.8% 18.3% 19.8%

16 Source: MoFEA budget books. Official data for actual expenditures at the level of disaggregation displayed in the table above could not be obtained, and budget figures according to the budget heads displayed could not be obtained for fiscal years prior to FY2006.

: Page 46

Sub sector 2005/06 2006/07 2007/08 2008/09R

Primary Education 55.76% 53.86% 49.14% 46.59%

Secondary Education 14.90% 13.16% 15.73% 9.80%

Vocational Training 1.27% 1.17% 1.71% 0.56%

Other Basic Education 5.15% 6.16% 4.48% 4.87%

Education to Total Govt. Budget 17.4% 18.8% 18.3% 19.8%

Higher Education to Total Education Budget 22.57% 25.37% 28.64% 33.06%

Primary education has been the main beneficiary of the education budget in absolute terms. As

summarised in Table 5 the sub sector share of the education budget in 2005/06 was 55.8% but has

since dropped to about 49.1% in 2007/08. In 2008/9 this was further reducing to 46.6%. Secondary

education recorded a decrease in 2006/7 to 13.2% from 14.95 in 2005/6. However this increased to

15.7% in 2007/8. The allocation to secondary education although increasing in absolute terms will

decrease in proportion to 9.8% in the revised budget estimates for 2008/09. In contrast higher

education has received an increasing share in both relative and absolute terms in the same period.

6.5 GOVERNMENT BUDGET ALLOACTION AND EXECUTION PROCEDURES

In order to understand the process of expenditure tracking, and design an appropriate methodology,

it is necessary to be familiar with the institutional context for budget execution. With this in mind,

this section seeks to clarify how wage and non-wage recurrent and development resources are

allocated and how budgets are executed in the course of a fiscal year as ministries, regions, councils

and schools implement their activities.

In general, ministries, regions, councils and schools have three main sources of financing for

expenditures on 'goods and services' – the national budget, the council budget (from self generated

revenues) and private funds such as 'off-budget' donor contributions and contributions from

households.

The annual cycle of budget allocation and budget execution in Tanzania begins when MoFEA

communicates indicative budget ceilings to the respective ministries, departments, agencies and

institutions. By law, the budget must be approved by parliament prior to the fiscal year, and budget

implementation begins on 1 July and the fiscal year ends on 30th June.

After approval of the budget by the Parliament, the MoFEA communicates the final budget ceilings

to ministries, departments, agencies and institutions.

The process of budget execution begins each year with the transfer of funds by the Bank of Tanzania

from the Consolidated Fund to the Paymaster General Expenditure Account. Subsequent stages of

budget execution are as summarized in figure 5 below and narrated subsequently.

: Page 47

Figure 5 - Flow of public funds for primary and secondary education

MoFEA

ACCOUNTANT

GENERAL

MoEVTPMO-RALG MoCDGCPO-PC

Vote 56 Vote 46 Vote 53Vote 66

District Council

Bank Accounts

RAS Regional

Sub-Treasury

Recurrent Development

Sub-vote

Holders

Secondary

Schools

TASAF

TEA = Tanzania

Education Authority

Vote 6220

Other Accounts

Private

Transfers

School Accounts

TEA

Pre-primary and

Primary Schools

Day Care

Centres

Folk

Education

Colleges

School

Account

6.5.1 TRANSFERS TO MINISTRIES, COUNCILS AND SCHOOLS

6.5.1.1 LEVEL 1 – TRANSFERS TO MINISTRIES (VOTE HOLDERS);

Once the resource envelope has been determined and allocations made of resources to different

ministries, MoFEA issues an Exchequer Issue Warrant which allows the Bank of Tanzania (BoT) to

transfer funds from the Consolidated Fund to the Paymaster General Expenditure Account.

The Accountant General then issues Exchequer Issue Notifications to Ministries in order to notify the

Ministries (MoEVT, PMO-RALG, etc) that funds are available.

On receipt of the Exchequer Issue Notification, the Exchequer Section (MoFEA) credits the

Paymaster General Cash Book for the relevant Ministry on the Central Payment System with the

amount transferred from the Exchequer Account.

The Accounting Officer at the Ministry (Permanent Secretary), on receipt of the Exchequer Issue

Notification, informs the relevant Warrant Holders; councils in the case of PMO-RALG; Headmasters

and TEA in the case of MoEVT; and Tanzania Social Action Fund (TASAF) in the case of the President’s

Office - Planning Commission (PO-PC) that resources are available via a Warrant of Funds

Notification Report. This report authorises each Warrant Holder to spend the funds allocated for

activities shown on the warrant.

The Warrant Holder authorises payments based on approved claims for payment up to the amount

available and for the purpose stated on the Warrant of Funds Notification Report. It is each Warrant

Holder’s responsibility to ensure that standard government procedures in respect of purchases are

observed.

: Page 48

6.5.1.2 LEVEL 2 - TRANSFERS TO COUNCILS AND OTHER INSTITUTIONS (SUB VOTE HOLDERS);

Transfers to Councils for Primary Education – For transfers to councils the Accounting Officer PMO-

RALG requests the Accountant General (AG) to release funds equal to the approved amounts by

producing a Payment Voucher in favour of each council and sends it to the Central Payments Office

for Telegraphic Transfer preparation. A Telegraphic Transfer is prepared in favour of the Council’s

bank accounts (Development and Recurrent) duly authorized by the Accountant General. These

transfers are then made through commercial banks (the National Microfinance Bank, in nearly all

cases).

These transfers are made with a copy of details of transfers routed to the Regional Administrative

Secretary. The Regional Secretariat is an extension of PMO-RALG in the regions. It has specialists to

support services in all sectors and specifically in this case education (Regional Education Officer). It

assists the PMO-RALG in executing its functions in the council by consolidating council reports at

regional level in line with statutory requirements.

Transfers to Secondary Schools - For transfers to secondary schools located outside Dar es Salaam

the transfers are executed as follows:

1. The MoEVT prepares a Warrant of Fund indicating the various items to be funded at the school;

2. A Telegraphic Transfer is effected by the Accountant General through a Bank payable to the regional sub-treasury where the school is located;

3. A copy of the Telegraphic Transfer, duly signed by the Bank is sent to Sub-Treasury Head Office at the AG Department. This facilitates the tracking of all funds sent to Sub-Treasuries.

4. At the same time, the original Warrant of Fund is remitted to the Headmaster as the Warrant Holder notifying him of the allocation of funds sent to the Regional Sub-Treasury. On the other hand, the Sub-Treasury receives a copy of the Warrant of Funds to support the funds remitted by Telegraphic Transfer.

5. Upon request, the Sub-Treasury will issue a cheque against funds of the School.

6.5.1.3 LEVEL 3 - TRANSFERS DIRECTLY TO SCHOOLS

All Government secondary and primary schools are eligible to receive funds allocated through the

Government Grants System. The grants receivable by schools are split into capitation grants and

development grants. Schools may also receive funds to cover other recurrent needs. Councils will

make disbursements of grant funds to primary schools through the School Development and

Capitation Grant Accounts.

For secondary schools the Regional Sub-Treasury effects all payments initiated by Heads of Schools

on the strength of fund balances maintained at the Regional Sub-Treasury.

6.5.2 OTHER TRANSFERS DIRECTLY TO SCHOOLS

There are instances where private institutions and individuals disburse funds and resources directly

to the schools. Where such disbursements occur, the procedure used to account for such

transactions is as follows:

: Page 49

√ Private funds, goods or services are directly given to the school;

√ Cash contributions should be collected according to procedures of collecting government revenue and deposited in the school Revenue Collection Account.

√ Contributions in kind are to be recorded in the schools’ records. Estimated values can be assigned to such contributions for reporting purposes. The school records show the value or amount and/or descriptions of goods, services or equipment received with relevant supporting evidence if available;

√ The schools include these resources received in its reports submitted to the council for subsequent and eventual recording by the Accountant General.

6.5.3 THE GRANT ALLOCATION SYSTEM FOR EDUCATION

There were three main types of grants allocated in FY2008 for councils; an Education Block Grant

(EBG), a Capitation Grant (CG) and Primary Education Development Grant which was allocated by

MoEVT outside the regular local government grant system.

The Education Block Grant was a recurrent grant consisting of three components;

1. A discretionary component for Personnel Emoluments (PE) which is predominantly teacher

salaries based on the number of positions allocated to each Council. The allocation took into

account the current commitments related to existing positions in the councils

(predominantly teachers) and policies to increase teacher allocations to some councils with

high pupil/teacher ratios.

2. A formula based element to cover Other Charges (OC).

3. A discretionary allocation for expenditures on special schools applicable to some councils.

According to the budget guidelines for FY2008, the allocation of the education block grant between

councils was 100% based on the number of school-aged children in each council (based on census

figures). In addition to the formula-based allocation, there was also an earmarked element of the

education block grant for funding special schools in specific councils (for example schools for the

blind). These amounts were not allocated by formula because not all councils have such facilities and

due to the special nature of this expenditure item. In addition, resources for examination fees as

well as resources for capitation spending at the school-level (i.e. the previous PEDP capitation grant)

were also allocated as earmarked OC allocations. These resources was allocated among councils on a

per-pupil basis. As a result, these earmarked amounts were shown separately in the grant

computation and was to be used for the funding of special schools, examination fees and capitation

spending only.

Councils were to allocate the OC amount of the education block grant firstly for council-wide

activities as needed. The balance of the funds was then to be distributed to the schools according to

the enrolment figures for capitation purposes (to supplement the earmarked capitation resources).

The following further conditions and limitations of the use of the education block grant was to be

observed:

√ Councils were to budget within the OC element of their grant for the cost of national

examinations, at an amount not less than the actual amount incurred in the previous year.

: Page 50

√ Councils were expected to provide TSh 3,000 (or 6,000 TSh17) per enrolled pupil in capitation

resources to the school level from the Education Block Grant.

√ School-level funding should be distributed between the schools in a council in accordance to

the number of pupils enrolled in each of the schools. This condition applied separately to PE

(staff postings) as well as OC.

√ School plans should be the basis for the use of OC resources at the school level. School level

plans and budgets were to be consolidated into those of the council.

√ School level OC resources (capitation funding) were to be used for: text books, teaching and

learning materials, maintenance, minor repairs, furniture, and school administration.

√ No more than 10% of school-level OC resources should be allocated for school

administration, including allowances and transportation.

√ School level OC resources should not be used for spending on capital infrastructure, such as

construction of classroom or teachers’ houses.

In addition to the Education Block Grant specific formula based Capitation Grant was also allocated

to councils equivalent to 5,000 TSh of the projected school age population in the district. It was

intended as a transfer to councils for onward transfer to schools as a capitation grant in addition to

the capitation grant allocation from the Education Block Grant .

In addition to the recurrent block grant and special capitation grant for primary education, a primary

education development grant amount was to be allocated to each councils by MoEVT, i.e. not part of

the Local Government regular allocation system but allocated at the discretion of MoEVT.

17 The budget guidelines for the fiscal year 2007/08 states to different amounts related to the same grant in different sections of the guidelines.

: Page 51

6.6 PRIVATE FUNDING FOR EDUCATION

According to the NBS Household Budget Survey for 2007 (HBS 2007) there is significant private

expenditure for education. Although the data sample do not allow cross comparison among regions

and districts, the results shows that expenditures varies across councils and households, in the latter

case according to the household income.

The HBS 2007 reflects growth in total household expenditure from 2001 to 2007. As illustrated in

figure 6, household expenditure has increased by as much as 40% in real terms and more so for Dar

es Salaam households than households in other urban areas and rural areas.

Figure 6 – Average monthly household expenditure (in 2007 TSh)18.

According to NBS data from the HBS 2007, households also spend more on education than seven

years ago and significantly more so for urban areas than rural areas. Across all households the share

of education expenditure to total expenditure has doubled since 2001. This is linked to the

significant growth in enrolment among others, for secondary education; growth in number of private

schools and also enrolment for higher education.

Private contributions to primary and secondary schools are discussed further in the following

sections based on the results from the sample survey.

18 Source: NBS Household Budget Survey 2007

-

10,000

20,000

30,000

40,000

50,000

Dar es

Salaam

Other urban

areas

Rural areas

2001 2007

: Page 52

7 SURVEY RESULTS PRIMARY EDUCATION

7.1 DATA FROM NATIONAL RECORDS AND SAMPLE FOR PRIMARY EDUCATION

In order to analyse revenue and expenditure data for councils and schools per student and teacher,

we needed to establish the actual number of students and teachers. The actual number of students

in accordance with school and council records vary depending on what time they are registered as

enrolled. For most primary schools enrolment in grade I is a continuing process starting at the

beginning of the school year and usually reaching its peak around March but in some cases the peak

enrolment at a school is not reached before May/June. There are also several dropouts during the

school year.

Comparing end year data from councils with end year data from school shows that there are 3.6%

more students according to council data compared to school data. The deviation in reporting was

observed for 39% of the schools in the sample. Since council data and projections are used for grant

allocations we have used the council data on enrolment when analysing capitation grant, capital

development grant and other grants per student from councils. We have however used school level

data when analysing actual school level costs per student.

When comparing statistics related to school employment provided by the council with data from the

schools, there are more school employees on the government payroll according to the council

records than employees on the payroll according to the schools. This was the case for 31% of the

schools in our sample. On the other hand, there are schools that claim to have more employees on

the government payroll than the number of employees paid according to the records in the councils.

This is because in some cases a teacher may be registered in the council payroll in a transitional

period after the teacher has actually moved to another school within or outside the council. The

excess number of government employees at the schools according to the council payroll when

adjusting for the above number of 'teachers in transit' compared to school data was 3.7%.

When analysing data on teachers, students and expenditure levels for mainland Tanzania in total we

have used national records even if they deviate from council records according to data collected

through this survey19.

The districts have been grouped into Dar es Salaam, other urban centres ('Other Urban') meaning

regional capitals, and rural districts (i.e. all others). Results showing totals for all schools based on

survey data are weighted averages in accordance with the sample's share of total schools in

mainland Tanzania.

19 Number of teachers and students according to council records deviate from the data presented by MoEVT EMIS data however the largest deviation found between Council and EMIS records was 3.2% for teachers and 4.3% for students.

: Page 53

7.2 FLOW OF RESOURCES FOR PRIMARY EDUCATION

7.2.1 CENTRAL GOVERNMENT ALLOCATIONS

According to official data, primary education was allocated 544 bln TSh through the state budget in

FY200820. With a total of 8.3 million students enrolled in government primary schools in 2008

according to MoEVT data the total Central Government allocation for education per primary student

was 66,646 TSh per student (52.5 USD)21.

Table 6 - Tanzania - State budget Primary Education 2007/200822

Recurrent Development Total In bln. TSh

MoEVT 12.7 14.4 27.1

PMO-RALG - PEDP - 0.2 0.2

Regions Vote 1.0 2.5 3.5

Transfers to councils 507.3 6.2 513.5

Total primary education 521.0 23.2 544.2

Percent distribution

MoEVT 2.4% 61.9% 5.0%

PMO-RALG - PEDP 0.0% 0.8% 0.0%

Regions Vote 0.2% 10.8% 0.6%

Transfers to councils 97.4% 26.6% 94.3%

Total primary education 100.0% 100.0% 100.0%

The amount allocated as education grants to councils was 526 bln. TSh (including grants charged to

the MoEVT budget outside the LGA grant allocation system).

The actual amount received by councils as education sector capital and recurrent grants for FY2008

was 473 bln. TSh (87% of the state budget for primary education). This amounts to 57,417 TSh per

primary student (45.2 USD).

The total council level expenditure for education in FY2008 was 495 bln. TSh i.e. 22 bln. TSh more

than the Central Government transfers for education received by the councils. This is due to the fact

that councils also have own revenue and receive general purpose grants from the state budget. Both

are discretionary sources of funding from which many councils spend part of the amount for

education. In addition councils also receive other transfers and direct payments by RAS and Central

Government ministries (in 52% of the districts)23.

20 MoFEA budget books 21 The budget allocation was based on an estimated 7.9 million students during FY2008 which is equivalent to 69,096 TSh per student. 22 Data from MoFEA budget books. 23 Official accounting records indicate that RAS contributed to 4% of LGA education expenditures and our survey data also indicate that MoEVT has made in kind contributions to primary schools although the value of these contributions are small (less than 0.03% of total school level expenditures).

: Page 54

In addition to primary education, adult education and education administration, many councils (as

many as 26% of the councils) also spent funds for secondary education to supplement efforts by

communities for construction of and equipping new secondary schools. In total the councils'

contribution to secondary education accounted for approximately 2% of total LGA education

expenditure.

7.2.2 EDUCATION GRANTS

7.2.2.1 THE GRANT ALLOCATION SYSTEM

State budget allocations in the form of education sector recurrent and development grants were the

main sources of funding for councils expenditure on education in FY2008. The grants were allocated

through a complex system of discretionary and earmarked grants.

As part of the budget preparation process councils were provided allocation ceilings for the different

types of discretionary and non discretionary grants they could expect to receive from the Central

Government. Some grants were formula based, some allocated on the discretion of Central

Government.

Based on these allocation ceilings communicated to the councils, they prepared their budgets.

However, during the budget preparation process the initial ceilings were adjusted as reflected by the

deviation between ceilings allocated and approved council budgets. As illustrated in table 7, this

change in the budget preparation process favoured rural councils on account of urban councils

within the overall resource envelope.

Table 7 - Education grants to councils - Initial ceilings and actual budget FY2008 (in bln. TSh)24

Councils Ceiling Council budget Approved budget /

ceilings allocated

Dar es Salaam 31.8 29.2 92 %

Other urban 66.3 62.0 93 %

Rural 427.5 435.6 102 %

Total 525.6 526.8 100 %

Once approved, MoFEA released the grants to the councils through the regional sub-treasuries. The

actual amount of each periodic release was decided upon by the Central Government (release

committee) taking into account actual available resources (revenue collection and external funding).

With this form of cash flow management the Central Government ensured that expenditures were

not executed beyond available resources i.e. a procedure for cash flow management to maintain

overall government fiscal targets for the year.

24 Based on PMO-RALG data from council financial statements (Login Tanzania database) and data on ceilings presented in budget guidelines.

: Page 55

In terms of education, there were three main types of grants allocated in FY2008; an Education Block

Grant (EBG), a Capitation Grant (CG) and Education Capital Development Grant which was allocated

by MoEVT outside the regular local government grant system.

The Education Block Grant was a recurrent grant consisting of three components;

1. A discretionary component for Personnel Emoluments (PE) that was predominantly teacher

salaries based on the number of positions allocated to each council. The allocation took into

account the current commitments related to existing positions in the councils

(predominantly teachers) and policies to increase teacher allocations to some councils with

high pupil/teacher ratios.

2. A formula based element to cover Other Charges (OC).

3. A discretionary allocation for expenditures on special schools applicable to some councils.

The Capitation Grant was a formula based grant initially allocated to councils equivalent to 5,000 TSh

of the projected school age population in the district. It was intended as a transfer to councils for

onward transfer to schools as a capitation grant.

The EBG component for Other Charges was allocated on the basis of students enrolled and intended

to cover various expenditure related to examination charges and other recurrent education

expenditure by the councils25. The remaining amount was added to the Capitation Grant transfers to

schools.

According to the budget call circular for 2007/2008 councils were expected to transfer an amount of

the EBG of minimum 3,000 TSh (or 6,000 TSh) per student as additional capitation grant resources26.

The total amount calculated as Capitation Grant from councils to schools was to be calculated on the

basis of the estimated number of students enrolled in the school.

As per guidelines this implies that the schools should have received 8,000 TSh or 11,000Tsh per

student enrolled (the amount depends on what section of the guidelines that were to be followed).

The distribution of the education sector allocations made in the budget call circular is illustrated in

figure 7 below.

25 The English version of the guidelines states that the allocation was based on number of school age population while the Swahili version states it was based on number of students enrolled in schools in 2007. The allocation table used by MoFEA appears to indicate that the allocation was made according to students enrolled. 26 In one section of the budget guidelines the level of capitation grant mentioned is 3,000 TSh per student, however, later in the same document the amount mentioned is 6,000 TSh per student.

: Page 56

Figure 7 - Distribution of education sector grants FY2008

As figure 7 illustrates, the grant element for salaries constitute the main element of the Education

Block Grant and the main component of total grants (80.6%). Other charges of the Education Block

Grant constitute the second main element (11.4% of total recurrent grants). The balance after the

councils have covered their expenditures was to be transferred to schools together with the

earmarked capitation grant as the total Capitation Grant for each school. The balance of the EBG

added to the capitation transfers to school will necessarily differ between councils and will result in

variations between councils in the actual amount of capitation grants per student.

7.2.2.2 HOW MUCH OF GRANTS ALLOCATED ARE TRANSFERRED TO COUNCILS

Once the budget is approved, funds are released to councils for payment of their salary and non

salary expenditures in education. As table 8 shows, actual approved allocations of education grants

have to a large extent been released. The exception is the Education Capital Development Grant

which is linked to progress in implementation of infrastructure projects (building of new schools,

other buildings, classrooms and school rehabilitation projects). Another feature is the lower level of

releases in particular for rural councils both in terms of capitation grants and capital development

grants.

Table 8 - Allocation and actual release to councils of education grants (TSh million) 27

Councils Education Block Grant Capitation Grant Capital Dev. Grant Total Allocation 496,883 25,484 30,574 552,942

Releases 445,276 18,664 12,428 476,368

% released

Dar es Salaam 99.1% 98.2% 85.9% 98.8%

Other urban 92.1% 96.5% 41.4% 90.5%

Rural 88.5% 69.9% 39.7% 84.7%

Total 89.6% 73.2% 40.6% 86.2%

27 Based on PMO-RALG data from council financial statements (Login Tanzania database).

EBG Salaries80,6 %

EBG special schools0,4 %

EBG other11,4 %

Capitation grant7,6 %

: Page 57

While the above may suggest that the allocation for the councils are broadly released as allocated

(with the few exceptions mentioned for rural districts), the level of releases for individual councils

display significant variations i.e. a significant reallocation of grants between councils is taking place

within the year.

The variation in releases compared to approved allocations is first and foremost due to level of

budget execution by councils. This is in turn related to their ability to employ teachers as planned for

and reflected in their budget, and their capacity to execute planned investments. Analyses related to

variation in releases will be presented in more detail in sections below.

7.2.2.3 EDUCATION GRANTS AND LEVEL OF EXECUTION

When comparing total education grant transfers, both recurrent and development grants, with

councils' total expenditure on education it shows that several councils spend less on education

expenditure than what they have received as grants for education, i.e. for several councils the grant

amounts were not spent or has been spent for other purposes than education.

When comparing education recurrent grant transfers with council recurrent expenditure on

education it also shows that several councils spend less on recurrent education expenditure than

what they have received as Education Block Grant and Capitation Grant for education, i.e. in several

councils part of the recurrent transfer for education is not spent or spent other purposes than

education.

To determine the amount of education grants not spent on education we assessed the amount of

education grant surplus that contributed the overall surplus in the councils (total amount of unspent

grants). In the councils where the grant surplus did not contribute to an overall surplus in the

council, it means it has been spent on a different sector or purpose than education.

Table 9 - Difference between Council education grant allocations and actual expenditure on education 2007/08 (in TSh)28

Education grants

fully spent on education

Education grants not fully spent

Education grants spent for different purposes

Total grant released 224,421,925,830 131,631,386,478 120,307,281,124

Surplus/deficit release (21,810,336,259) 11,829,105,880 28,898,716,418

Share of release 9.7% 9.0 % 24.0 %

Number of councils 65 34 32

Table 9 presents the difference in total state budget grants for education and total education sector

expenditure by the councils. The councils have been grouped into three categories;

1. Councils that spend more on education than what they have received in total education

sector grants (recurrent and development).

28 Based on PMO-RALG data from council financial statements (Login Tanzania database). Detailed data were available for 131 out of the 132 councils and excluding Dar es Salaam City Council expenditures.

: Page 58

2. Councils that spend less on education than what they have received in total education sector

grants with the surplus remaining unspent.

3. Councils that spend less on education than what they have received in total education sector

grants and has spent the savings in education sector grants on other sectors/purposes than

education.

In 65 councils the education grants received have been less than actual expenditures for education

i.e. they have also used other sources of funding for education expenditure (like local revenue,

discretionary grants, and in some cases also used other sector grants).

In 66 councils the educations grants received exceeds actual expenditures for education i.e. the

balance of grants over expenditures means the grants have not been fully spent (contributed to an

overall surplus of total grants over expenditure) or used for a different purpose or sector than

education29. Out of these councils, 32 have used parts of education sector grants for a different

purpose than education. These councils are to be found in all regions of mainland Tanzania. The

amount of education grants used for non-education purposes is 28.9 bln. TSh (6.1% of total

education sector grants transferred to councils or 24.0% of the education sector grant transfers to

these 32 councils).

7.2.3 RESOURCES FOR PRIMARY SCHOOLS

As mentioned in sections above, councils receive grants from the state budget which are used for

funding of salaries, non salary education expenditures, investments in education infrastructure, and

finally, transfers of funds to schools as capitation and development grants.

Analyses of allocations for individual schools are based on survey data since there are no

consolidated information of councils' transfers to and expenditures for individual primary schools.

The survey data were collected for each school from the accounts of the district council as well as

from the school cash books, bank statements and records.

Figure 8 - Average council spending on schools – distribution by main cost items

29 After adjusting for unspent balances remaining in district accounts.

LGA Salaries

84 %

LGA other

expenses7 %

LGA Grants to

schools9 %

: Page 59

The average distribution in allocation by the councils for different cost items is shown in figure 8. In

the figure 'LGA other expenses' are contributions in kind by the councils like various teaching

materials and other inputs procured by councils and distributed to the schools. The main item of

expenditure is teacher salaries. As will be presented further in the sections below it is also

allocations and transfers for teacher salaries that is the main component when assessing losses in

resource flows and variations in distributions of resources between councils.

7.2.3.1 ALLOCATION OF TEACHERS AND SALARIES

As mentioned above, teachers are the main input to primary schools in financial terms. To assess

actual resource flows to schools we have analysed the extent to which the teachers on government

payroll are present at the schools. Some teachers may be absent from schools for a variety of

reasons;

In a transitional period they may appear on a council/school payroll even after they have

been employed at a different school in the same council or in a school in a different council.

They may have been granted leave of absence for training or postgraduate education and as

such are absent from the school for a longer period of time (in our sample as much as three

years in some schools).

They may be absent due to longer term sick leave or similar reasons preventing them from

attending the school.

They may be serving in other parts of the civil service on a temporary or permanent basis

e.g. to support council administration, but with the salary still charged by the council as

teacher salary for the school.

They are absent for a reason unknown to the council and the school.

When assessing teachers on payroll compared to teachers actually attending the school we have

adjusted for the number of teachers that in a transitional period continue to be accounted for as

present at a school despite that they have been transferred to a different school. The other reasons

for teachers' absence proved to be difficult to determine since in most cases school head teachers

were not aware of the reasons why there were absent in a longer period of time. Only in few cases

were they aware of teachers being absent because they were granted leave of absence for different

reasons. The results of the analysis presented below are accordingly a mix of teachers' absence due

to eligible reasons as well as teachers being absent for unknown reasons.

The number of staff on government payroll recorded by the councils deviates from the number

recorded at schools. The difference is estimated at 3.7% according to survey data30. More significant

however is the number of teachers on the payroll compared to the number of teachers actually

present in the school.

With the school level survey tool, two sets of data were collected;

30 The estimate has been adjusted for the number of teachers on the payroll that have been transferred to a different school within the council or to a school within another council but who in a transitional period feature on the payroll of the school and/or council.

: Page 60

1. the number of teachers on government payroll allocated to the school and the number of

teachers absent the full school year, and

2. the number of teachers present less than 50% of the year (recorded as 50% part time

teacher in our survey) and the number of teachers present more than 50% of the year

(recorded as full time teacher).

This approach will eventually give an estimate of full time teaching above what is the reality since

part time teachers with less than 50% are considered as present 50% of the time while all teachers

present more than 50% of the time are counted as full time teachers.

To qualify the data further we also collected data on number of teachers on the payroll in May 2008

and the number of these teachers actually present in the school during the same month. In total

these data led to a fair estimate of the number of teachers on government payroll who is actually

performing in the schools.

The results are presented in table 10 below with an estimate of the equivalent cost of teachers not

present at the school. In financial terms this loss has been estimated at 845 million TSh equivalent to

13.9% of total teacher costs or 9.5% of total primary school expenditure by councils. If the figures

are used as proxy for all primary schools in mainland Tanzania, the total financial loss of teacher

inputs for mainland Tanzania was 47.0 bln TSh in total (equivalent to 37.8 million USD).

Table 10 - School employees on government payroll - survey data

Councils Payroll according

to council Payroll according

to school Payroll

present at school Estimated cost

teachers absent in million TSh

Dar es Salaam 1,000 993 887 239

Other urban 736 712 644 115

Rural 2,173 2,101 1,875 491

Total 3,909 3,806 3,406 845

The issue of teachers on council payroll not performing at the school was observed in 49% of all

schools in the sample and in all councils. However, in most councils the number of teachers absent

compared to the number on the payroll was less than 10%. In 13 councils more than 10% of the

teachers were absent of which in 3 councils more than 20% were not performing at the school.

7.2.3.2 FLOW OF CAPITATION GRANT TO SCHOOLS

Comparing council records with school data for the same schools shows that what has been

accounted for in council records as capitation grant transferred to the Micro Finance Bank for

distribution to individual schools is more than what was received by schools according to their cash

books and bank statements. This observation is illustrated in table 11 below.

Table 11 - Average capitation grant per student according - survey data (in TSh)

Councils Capitation grant from council Capitation grant received by school

Dar es Salaam 4,405 3,736

Other urban 4,682 4,192

Rural 4,574 4,251

Total 4,570 4,189

: Page 61

According to the data the schools received approximately 8.3% less than what was transferred from

the councils. However, these averages hide the fact that there are significant variations between

councils and schools. A deviation between council and school accounts was found in 41% of the

schools and in 19 of the 27 councils in the survey. In 30% of the schools the deviation was more than

100,000 TSh between what was accounted for as allocated by the council compared to what was

actually received by the school. In 8% of the schools the deviation was more than 1 million TSh.

If using the above as a proxy for deviation in all schools of mainland Tanzania the total amount

would be 3.2 bln. TSh (approximately 2.5 million USD).

7.2.3.3 FLOW OF CAPITAL DEVELOPMENT GRANT TO SCHOOLS

When comparing capital development grant transfers according to council records with school

records of receipts as recorded in the cash books and bank statements, the same picture emerges as

for the capitation grant, i.e. council records present a higher amount than schools. However, in this

case there are schools that have not recorded any development grant as received even if the council

has accounted for the grant as disbursed to the school.

Table 12 - Average development grant per student - survey data (in TSh)

Councils Council records School records

Dar es Salaam 1,279 624

Other urban 1,597 1,389

Rural 1,496 1,144

Total 1,486 1,120

According to council records as many as 48% of the schools were allocated Capital Development

Grants. According to school records 18% of the schools have recorded receipt of these development

grants. The deviation between total development grant accounted for by the council compared to

school records is approximately 27%. If using this as a proxy for deviation in all schools of mainland

Tanzania the total amount would be 3.0 bln. TSh (approximately 2.4 million USD).

7.2.3.4 TO WHAT EXTENT ALLOCATIONS FOR PRIMARY EDUCATION REACH SCHOOLS

The results of the above analysis can be summarised into the following;

Total Central Government allocations to councils for education per primary student was

66,646 TSh (52.5 USD).

The actual amount released by treasury and received by councils as education sector capital

and recurrent grants for FY2008 was 57,417 TSh per primary student (45.2 USD).

In 66 councils, expenditure for education exceeded the total education grants received.

In 34 councils the total education grants received was not fully spent. This constituted 1.2%

of the total education grants transferred to all the councils.

In 32 councils part of the education grants received was also spent on other

sectors/purposes than education. This constituted 6.1% of the total education grants

transferred to all the councils.

: Page 62

The net amount of total education grants transferred to councils and used for education was

52,541 TSh per student (41.4 USD) equivalent to 78.8% of the grants allocated and 91.5% of

the grants actually transferred to the councils.

This is summarised in table 13 below.

Table 13 Education grant allocations, releases and actual expenditure

MoFEA Education grant transfers and council expenditures Per

student in TSh

Per student in USD

Percent of approved

MoFEA Education grants - Approved budget 66,646 52.5 100.0%

MoFEA Actual releases to councils 57,417 45.2 86.2%

- of which releases not executed by the councils 794 0.6 1.2%

- of which releases used for other sectors by the councils 4,082 3.2 6.1%

MoFEA Net grant releases executed by councils for education 52,541 41.4 78.8%

Council expenditure on education 59,697 47.0 89.6%

- of which MoFEA grant releases executed by councils 52,541 41.4 78.8%

- of which own revenue and transfers from others 7,156 5.6 10.7%

The total council expenditure for education was 59,697 TSh per student in government primary

schools (47.0 USD). The difference between education grants per student transferred and actual

expenditure per student is made up of other resources received by the councils than education

grants from Central Government. These are councils' own revenues, contributions from RAS and line

ministries as well as use of other grants by councils (general and other sector grants received by the

councils from Central Government).

The above gives a full account of education sector resource allocations and transfers from Central

Government to councils. However, it does not give account of what amount is subsequently

transferred and used for the benefit of primary schools.

The first step was to identify how much of total council expenditure was accounted for as primary

education. Actual expenditure by councils on primary education was 90.1% of total education

expenditure equivalent to 53,788 TSh per student (42.4 USD).

Table 14 - Council education expenditure

Council education expenditure Per student in

TSh Per student in

USD Percent of total

Total education 59,696 47.0 100.0%

- of which administration 3,790 3.0 6.3%

- of which adult education 1,099 0.9 1.8%

- of which secondary education 1,019 0.8 1.7%

- of which primary education 53,788 42.4 90.1%

The second step was to determine how much was spent for direct benefit of primary schools.

Of the amounts recorded by councils as primary education expenditure, 84% were for salaries of

school employees equivalent to 44,910 TSh per primary student (35.4 USD). However, not all

: Page 63

teachers on the payroll are actually present at the school, which represents a net loss of 5,779 TSh

per student (4.6 USD).

Capitation grants and development grants transferred to primary schools account for 12% of total

primary education expenditure equivalent to 6,436 TSh per student (5.1 USD). However, the amount

actually received by the schools was 6,046 TSh per student (4.5 USD) i.e. a deviation between

transfers and actual receipts of 746 TSh per student (0.6 USD).

In addition to the above, councils spent another 2,443 TSh per student for primary education as

contributions in kind to the primary schools of which 668 TSh per student were payments for

contributions in kind to schools in the form of text books, desks, chairs and other inputs.

Table 15 - Council education expenditure for primary schools

Council expenditure primary schools Per student

in TSh Per student in

USD Percent of

total Council expenditures primary education 53,788 42.4 100.0%

Council expenditures teachers 44,910 35.4 83.5%

- of which teachers present at schools 39,131 30.8 72.7%

- of which teachers absent from school 5,779 4.6 10.7%

Council transfers Capitation Grant 4,950 3.9 9.2%

- of which Capitation Grant received by schools 4,570 3.6 8.5%

- of which Capitation Grant not received by schools 380 0.3 0.7%

Council transfers Capital Development Grants 1,486 1.2 2.8%

- of which Development Grant received by schools 1,120 0.9 2.1%

- of which Development Grant not received by schools 366 0.3 0.7%

Other primary school expenditure by councils 2,443 1.9 4.5%

-of which contributions to schools according to councils

668 0.5 1.2%

Total primary school receipts cash and in kind 45,489 35.8 84.6%

The above leads to the following conclusions as concern overall resource allocation and use for

primary schools;

The education grant allocation in FY2008 by MoFEA was 66,646 TSh per student (52.5 USD).

Of the total amount used for education by the councils, 90.1% was used for primary

education.

Using the above as a proxy for the primary education share of education grants allocated,

the total amount of grants allocated to primary education was 60,050 TSh per student (47.3

USD).

The amount actually spent for the direct benefit of primary schools (salaries of school

employees, capitation and development grants as well as in kind contributions from

councils) was 45,489 TSh per student (35.8 USD). The amount constitutes 68.6% of the

amount initially allocated by the Central Government under the budget votes for primary

education. It constitutes 84.6% of total primary school expenditure as accounted for by the

councils.

The difference between Central Government allocation under the budget sub-votes for primary

education and actual spending directly benefitting primary schools can be explained by;

: Page 64

86.2% of the amount allocated was released.

Parts of the releases were for adult education, education administration and also used for

secondary education.

34 councils did not spend the full amount released.

32 councils spent part of the amount released for other purposes than education.

Councils paid salaries for teachers not present in primary schools.

Capitation and development grants did not reach many of the schools with the full amount

transferred to them.

In order to raise levels of resources for primary education and schools as reflected by the budget

allocations the results of our analyses suggest some of the following to be considered;

Capacity in execution of education grants appears to be a constraint for 34 councils. These

councils should be devoted specific attention to identify the constraints in execution. As will

be shown in sections below, part of the constraint appears to be their ability to attract

and/or retain teachers allocated to them and they generally have lower expenditure per

student than the average.

Another 32 councils spent part of the education grants released for other purposes than

education. This can be partly attributed to less than planned capacity in execution of

education expenditure, but also changes in sector priorities within the year by the councils

since budget allocations for education is shifted towards other sectors.

The cost of teachers not performing in schools represents the most significant efficiency loss

for primary schools. Whether teachers are absent for eligible or non eligible reasons, teacher

attendance should be subject for closer monitoring to raise the level of attendance.

Both Capitation and Development Grants represent small shares of total school level

expenditures but nevertheless significant shares of non-wage school level resources. There

is a scope to raise the level of allocation of these grants to schools by reducing the amount

of school contributions managed by councils or ensure that councils transfers more of the

capitation grant to schools as initially demanded in accordance with budget guidelines.

Schools are the intended beneficiaries of the school grants (capitation and development

grants). However, since it is effectively left to the discretion of councils to decide how much

each school shall get it leads to very diverse resource levels for schools, both between

councils and between schools within some councils. To ensure that grants actually reach

schools with the intended amount, they could be granted directly to schools rather than

councils transferring the at their own discretion remaining balances after accommodating

other expenditures within the councils.

Parts of the above are also linked to how resources are distributed between councils and schools

which will be analysed further in the following sections.

: Page 65

7.3 DISTRIBUTION OF RESOURCES

7.3.1 ALLOCATION AND TRANSFERS OF GRANTS TO COUNCILS

Analyses already during the inception phase showed that the distributions of funds vary between

regions, but even more significantly between councils in each region. Councils influence resource

allocation and expenditure for primary education, regions do not. Accordingly, while there are

variations in resource allocation between regions, regional variations are of less interest from an

analytical point of view since education grants for primary education is transferred to councils, not

regions.

Primary education spending by councils is accounted for under different budget heads. These can be

classified into three broad categories; personnel emoluments (salaries and allowances for teachers

and other staff working in a school), non-wage spending on various goods and services provided as

inputs to schools (such as teaching materials, school construction, maintenance and rehabilitation,

equipments, etc.) as well as cash transfers to schools in the form of capitation grant and capital

development grants. The latter is transferred to school accounts with the Tanzania Micro Finance

Bank from which schools make payments for inputs according to specific guidelines for the use of

the funds (eligible expenditures).

Table 16 - Allocation of education grants to councils (TSh per student)

Councils Allocation Std Dev Min Max Dar es Salaam 70,887 4,272 66,208 74,578

Other urban 84,342 26,564 13,485 157,682

Rural 69,801 22,266 13,182 187,696

The allocation of education grants , both recurrent and development, varies considerably between

Councils with allocations between 13 000 and 188 00 TSh per student enrolled as illustrated in table

16. The main factor determining level of allocations to a council is the number of teachers on the

payroll. Councils with low P/T ratios are allocated higher levels education grants per student, i.e.

they have more teachers per student than other councils.

Allocations for Personnel Emoluments (PE) constitute the major share of the Education Block Grant

(87%) and total education grants (including capitation grant and education capital development

grants). Allocations are made on the basis of the number of employees already on the payroll in the

council. However, councils with lower P/T ratios are allocated additional amounts to increase their

employment of teachers to gradually achieve national targets for P/T.

: Page 66

Figure 9 - Council allocation of education grants by pupil/teacher ratio (TSh per student)

Since the figures also include allocations of development grants, it could be assumed that the

variations are to some extent linked to enrolment rates, i.e. councils with high P/T rates are awarded

a higher share of development grants to increase investments in new classrooms and school

buildings to reduce size of classes and employ more teachers. However, analysis of data does not

suggest that level of capital development grants allocated to councils are linked to their level of

enrolment nor P/T ratios but were allocated according to other criteria.

7.3.2 LEVEL OF EXPENDITURE BETWEEN COUNCILS

Education sector expenditures constitute the major share of council level expenditures with higher

share for rural councils than urban councils. The expenditure per capita in Dar es Salaam includes

the Dar es Salaam City Council which cater for a number of general services (but not education

which is accounted for under the three district councils within Dar es Salaam). Dar es Salaam and

other urban areas have a lower school age population compared to its total population than rural

councils.

Table 17 - District level education expenditure and total expenditure (per capita in TSh)

Councils Total expenditure Total education expenditure Education share of total

Dar es Salaam 27,308 10,910 42%

Other urban 35,793 16,275 45%

Rural 25,647 12,551 49%

As illustrated in table 18, while Dar es Salaam displays a lower expenditure per capita and lower

share of education expenditure for the above mentioned reasons, other urban councils have more

public resources in general, and for education in particular, than rural councils despite that the latter

allocate a higher share of their spending for education. Accordingly, councils with more resources in

general per capita also display higher level of spending per capita on education. In general urban

councils have more resources per capita and thus the level of education spending is also higher

despite that the share of spending for education is less than for rural councils.

-

20 000

40 000

60 000

80 000

100 000

120 000

140 000

160 000

180 000

200 000

- 20,0 40,0 60,0 80,0 100,0 120,0 140,0

Council P/T rate

Co

un

cil a

lloca

tio

n p

er s

tud

ent

in T

sh

: Page 67

Council expenditures per student enrolled display significant variations between councils. The

variations are linked to level of grant allocations and capacity to execute grants allocated (ref.

sections above).

Table 18 - Education expenditure councils (TSh per student)

Councils Expenditure Std Dev Min Max Dar es Salaam 72,850 26,264 52,906 102,608

Other urban 91,636 47,759 39,908 256,373

Rural 60,294 22,742 15,290 155,112

Some councils are allocated a higher level of grants per student since they from the outset have a

higher P/T ratio, i.e. an attempt to give them needed resources to employ more teachers and thus

reduce the P/T ratio. However, many of them display a lower level of capacity in execution of these

grants than the average for all councils. Despite they are given the opportunity to among others

employ more teachers by being allocated more than average grant per student, they are not able to

utilize the grants in full while Dar es Salaam and in particular other urban councils spend more that

the amount initially allocated.

Figure 10 displays level of budget execution and P/T ratio per council. Many of the councils with high

P/T ratios show a lower level of budget execution. They are to be found among rural councils. These

rural councils are not able to utilise the resources allocated and thus reduce the P/T ratio since they

are among others not able to employ the additional teachers budgeted for.

Figure 10 - Council level of budget execution by pupil/teacher ratio

This observation is further confirmed when assessing actual expenditure per student and level of

budget execution. Councils with high level of expenditure per student display a higher level of

budget execution (percent of education budget spent).

Councils with high P/T ratios, low levels of expenditure per student and low levels of budget

execution are found in all regions. They are rural councils among the 34 councils identified with

lower overall expenditure per student than initially allocated through the budget.

While some councils received less than 50% of the amount budgeted due to low level of execution,

some received 30% more than of the amount budgeted i.e. during the fiscal year a reallocation of

-

0,2

0,4

0,6

0,8

1,0

1,2

1,4

1,6

- 20,0 40,0 60,0 80,0 100,0 120,0 140,0

Leve

l of b

ugde

tex

ecut

ion

P/T ratio

: Page 68

the Education Block Grants took place between councils due to different capacities in executing the

amounts allocated.

Figure 11 - Council level of budget execution by expenditure per student (in TSh)

The in-year reallocations weakens the budget as a management tool for actual resource allocation

and the various formula based and discretionary allocations initially made have limited impact on

execution in many councils. This reallocation does not level out the initial unequal distribution of

total resource on per capita basis, nor per primary school student. It means that attempts to level

out resource allocation between councils during the budget process had limited impact on budget

execution.

The level of council recurrent budget execution is linked to a number of factors of which the

following are the main ones;

Councils have different levels of capacity in executing their budget. Since release of grants is

linked to progress in expenditures, actual level of budget execution will determine level of

releases. Councils with lower levels of budget execution also receive less of the amount

initially allocated compared to others. Councils which display high level of execution and

even above the level initially allocated receive additional grants within the budget year.

Attempts to equalise actual spending per student across councils is not a question of

changing initial allocations per council since councils with high P/T ratios are not able to

effectively utilize the additional resource. It requires additional technical assistance and

supervision to assist them in implementing the resources allocated.

Education personnel emoluments (predominantly teachers' salaries) constitute 87.2% of the

Education Block Grant calculated on the basis of approved positions in the councils. From

the outset the main explanation for the variation in allocation of block grants is linked to the

variations in allocation of teachers. Movement of teachers across district boundaries

changes the actual release of a major share of the block grant which results in some councils

receiving far less than the initial grant allocation while others receive more than the grant

initially allocated i.e. actual spending is determined by the ability to employ teachers not the

level of initial allocation.

As will be discussed further in sections below, this is linked to the issue of allocation of

teachers. For equalisation of allocation between councils to be actually implemented (raising

levels of spending by rural councils with low level of budget execution), it will require

-

0,2

0,4

0,6

0,8

1,0

1,2

1,4

1,6

1,8

- 50 000 100 000 150 000 200 000

Leve

l of

bugd

etex

ecut

ion

: Page 69

additional efforts in employing and retaining teachers e.g. specific incentives for teachers

serving in rural and remote schools.

In sections below we will be presenting in more detail distribution related to the different education

sector recurrent grants i.e. the salary component of the Education Block Grant, the none salary

component of the Education Block and the additional Capitation Grant and the Capital Development

Grant.

7.3.3 ALLOCATION OF TEACHERS

7.3.3.1 ALLOCATION OF TEACHERS BETWEEN COUNCILS AND SCHOOLS

Allocation of the salary component of the Education Block Grant is determined by the number of

teachers already employed and the number of new positions allocated across districts. The

allocation of new positions is based on national targets for P/T ratio (in 2008 the national target for

primary schools was 45). New teachers entering the market (estimated number of new graduates

from teacher training colleges) are distributed across councils based on their prior year P/T ratio in

an attempt to reduce the P/T ratios for councils with a ratio above the national target.

Councils' allocations for teachers salaries are the main component of the education sector budget.

The total allocation per student is linked to student/teacher ratios31. The student/teacher ratio is on

average higher in rural schools than urban schools and accordingly the grant release per student

enrolled is higher in urban councils than rural councils.

Table 19 - Student/teacher ratio - average by council - MoEVT records

Councils Student/teacher ratio Std deviation Min Max

Dar es Salaam 42 6 36 47

Other urban 43 10 29 78

Rural 60 14 34 121

There are major variations in the student/teacher ratios between councils, in particular between

rural councils as indicated by the standard deviation and min/max data in table 19 above. There are

councils with an average of as many as 121 students per teacher at the one end and at the other end

councils with as few as 29 students per teacher on average between their schools.

The student/teacher ratio between schools in the sample subject for our survey displays a further

variation as illustrated by table 20 below.

31 There is also non teaching staff like guards and other care takers on the council payroll however according to school level data they constitute only 2% of the total number on the payroll.

: Page 70

Table 20 - Student/teacher ratio - average by schools - survey data

Councils Student/teacher ratio Std deviation Min Max

Dar es Salaam 41 16 14 71

Other urban 41 11 10 61

Rural 59 31 11 283

While the average in our sample suggests lower P/T ratio than the national records, it also displays

the significant variations between schools as reflected by the standard deviation and min/max

figures in our sample. It confirms that teacher allocations remain favourable to urban councils and

schools. It also confirms the observation that employment of teachers is not just an issue of

allocating sufficient PE ceilings to councils but also a question of council allocation of teachers to

schools.

7.3.3.2 ALLOCATION OF NEW TEACHERS

In 2007 MoEVT estimated the number of new teacher graduates that could be deployed to schools

in 2008. The allocation of these teachers to councils was made relative to the council P/T ratio for

the school year 2007. Councils with less than 45 students per teacher were not awarded any new

teachers among the newly graduates. The other councils were allocated a share of the newly

graduates relative to their P/T ratio. This was done as an attempt to level out the inequality in P/T

ratios between councils. When comparing the planned allocation with the actual change in teachers

in the councils in our sample, the allocation appears to have had limited impact.

Table 21 - Planned and actual allocation of new teachers 2008 - MoEVT and survey data

Councils Planned allocation of new graduated teachers 2008

Actual change in teaching staff according to survey data

Dar es Salaam 182 441

Other urban 15 20

Rural 1,271 444

The net impact of the allocation was less than the number of newly graduates allocated to the

councils as illustrated in table 21. Of the 1,468 new graduates allocated, the net increase in number

of teachers was 905 equivalent to 61% of the planned allocation. Furthermore, the net increase was

significantly higher than planned in Dar es Salaam and significantly lower in rural districts . For many

of the schools in the rural councils the recruitment of new teachers were not sufficient to maintain

the P/T ratios from the previous year.

The above further explains that even if rural councils are allocated grants (the PE component of the

Education Block Grant) to employ an increasing number of teachers to achieve a lower P/T ratio like

their urban counterparts, they are not able to i.e. the majority of the newly graduated teachers do

not take up positions allocated in rural schools. The net result is low level of budget execution i.e.

low level of spending compared to the approved budget and thus rural councils and schools are not

able to utilise the central government grant allocation as intended. The opposite is the case for

urban councils. They employ more teachers than initially allocated and thus require additional

resources to pay them and subsequently actual releases of education grants exceeds what they were

initially allocated.

: Page 71

The above confirms that the key to promote equality in resource levels for primary education

between councils and schools are closely linked to ability of councils and schools to employ and

retain teachers. Budgeting for new teachers and/or changing grant allocations between councils

does not change inequality in actual execution unless specific measures are applied to enable rural

and remote councils and schools to actually employ new teacher allocated and retain them i.e. like

the introduction of incentive schemes for teachers serving in rural and remote councils and schools

as mentioned in sections above.

7.3.3.3 COST OF TEACHERS

Average cost per teacher at schools varies with relative higher expenditure per teacher for urban

councils than rural councils. In table 22 the average cost per employee on the payroll at a school is

displayed. The variations in costs per teacher are due to several factors. One is that some of the

lower paid non teaching staff is included and not all schools have additional none teaching staff on

the payroll. Another is that in some rural schools teachers only work part time but still receive basic

salaries as teachers. The most significant cost however is related to teachers on payroll that was not

attending the schools. For these schools the total cost per teacher then becomes significant as

illustrated by a cost per teacher as high as 6.6 million TSh in one rural school (maximum figure).

Table 22 - Cost per teacher according to PMO-RALG and MoEVT records (in TSh)

Councils Cost per teacher Std deviation Min Max

Dar es Salaam 2,885,772 695,030 1,594,480 3,089,259

Other urban 2,537,396 758,243 827,735 4,813,775

Rural 2,415,674 1,176,553 1,307,927 6,624,149

The same distribution of costs between councils is confirmed in our sample of schools although it

displays generally a lower level of cost per teacher. The difference can be attributed to sample

errors. Since the difference between national records and sample is systematic across all council

averages then it is likely first and foremost due to differences in council payroll data and teachers on

the payroll actually allocated to a school.

Table 23 - Cost per teacher according to survey data (in TSh)

Councils Cost per teacher Std deviation Min Max

Dar es Salaam 2,231,612 478,102 1,209,314 2,697,280

Other urban 2,160,334 459,018 985,013 2,791,846

Rural 1,920,748 521,880 1,243,272 4,254,552

The allocation of teaching and non teaching staff by level of education at school level is displayed in

table 24. Teachers in urban schools have generally a higher level of education than in rural schools. A

majority of the teachers are licensed teachers with grade A but as many as 15% of teachers on the

payroll in rural schools have lower grades. This may serve to explain the higher average expenditure

per teacher on the payroll in urban schools observed above.

: Page 72

Table 24 - Teachers by level of education according to survey data

Councils Degree Diploma Grade A Grade B Grade C Non teaching Total

Dar es Salaam 0.4 % 7.7 % 90.2 % 1.4 % 0.1 % 0.2 % 100.0 %

Other urban 0.3 % 3.5 % 92.5 % 1.8 % 1.4 % 0.6 % 100.0 %

Rural 0.0 % 0.8 % 84.4 % 11.0 % 3.3 % 0.5 % 100.0 %

Total 0.2 % 3.1 % 87.4 % 6.8 % 2.1 % 0.4 % 100.0 %

It further serves to explain also the variations in costs per student between councils i.e. urban

schools have a higher number of teachers per school and they have higher qualifications while many

rural schools display higher P/T ratio from and have lower qualified teachers.

7.3.3.4 TEACHER EMPLOYMENT DETERMINES RESOURCES FOR PRIMARY SCHOOLS

Allocation of teachers is a main determinant for resource allocation between councils and schools.

As long as existing staff positions are maintained and new teachers' entries are not effectively

distributed to change the allocations of teachers; both in terms of numbers and qualifications, there

will be limited prospects for changes in allocations between councils and schools. The former is a

question of approach to allocation of teachers between councils by the Central Government (PE

ceilings). The latter is a question of intra council allocation of teachers between schools and their

opportunity to attract teachers in rural and remote areas.

After conducting a stepwise regression testing the combination of different variables to explain the

variations in P/T ratios, the distance from urban centres for rural schools appeared to be the variable

most significantly correlated to the P/T ratio (ref. figure 12 below).

This confirm observations presented in previous sections i.e. that changing allocations and assisting

councils with high P/T ratios will in addition require specific incentives for teachers in schools located

in remote places with the level of remoteness measured by their distance to urban centres. These

incentives could be directed at teachers to take up positions allocated by higher PE ceilings but

today not executed since councils and schools are not able to employ them or retain teachers

employed. The incentives could be in the form of a diversified enumeration package for teachers

linked to the schools' location.

Figure 12 - Rural schools - P/T ratio and distance to council headquarter - survey data

-

50

100

150

200

250

300

- 100 200 300 400 Distance urban centre

P/T

rat

io

: Page 73

The incentive would require a two-step approach similar to other countries practising diversified

remuneration packages for the same positions. The first step would be to identify councils with low

level of P/T. The second to ensure that these incentives are applied by councils to equalise P/T ratios

among the schools within the councils.

7.3.4 EDUCATION GRANTS FOR NON-WAGE INPUTS

The amount of non-wage inputs to schools is to a large extent determined by the amounts of

capitation and development grants transferred to schools from the councils. In addition some

schools also mobilise funds and inputs in kind from parents, teachers, Non Governmental

Organisations (NGOs) and from other contributors. In this section we will focus on cash and in kind

contributions from councils to schools using school data from the survey. Private contributions

(including NGOs) will be discussed further in the sections below.

There are two main grants transferred to schools from the councils; Capitation Grant and Capital

Development Grants. In total these grants constitute 99.8% of all government grants to the schools

in the sample32. The former is to support non-wage inputs to schools, the latter for school

infrastructure improvements. These grants constitute 9% of average school resource inputs.

7.3.4.1 TRANSFER OF CAPITATION GRANTS TO SCHOOLS

Capitation grants are on average equally distributed among schools in a council, however in some

councils there are major exceptions with significant allocations for some individual schools while

others receive very low levels of capitation grants per student. Average capitation grant per student

received by the schools in our sample is 4,189 TSh per student (3.3 USD)33 which is below the

indicative levels presented in the budget guidelines provided to LGAs for FY2008 when combining

the capitation component of the Education Block Grant and the additional formula based capitation

grant by MoFEA/PMO RALG34.

32 The remainder are grants from the Tanzania Social Action Fund (TASAF) which constitute only a minor share of grants to the primary schools in the sample and was provided to only a few schools.

33 In the PETS 2004 covering the calendar years 2002 and 2003 the average capitation grant per student was estimated at 5,800 TSh in nominal terms (equivalent to 5.6 USD per student).

34 As previously mentioned the 'capitation grant' intended for schools consists of two parts; the balance of the Education Block Grant after a council has accommodated own expenditures and an additional formula based capitation grant. According to the budget guidelines the former was to be a minimum of 3,000 TSh per student (or 6,000 Tsh per student pending which section in the guideline is to be followed) and the latter was allocated with 5,000 Tsh per student enrolled according to the Swahili version of the guidelines and as per MoFEA data (although the English version of the guidelines stated they were allocated on the basis of estimated school age population in a council).

: Page 74

Table 25 - Average Capitation Grant per student to according to survey data ( in TSh)

Councils Average per student Std deviation Min Max

Dar es Salaam 3,736 1,593 1,805 10,104

Other urban 4,192 2,005 2,083 11,765

Rural 4,251 2,498 1,260 19,236

In 18 of the 21 councils in our sample the level of Capitation Grant per school is linked to size of

school and P/T ratio, i.e. rural schools with a relatively low number of teachers per student are

compensated with higher than average level of capitation grants.

As mentioned above P/T levels and teacher inputs are linked to location of the schools. It means that

for these councils the rural schools with higher P/T ratios than the average and distant location from

the council headquarters receive more in capitation grants per student. However, there are some

major exceptions; 4% of the schools in the sample received exceptionally high levels of capitation

grants per student (above 10,000 TSh per student) compared to the average. These schools are to be

found in one town council and four district councils. These councils also display extremely uneven

allocations of grants between schools.

The capitation grants to schools represent approximately 93% percent of public funding for school

teaching materials and other non-wage inputs as well as contributions to administrative expenses.

The remaining 7% of council contributions are in kind because some councils visited did not transfer

the full amount of capitation grants allocated to schools but instead procured text books and other

inputs on behalf of some schools. This was partly to assist schools, particularly in remote areas in

acquiring inputs, partly because some schools for various reasons were considered to have too weak

oversight in managing cash and partly because procurement of school inputs on behalf of several

schools could achieve better economy of scale.

.

Table 26 - Average Capitation Grant and in kind contributions by council to schools per student according to survey data ( in TSh)

Councils Average per

student Std deviation Min Max

Dar es Salaam 4,461 1,354 2,824 10,104

Other urban 4,872 2,116 2,138 14,206

Rural 4,893 2,465 1,360 20,117

When including the value of in kind contributions to schools by councils the total resource per

student for non-wage inputs display less variation for schools in Dar es Salaam as observed in table

26 above which only displays cash contributions. But the variation increases for schools in other

urban areas and rural schools as reflected by the change in standard deviation in table 27 below.

Although for some schools the in kind contributions serve to level out their lower per student

capitation grant in cash, for others they do not. In the latter case the schools are already receiving

larger than average capitation grants and simultaneously receive higher value of in kind

contributions from the council.

: Page 75

7.3.4.2 TIMING OF DISBURSEMENTS OF GRANTS TO SCHOOLS

Timing of release of capitation grants is also an issue for schools if they are to provide text books and

teaching materials in a timely manner within the school year.

Table 27 - Days between date of council transfer and date credited school accounts for the first

Capitation Grant release (average number of days).

Councils Days from council transfer to credit school account

Max days capitation grant

Days of the fiscal year before council transfer

Max days fiscal year

Dar es Salaam 16 24 87 193

Other urban 18 64 195 300

Rural 41 318 167 291

Total 35 318 162 300

Table 27 presents the average number of days between the date according to council records that

the first Capitation Grant release to schools was made and the date it was recorded as received

according to school cash books and bank statements. It also presents the number of days from the

beginning of FY2008 (1 July 2007) to the date the council released the first tranche. Most councils

prepare a payment advice to the bank for distribution of capitation grants from the council bank

account to individual schools covering all schools in the councils, however some councils prepare

specific payment advise for individual schools at different point in time.

For 11% of the schools the date between council records and school records was less than one week.

For 36% of the schools it was less than 2 weeks and for 60% less than 3 weeks. If assuming that there

is a delay from when the council registers its payment advice to the time it is actually submitted to

the bank and then allowing normal banking days for processing of the transfers to individual schools

then it can be claimed that for 60% of the schools they receive the capitation grant within an

acceptable timeframe from the date of the transaction is being processed by a council. For the other

councils/schools it may be claimed to be delays, and in some cases significant delays, either in the

time the council takes in preparing the payment advise, the time in actually submitting the payment

advise to the bank and/or the number of days the bank spends on processing the transfers to the

individual schools.

Since all councils have remaining balances from previous years unspent grants from central

government, they are in most cases not relying on current fiscal year transfers from the central

government to make the first tranche release of the capitation grants to schools. Accordingly, the

time from the beginning of the fiscal year to the time the school is actually credited the capitation

grant gives an indication of council timeliness in budget execution. While councils in Dar es Salaam

on average recorded their capitation grant transfers already during September 2007 in the FY2008,

other urban and rural councils on average recorded their first transfers in January/February 2008 (at

the beginning of the next school year). Some of the councils make several transfers, other only one

or two transfers per year. However, some schools did not receive their FY2008 capitation grant

before the end of the fiscal year (mid-term of the school year).

In summary the observations presented in this section show that the attempts to equalise non-wage

resources through formula based allocations to councils have limited impact on the actual allocation

: Page 76

of capitation grants to schools. Even if the budget guidelines suggest that councils should allocate a

minimum threshold of 3,000 TSh (or 6,000 TSh) per student enrolled to schools, this is not

implemented by the councils and actual transfers to schools vary significantly. Furthermore, several

schools face significant delays in receiving the grants and many receive the first grant allocation

several months into the next school year.

To address the above issues related to uneven distribution of grants and delays in transfers it

suggests a revisit to the approach in allocation and transfers of grants to schools. On option similar

to many other countries could be to allocate 'block grants' for non-wage inputs directly to schools. It

implies that spending decisions is left to the school management. One important element for

ensuring that spending decisions reflect intended use of the resources is an effective decision

making arrangement and oversight function by the beneficiaries; the students represented by their

parents.

There is already a functioning oversight system for primary schools in mainland Tanzania as our data

suggest.

Table 28 - School Management Committees and school publishing receipts/expenditures.

Councils Schools with

SMC SMC that met

Average number of SMC meetings

Schools that publish accounts/receipts

Dar es Salaam 100% 100% 6 100%

Other urban 100% 100% 5 100%

Rural 100% 100% 6 98%

As table 28 illustrates, all schools in the sample had School Management Committees (SMC) and all

of them met at least once during the school year. On average SMCs met six times per year. In only

13% of the schools the SMC met less than four times per year (less than every quarter).

7.3.4.3 TRANSFER OF DEVELOPMENT GRANTS TO SCHOOLS

The capital development grants constitute on average 21% of cash contributions to schools from

councils. They are allocated to schools based on rehabilitation and investments needs. In many

councils some schools receive capital development grants one year with schools changing each year.

This is reflected in table 29 below by the high standard deviation compared to the average. Of the

schools in the sample, 18% received capital development grants while the remaining 82% did not

(the minimum being zero). For the schools receiving the grant the amount can be significant since it

is to cover expenses for desks and chairs as well as construction of classrooms, latrines and other

infrastructure.

: Page 77

Table 29 - Average Capital Development Grant from councils to schools (in TSh per student) -

survey data

Councils Development grant per student Std. Deviation Min Max

Dar es Salaam 624 1,500 - 6,134

Other urban 1,389 4,447 - 24,914

Rural 1,144 4,485 - 38,441

Some councils procure equipment and/or pay contractors directly for construction of classrooms

and other infrastructure. These contributions are predominantly provided to schools that do not

receive capital development grants in cash. The in kind contributions from councils however

constitute only a minor share of total contributions to schools capital expenditure as shown when

comparing the cash contributions presented in table 29 above and total capital development

contributions presented in table 30 below.

Table 30 - Average Capital Development Grant and in kind contributions to schools from councils (in TSh per student) - survey data

Councils Development grant per student Std. Deviation Min Max

Dar es Salaam 624 1,500 - 6,134

Other urban 1,476 4,579 - 25,862

Rural 1,275 4,563 - 38,441

Total 1,231 4,339 - 38,441

While development grant contributions in cash were transferred to 18% of the schools in the sample

the combined cash and in kind contributions were transferred to 24% of the schools in the sample.

7.4 PRIVATE CONTRIBUTIONS TO PRIMARY SCHOOLS

The survey tool for schools also included questions about parents' contributions. The data on

parents' cash and in kind contributions are stemming from school records. In terms of in kind

contributions from parents these are estimates based on head teacher and school management

committee estimated value of the contributions.

In total the private contributions to schools including contributions from NGOs, parents and other

private donations constituted 3.7% of total school cash and in kind contributions to schools. Their

contributions accounted for 28.4% of non-wage resources. Cash and in kind contributions from

parents only constituted 1.0% of the total contributions to schools and 7.7% of non-wage resources.

: Page 78

Figure 13 - Value of cash and in kind contributions to schools excl. cost of teachers on Government payroll - survey data

It is fair to assume that these figures are underestimated since many schools do not keep full records

of such contributions, and in particular for in kind contributions35. Our data shows that in 30% of the

schools parents made contributions in cash to pay for extra classes (extra classes were offered in

63% of the schools), pay for employment of additional teachers in addition to teachers on the

payroll (in 14% of the schools), and to pay for various teaching materials and equipment (like desks,

chairs, etc.) to supplement contributions in cash and in kind from the councils. In 12% of the schools

parents also contributed with in kind contributions, i.e. the same type of inputs procured by the

school with cash contributions but in this case it was the parents that organised collection of funds

and procured the inputs.

As illustrated in figure 13 above, parents contributions constitute a small share of contributions to

schools. The main contributions are the cash contributions from councils in the form of capitation

and development grants (63% of total contributions). Furthermore, the councils also contribute with

in kind contributions by procuring teaching materials and equipment which they supply to the

schools (8% of total contributions). NGOs and other community organisations, private companies

and individuals also make major in-kind contributions to schools (20% of total contributions).

Nevertheless, for some schools and in particular for rural schools parents are in many cases a

significant additional resource for schools to acquire non-wage inputs as well to pay for additional

teachers and others not on the government payroll as reflected in table 31.

35 An attempt was made to consolidate these data with the data from the HBS 2007, however, the HBS 2007 survey did not distinguish between household expenditure related to different types of schools (pre-primary, primary, secondary and tertiary). While an indirect method could have been applied to estimate household expenditure for education using household member age groups to distinguish between different levels of education, it would have made the individual education level strata too small and HBS data would not have enabled adjustment for the fact that some students are enrolled in schools below their age group.

School LGA grants

63 %

Parents cash

contributions2 %

Other cash

contributions2 %

Parents in kind

contributions5 %

LGA in kind

contributions8 %

Other in kind

contributions20 %

: Page 79

Table 31 - Average contribution per school (in TSh) - survey data

Councils Parents cash contributions

Parents in-kind contributions

Contributions others In-kind contributions

council

Urban schools

143,700 57,763 1,569,791 235,809

Rural schools 85,870 312,886 758,082 386,374

Total 99,152 254,289 944,517 351,792

Table 31 shows the value of cash and in kind contributions (excluding payment for teacher on

government payroll) per school with averages for all urban and rural schools. Urban schools receive

more contributions in cash rather than in kind as compared to rural schools where school

management committees/parents appear to a larger extent manage the procurement of the inputs.

Urban schools also enjoy a significantly higher value of NGO and other private contributions than

rural schools. On the other hand, rural schools rely more on parents contributions than urban

schools.

7.5 SCHOOL LEVEL EXPENDITURE

While the sections above have presented analyses of resource allocations and transfers from Central

Government to councils and from councils to schools as well as private contributions, the following

sections will analyse structure of expenditure at school level.

Spending on teachers on government payroll is managed by the councils. Spending on non-wage

inputs is made by the schools using their cash receipts in the form of capitation and development

grants as well as cash contributions from others. In addition they receive in kind contributions from

the councils, parents and others.

Cash contributions managed by schools are used for funding of 67% of school inputs. The other 33%

are contributions in kind from the councils, parents and others. The distribution of expenditure from

these contributions are presented in figure 14.

The main cost item at school level is construction costs which include construction of new

classrooms, administration buildings, staff houses, latrines, etc. In total construction costs accounted

for 30% of school level expenditures . These costs are partly funded by cash contributions to schools

(51%) or paid by councils or other contributors directly to contractors (49%).

The other main items are text books and other teaching materials which constitute 29% of total

school level expenditures36. These cost items are for the majority of schools paid for by cash

contributions to schools (92%). Other teaching tools and inputs like exercise books,

pens/pencils/chalks constitute another 12% of school level expenditure which is paid for entirely by

the schools from cash contributions. Desks and other equipment to schools constitute 6% of school

level expenditure of which 52% is paid for by the school while 48% is provided as in kind

contributions from the councils and others.

36 In addition to text books it includes globes, math kits and slates.

: Page 80

Figure 14 - School level distribution of expenditure by main cost items (averages for all schools) -

survey data

Administration and other expenses constitute 9% of school level expenditure. This covers among

others travel and other allowances to SMCs and teachers. While it proved in some cases that part of

these allowances was for teachers performing extra classes the amount for this service could not be

determined.

Some schools also employ teachers directly in addition to teachers on government payroll and some

keep separate accounts for extra allowances to teachers on government payroll for performing extra

classes. The number of extra classes constitute 19.3% of the total number of classes taught in a week

which further supports the observation that parts of the administrative costs are also paid as

allowances to teachers for performing these extra classes.

The above cost items are presented in more detail in table 32 with distribution of costs per student

for schools in Dar es Salaam, other urban areas and schools located in rural councils.

Table 32 - Average cost per student (in TSh) - survey data

Councils Teachers

Gov. payroll

Employees not on payroll

Text books, teaching

tools

Exercise books, pencils,

etc.

Exam materials

School meals

Constr. Rehab. Maint.

Desk/ chairs/ tables

Admin/ other

Dar es Salaam

- 244 1,870 730 484 198 947 416 394

Other urban

7 367 1,845 772 422 228 2,470 368 385

Rural 102 274 1,860 730 476 62 2,042 351 664

Construction, rehabilitation and maintenance constitute the major cots item for rural schools and

other urban schools. Rural schools also spend relatively more on additional allowances for teachers

on government payroll and administrative expenses which are assumed partly to fund the same type

of expenses.

Teachers on payroll1 % Employees not on

payroll

4 %

Text Books, teaching guidelines teaching tools

29 %

Chalks Exercise Books, pens, pencils, etc.

12 %

Examination papers materials

7 %

School meals 2 %

Construction Rehabilitation Maintenance

30 %

Desk/chairs/tables and repair

6 %

Administration and other expenses

9 %

: Page 81

Combining the above observations with cost of teacher salaries (ref. previous sections) the following

conclusions can be made;

Urban schools spend less on teachers performing extra classes but on average have higher

number of teachers per student on government payroll than rural schools.

Other urban and rural schools spend more resources on upgrading their facilities37. It is in

these councils that there has been a significant increase in investments in classrooms to

accommodate a desired growth of enrolment.

Rural schools spend more on employment contract teachers and allowances to teachers on

government payroll in an attempt to employ and retain teachers but the net result is still a

far lower number of teachers per student than other schools.

The above averages between schools in Dar es Salaam, other urban areas and rural districts do not

display the variations in costs for the different schools. The variations are significant and linked to

level of grants transferred. The average cost per student for text books may serve as an illustration.

Table 33 - Average cost of text books and other teaching materials per student (in TSh) -survey

data

Councils Development grant per student Std. Deviation Min Max

Dar es Salaam 1,870 545 963 3,329

Other urban 1,845 647 758 3,333

Rural 1,860 1,519 163 15,190

In table 33 we have presented the average cost for text books and teaching materials per school in

Dar es Salaam and other urban and rural schools. The average cost per student varies significantly

between schools and most prominently for rural schools (reflected by the high standard deviation).

Among the rural schools 4% spent less than 500 TSh on text books and teaching materials per

student, 13% of the schools spent less than 1000 TSh per student and 31% less than 1,500 TSh per

student. This stands in contrast to urban schools in which 94% of the schools spent more than 1,000

TSh per student with the lowest expenditure per student in a school as displayed in the table above.

Level of expenditure is linked to level of grants to the councils and subsequently from the councils to

the schools i.e. the main source of funding for school expenditures. As the above observations show,

there are schools with very limited resources for non-wage inputs and many of them have

simultaneously few and less qualified teachers than others. Even with their ability to mobilise

contributions from parents these contributions do not compensate for the low level of transfers

from councils.

7.6 SCHOOL LEVEL RESOURCES AND PERFORMANCE

The main indicator for assessing school performance has been number of students performing and

passing Primary School Leaving Examination (PSLE). An attempt to analyse correlation between PSLE

37 This may be due to the fact that the average age of the schools in our sample is highest for the rural schools followed by other urban schools

: Page 82

performance and school inputs has been assessed against different types of inputs including various

school infrastructure indicators.

In the following we present results from analysis of council data for all councils in mainland Tanzania

based on our database compiled from different national records such as NECTA school examination

results, MoEVT EMIS database, PMO-RALG data on expenditures by Councils and NBS poverty data

by council38.

Table 34 - Selected council indicators (averages for each group of councils) - data from official

records

Councils PSLE national ranking per

school P/T

Education exp. per student

Recurrent exp- per student

Rural population

Population classified as

poor

Dar es Salaam 2,874 45 72,850 65,114 6% 18%

Other urban 4,562 44 75,322 70,381 32% 24%

Rural 7,238 64 61,476 57,008 90% 39%

In the above table we present a number of indicators for each group of councils. The PSLE ranking is

based on different score and pass levels for schools according to NECTAs rating of the individual

primary schools. The lower the number the higher the school rank compared to others as measured

by examination results i.e. for a council it may be interpreted as average rank of all schools in the

council. A low number would indicate that the council has a high number of better performing

schools compared to others.

The result suggests that on average urban councils (including Dar es Salaam) have better performing

schools (as measured by their average PSLE rank), they have on average lower P/T ratios, spend

more on recurrent expenditure for schools (wage and non-wage inputs). Urban councils also have a

lower poverty rate than rural councils.

One input which is assumed to be a major factor for school performance is the deployment of

teachers as measured by the P/T ratio. Council averages seem to suggest that deployment of

teachers are correlated with school performance when analysing council averages which also serve

to explain that better performing schools have higher recurrent expenditures per student.

Figure 15 - PSLE ranking and P/T ratio - council averages - data from official records

38 The NBS data are for the year 2007.

-

20

40

60

80

100

120

140

- 2 000 4 000 6 000 8 000 10 000 12 000 14 000

All Councils

PT ratio/PSLE ranking

-

20

40

60

80

100

120

140

- 2 000 4 000 6 000 8 000 10 000 12 000 14 000

Rural Councils

PT ratio/PSLE ranking

: Page 83

This is illustrated in figure 15 above, lower P/T ratio (and thus higher PE spending per student)

generally gives a better average rank of PSLE pass rate for schools in the council. As mentioned in

the sections above however, the specific indicators related to education vary significantly between

rural councils as well as between schools within rural councils e.g. the correlation between one

indicator like P/T ratio (spending on teacher salaries per student) does not alone serve to explain the

level of PSLE ranking.

While spending on non-wage inputs appears to influence school performance measured by average

school PSLE ranking in a council, again there are significant variations, first and foremost among rural

schools as illustrated by figure 16 below.

Figure 16 - PSLE ranking and council non-wage spending for education per student (in TSh)

Another dimension typically referred to in many analyses is poverty rate. Level of service delivery

and service delivery outcomes are assumed to be influenced by levels of poverty i.e. that school

performance is correlated with the general level of welfare in a council. This is the case when

assessing the urban versus the rural councils, however communities within the rural councils are

highly diversified in terms of poverty levels and so is the quality of schools, P/T ratios and overall

resource inputs in financial terms and these do not appear to be inter-correlated. This is illustrated

by figure 17 with the large variations in average school PSLE ranking and council aggregate poverty

level.

Figure 17 - Average school PSLE ranking and council poverty rates - rural councils

We do not have access to more disaggregated data on poverty rates and other indicators within

councils and hence no opportunity to test these data related to communities (wards, village level)

within councils. However, the data from the sample survey allow more in-depth analysis of school

performance along a number of dimensions.

-

2 000

4 000

6 000

8 000

10 000

12 000

14 000

- 20 000 40 000 60 000 80 000 100 000 120 000

All Councils

PSLE rank/OC exp per student

-

2 000

4 000

6 000

8 000

10 000

12 000

14 000

- 20 000 40 000 60 000 80 000 100 000

Rural Councils

PSLE ranking/OC expenditure per student

-

2 000

4 000

6 000

8 000

10 000

12 000

14 000

- 0,10 0,20 0,30 0,40 0,50 0,60 0,70

All Councils

PSLE rank/ poverty rate

-

2 000

4 000

6 000

8 000

10 000

12 000

14 000

- 0,10 0,20 0,30 0,40 0,50 0,60 0,70

Rural Councils

PSLE ranking/Poverty rate

: Page 84

The survey included collection of information of school physical facilities. These facilities included

number of classrooms, teacher staff houses, playgrounds, number of student desks and chairs,

teacher tables, blackboard, access to drinking water, electricity, type and number of latrines, etc. All

these physical facilities could be analysed separately but analysis for each infrastructure component

individually maintains a large unexplained variance39. A stepwise regression introducing the selected

infrastructure indicators reduced the unexplained variance significantly.

Empirical evidence suggests that many of the above inputs have different relative impact on learning

environment and different in different countries and regions within a country40. In some countries

national standards are applied in relation to all these indicators. The indexes are usually designed

based on analysis of the likely relative impact of each component on performance to measure school

quality through a composite index. In some countries this also includes qualifications of headmasters

and other teaching staff as well as the more conventional indicators such as P/T ratio.

We have grouped the schools according to their combined availability of all facilities by designing a

simple index to express 'school quality'. In terms of school level facilities like electricity and water

we have applied a binary rating. In terms of other infrastructure inputs like desks, classrooms,

latrines, etc. we have ranked the schools according to their score along each dimension and added

the score for each dimension as an overall index of school physical facilities. It has been beyond the

scope of this survey to develop or suggest the content of a more accurate index for mainland

Tanzania through multiple regression hence the application of a simple 'school facility index'.

In the following we will present the results from the sample allowing a more in-depth analysis of

what determines school performance, among others using the above index. In this case we use PSLE

student pass rates as the performance indicator and have included the data on infrastructure

facilities and school location (distance to council headquarters) in addition to the ones presented for

council averages.

Table 35 - Selected school indicators (averages for councils) - sample survey data

Councils PSLE pass rate P/T ratio (teachers attending school)

School exp. per student

Rank school facilities

Dar es Salaam 0.74 40 58,176 114

Other urban 0.67 44 59,870 117

Rural 0.53 62 49,394 147

Judging from the above table urban schools generally show a higher performance as measured by

pass rates for PSLE, they have, as also discussed in sections above, on average significantly lower P/T

ratios. They spend more money per student (due to higher P/T ratios) and facilities are generally of

higher quality as measured by availability of equipment per student and different type of facilities at

39 Using coefficient of determination in regression showing variance between variables included and not in the regression model.

40 Ref. Oliveira, J. and J. Farrell. 1993: Teacher Costs and Teacher Effectiveness in Developing Countries and Schneider, M. 2002: School Facilities and Academic Outcomes. National Clearinghouse for Educational Facilities. Washington DC. 2002.The Political Economy of Educational Reforms and Capacity Development in Southeast Asia, Springer Netherlands, 2009.

: Page 85

the schools (the average ranking of school facilities). For rural schools the aggregates hide

significant variations between them i.e. the P/T ratio, level of wage and non-wage spending as well

as 'quality of facilities' do not serve to explain variations in school performance across rural councils.

The overall factor to explain variations between schools in rural councils appears to be its location;

measured by its distance to council headquarters. In table 38 below we have grouped the rural

schools according to their location.

Table 36 - Selected school indicators by school distance to council headquarters - survey data

School distance to council HQ

Percent passed PSLE

P/T ratio Council exp. per

student School exp. per

student Rank

facilities

Rank facilities

excl. staff houses

<20 km 0.58 50 57,226 5,604 146 147

20-50km 0.51 66 45,324 4,056 147 148

50-100 km 0.50 60 49,674 4,710 142 135

>100 km 0.50 71 47,929 7,201 175 179

The above table serves to illustrate that the rural/urban dimension within a rural council have impact

on overall school level performance and also on P/T levels. The most remote schools (more than 100

km to council headquarters) have less resource inputs in terms of teachers and lower quality of

facilities. Even if they spend more money on non-wage items per student than other schools it

appears not to compensate for the lower P/T ratio and influence the performance rate measured by

PSLE pass rate. Schools located within 20 km of the council headquarters are allocated more

resources per student because they have higher P/T ratios and have better school facilities as

measured by the average ranking of the schools.

The above observations may serve to support a recommendation that teacher allocations to rural

and 'remote' schools is an issue that needs to be addressed (ref. sections above on teacher

allocations). In the table we have also included ranking of facilities without staff houses. It shows

that for the remote schools staff is a major input for their ranking according to the school facility

index (by the change in ranking). For schools within the proximity of 50–100 km of the council

headquarters houses are not a major component of their index (since they then have better ranking

of facilities than if staff houses were included). Despite this they have lower P/T ratios and PSLE pass

rates than remote schools. The analysis suggests that the focus on infrastructure for teachers is not

alone sufficient to attract teachers to rural schools i.e. other factors may be equally important to

attract teachers to rural and remotes schools.

7.7 DISTRIBUTION AND PERFORMANCE RELATED TO GENDER

In this section we will present results from analysis disaggregated by gender with focus on school

performance related to enrolment of girls and employment of female teachers.

7.7.1 ENROLMENT OF GIRLS

The MoEVT data on girls' enrolment suggest that there is a fairly equal enrolment of boys and girls as

illustrated in table 37 below. Among rural councils however there were councils with lower share of

: Page 86

girls enrolled as evident by lower average and the enrolment rate at 40% for enrolment rates for two

rural councils.

Table 37 - Percent girls enrolled in schools (council averages from national records) - council data from official records

Councils Girls enrolled Std Dev Min Max

Dar es Salaam 50 % 1 % 50 % 51 %

Urban 50 % 1 % 49 % 51 %

Rural 49 % 2 % 40 % 52 %

Total 49 % 2 % 40 % 52 %

Average school performance in the councils measured by pass rates is not correlated with girls'

enrolment and pass rates are equally distributed between boys and girls. The same is the case for

P/T ratios and allocation of resources per student. Level of effort and school performance is not

correlated with enrolment of girls when assessing aggregate council data.

On the other hand, the data from the sample of schools reveals that there are significant variations

between schools in the councils, in particular among schools in rural councils. In rural councils there

are schools with girls enrolled as few as 25% of total enrolled while at the other end schools with as

many as 75% enrolment of girls.

Table 38 - Percent girls enrolled in schools (school averages) - survey data

Councils Girls enrolled Std Dev Min Max

Dar es Salaam 49% 4% 32% 53%

Other urban 50% 3% 41% 58%

Rural 49% 5% 25% 75%

Total 49% 5% 25% 75%

Similar to council aggregates, data on school performance in the sample measured by pass rates is

not correlated with girls' enrolment and pass rates are equally distributed between boys and girls.

Also as suggested by the council aggregates, P/T ratios and allocation of resources per student is not

correlated with enrolment of girls in schools. However, as discussed in sections above, the remote

and rural schools have the highest P/T ratios, the lowest scores on school facilities and performance

rates (measured by pass rates) and it is among these schools the lowest enrolment rates of girls are

to be found (with as low as 25% girls of total students enrolled). It means that a specific focus on

these schools will also have impact not only on primary education in general but on girls education

in particular. These schools are to be found in all rural councils of all regions in our sample.

7.7.2 TEACHERS

The gender equation related to teachers has been assessed in relation to school performance and

resource allocations. Since teacher absence is the major source of efficiency losses in primary

education we have included specific analysis on gender related to this issue.

Urban schools on average have a higher number of female teachers to total number of teachers. As

illustrated in table 39 below, there is a significant gender bias towards female teachers in our sample

in urban areas (77%) compared to the average of rural schools (36%). The gender distribution of

: Page 87

teachers however varies between schools with some schools that only have female teachers while in

rural councils some schools only have male teachers.

Table 39 - Teachers by Gender in primary schools (average by schools in a council) - survey data

Councils Female teachers Std Dev Min Max

Dar es Salaam 77% 22% 11% 100%

Urban 60% 22% 14% 96%

Rural 36% 26% 0% 100%

Total 44% 29% 0% 100%

Cost per teacher is not linked to gender distribution, i.e. there is no evidence in the data to suggest

that level of enumeration is linked to gender distribution of teachers. There is no evidence to

suggest that other inputs and school performance are linked to distribution of teachers by gender

when assessing gender distribution among teachers across schools in all councils. However, the

gender distribution is an issue related to location of schools in rural councils. As illustrated in table

40, rural and remote schools have high P/T ratios and simultaneously a low share of female teachers

compared to others.

Table 40 - Teachers by Gender in primary schools and P/T ratio (average by schools in a council and school location) - survey data

Councils Female teachers P/T ratio

Dar es Salaam 77% 40

Urban 60% 43

Rural schools

<20 km 52% 49

20-50km 36% 66

50-100 km 25% 60

>100 km 27% 70

Combining these findings with findings of school performance, it is evident that location of schools in

rural councils determines ability to employ teachers, and in particular female teachers. In total this

has impact on school performance as measured by lower PSLE pass rates despite that they do not

necessarily have less resources for non-wage inputs.

: Page 88

8 SURVEY RESULTS SECONDARY EDUCATION

8.1 NATIONAL RECORDS AND SURVEY DATA FOR SECONDARY EDUCATION

To analyse allocations and expenditure per teacher and student it was important to establish the

number of teachers and schools in the sample. The official records by MoEVT of students and

teachers, among others for allocation of grants, deviate from the records presented by the Regional

Education Officers (REOs) in the sample regions. Furthermore, these figures deviate from the records

at school level. The records from the REOs present 9.1% more students than what has been

recorded as enrolled by schools in our sample. On the other hand the REOs have recorded 90.5% of

the teachers recorded by the survey tools at the schools (teachers on government payroll). When

presenting results from national records with the data from the sample part of the deviations found

can be explained by these statistical errors in national and regional records. This has been taking into

account in the analyses.

Some of the national records providing financial data proved to be incomplete in order to use them

for detailed analysis. The regional sub-treasury data contain records of financial transactions at the

school levels by expenditure category and this would have been the most reliable basis for capture

of regional financial data. However, for FY2008 it proved to be incomplete in that not many schools

had been captured in the FMIS database and in a format required for analytical purposes. Financial

data are not collected, reported and analysed in such a way to allow the focus to be at the school

level for both wage and non-wage expenditures.

As for payroll records obtained from MoFEA/MoEVT, the recording of payroll data is by pay station.

In terms of secondary education some selected secondary schools acts as pay stations and makes

payments of salaries on behalf of several schools i.e. one school is pay station for a number of other

schools. The regional sub-treasuries do not have data on personnel emoluments. The only source for

tracking PE payments to an individual school is the secondary school acting on behalf of other

schools as pay station. The only option for tracking these resources would then have been to collect

data from either each individual teacher or pay station and compare the amount received every

month (or for selected months) with the amount recorded as transferred by MoFEA for each

individual teacher, a task far beyond the scope of and resources allocated to this survey.

In the following we present the results from analysis of national records available and survey data of

secondary schools. For expenditure data the focus will be on 'non-wage' grants and expenditures for

reasons as explained above. The schools have been grouped according to their district locations (like

for primary schools) i.e. Dar es Salaam, other urban schools (located in regional capitals) and rural

schools (all other schools). In the case of secondary schools initial analyses also indicated that type

of school 'ownership' is also a major determinant for the structure and level of resource allocation.

Accordingly, the results have also been presented for community and government owned schools

respectively.

Results showing totals for all schools based on survey data are weighted averages in accordance with

the sample's share of total schools in mainland Tanzania.

: Page 89

8.2 RESOURCES FOR SECONDARY EDUCATION

The State Budget allocation for secondary education under the budget heads of MoEVT was TSh

170.2 bln in FY2008. With 1,128,711 students enrolled in secondary schools the equivalent per

student allocation was TSh 150,792 (118.7 USD).

The MoEVT budget for secondary education had 20 specific sub vote items under which specific

amounts were allocated. These items can be summarised into five expenditure categories; Personnel

Emoluments (TSh 79.1 bln), , Teaching Materials (TSh 6.1 bln), Transfers (TSh 20.4 bln), School Meals

(TSh 15.0 bln) and (other) Non-wage spending (TSh 49.6 bln).

Table 41 - State budget allocations and releases for secondary education 2007/8 (TSh millions)

Expenditure item Approved estimates

Percent Funds released Percent

Personnel emoluments 79,125 46% 79,122 51%

Teaching materials 6,068 4% 3,426 2%

Transfers 20,370 12% 18,070 12%

Meals 14,996 9% 10,967 7%

Other non-wage spending 49,642 29% 44,866 29%

Total 170,201 100% 156,451 100%

The actual release of funds for secondary education was 156.5 bln TSh (91.9% of the budget was

executed) equivalent to 138,610 TSh per student (109.1 USD).

Secondary school heads receive finances for their operations through Regional Sub-Treasury Offices

on authority of a Warrant of Funds. These funds are disbursed upon request by the Head of School

directly from the sub-treasury’s offices. Secondary schools also generated own funds through fees

and parents' contributions. The schools implemented minor business operations in the name of 'Self

Reliance Projects' from which they generated additional income to the schools. Schools were also

eligible to receive public funds through other institutions with different financing arrangements,

such as councils or TASAF.

Table 42 - State budget releases and expenditures for secondary education 2007/8 (TSh millions)

Expenditure item Funds released Expenditure Allocation/ expenditure

Personnel emoluments 79,122 79,093 100.0%

Teaching materials 3,426 3,425 100.0%

Transfers 18,070 18,071 100.0%

Meals 10,967 13,320 121.5%

Other non-wage spending 44,866 45,220 100.8%

Total 156,451 159,130 101.7%

The total amount spent for secondary education according to MoFEA records was TSh 159.1 bln

equivalent to a central budget execution of 93.5%. This is equivalent to TSh 140,983 per student

enrolled (111.0 USD). The higher level of spending than Central Government releases is due to

revenues generated by secondary schools (fees) and additional funds from councils spend on

secondary education.

: Page 90

The non-wage spending constituted 50.3% of total expenditures; equivalent to 70,909 TSh per

student (55.8 USD).

8.3 SCHOOL LEVEL RECEIPTS AND EXPENDITURES

8.3.1 MAIN SOURCES OF REVENUE FOR SCHOOLS

In the following we present data on cash contributions to schools in total from which they pay for

expenditures other than for personnel on the government payroll. In the following sections we will

present cash contributions before presenting structure of expenditures at school level for all these

resource inputs in total.

Figure 18 - Cash contributions to secondary schools – survey data

The distribution of cash contributions by source is presented in figure 18 excluding personnel on

government payroll for secondary schools. The main source of funding is MoEVT who contributed

with 75.1% of all cash receipts for the schools. Parent contributions in the form of various fees and

other cash contributions accounted for 18.7% with the balance made up of smaller contributions

from councils, NGO's and other private donations.

The cash transfers from MoEVT by type of school and location is presented in table 43 below.

Table 43 - Secondary schools cash receipts from MoEVT per student (in TSh) – survey data

Schools Average cash contributions MoEVT Std Dev Min Max

Community 66,589 65,416 5,906 266,667

Government 28,293 2,149 26,774 29,813

Schools Average cash contributions MoEVT Std Dev Min Max

Dar es Salaam 28,896 23,152 5,906 82,485

Other Urban 56,172 57,736 14,533 253,276

Rural 79,211 71,303 12,346 266,667

Total 65,373 64,708 5,906 266,667

Community schools have been prioritised compared to government schools as concerns cash

contributions from MoEVT. Since there is a higher share of community compared to government

schools in rural areas, rural secondary schools received a significant higher per student cash

contribution than urban schools. The contributions include both formula based (per student)

MoEVT75,1 %

Councils1,9 %

RAS0,2 %

Parents18,7 %

From others4,2 %

: Page 91

transfers as well as discretionary transfers for school improvement of physical facilities and

equipment. Accordingly, the variation between schools are significant in terms of receipts, in

particular for community schools of which many have been established in the last few years and

some opened before completion of all infrastructure facilities (classrooms and other facilities were

still being constructed for several of the schools in 2007 and 2008)41.

Different schools were allocated different types of grants among the 14 types of formula based

grants with fixed amounts per student or school. Many are allocated to the schools as development

(infrastructure) grants for upgrading of facilities (new classrooms, administration buildings,

laboratories, etc.) with a fixed amount per school. The two main grants allocated on a per student

basis are school fee subsidy and learning grant. The others are allocated on a prorate basis to

schools based on estimated costs of facilities (like cost per number of classrooms to be constructed,

cost of administration building, library, laboratory, staff houses, student hostels for boarding

schools, etc.).

The difference in allocation criteria and formulas for all these different grants serves to explain the

variance in grants per student to individual schools. While there were significant variations in

transfers of capitation grants to primary schools and significant deviations from the levels indicated

in budget guidelines, the majority of the secondary schools received the grants initially allocated

with some exceptions. During the year a reallocation was been made between some secondary

schools, among others because some schools eligible for grants were initially not included in the

allocation42. These variations between allocations and actual releases are displayed in figure 19.

Figure 19 - Grants received by secondary schools in percent of initial MoEVT allocations - survey

data

A more detailed account of the types of grants allocated and actual receipts will be presented in the

sections below.

41 When comparing the average contributions in our sample with the average of non-wage expenditure under the secondary education vote of MoEVT the deviations can partly be explained by that MoEVT uses part of the budget for non-school expenditures (administration and other expenditures). On the other hand our sample has fewer students than what is displayed in national records. Finally, the deviations may be explained by statistical errors (sampling). 42 In our sample two schools were among the schools without any initial allocation but were allocated funds by releases during the year. Both schools were established at the beginning of 2007.

0 %

50 %

100 %

150 %

200 %

250 %

300 %

350 %

400 %

450 %

500 %

Per

cen

t o

f M

oEV

T al

loca

tio

n r

ecei

ved

Schools

: Page 92

Contributions from parents are the other major source of funding. The contributions from parents

per student are almost at the same level for community and government schools. However, parent

contributions are significantly higher for rural schools than urban schools. Total rural schools enjoy

higher level contributions per student than urban schools both in terms of government grants

(MoEVT) and parent contributions.

Table 44 - Secondary schools cash receipts from parents per student (in TSh) - survey data

Schools Average cash contributions from parents Std Dev Min Max

Community 24,105 26,562 1,923 159,505

Government 26,620 14,094 16,654 36,586

Schools Average cash contributions from parents Std Dev Min Max

Dar es Salaam 19,457 10,370 8,352 44,452

Other urban 18,341 11,253 1,923 46,009

Rural 27,735 32,294 5,409 159,505

Grand Total 24,185 26,195 1,923 159,505

The contributions display significant variations between schools. School fees are composed of

different elements and some parents are granted exemptions due to low ability to pay. As many as

95% of the schools (including some government schools) in our sample received a fee subsidy to

compensate for fee exemptions granted to parents with low ability to pay. The data suggest that this

is the case also for schools were parents make a major cash contributions to the school as voluntary

or non-regulated contributions.

Only government schools received a food subsidy for students to compensate for exemptions

granted to parents with low ability to pay.

In kind contributions from MoEVT, parents and others constitute only a minor share of total

contributions to secondary schools judging from the data in our sample. While MoEVT contracted

larger civil works and paid suppliers directly from its budget in FY2008, this applied first and

foremost to major construction at new schools and in total for all secondary schools in FY2008 the

amount was insignificant compared to the total MoEVT secondary education expenditures (less than

1% of total MoEVT non-wage expenditures).

8.3.2 ALLOCATION AND COST OF TEACHERS

The number and quality of teachers are considered as important inputs in a student learning

environment. This section examines the allocation of teachers from the perspectives of the central

level and at the school level. Cost of teachers on government payroll could not be assessed since

both salary and allowances are paid individually to each teacher through a centralised payroll system

via specific secondary schools acting as a pay station on behalf of others. Accordingly, data on

remuneration for teachers by school was not available with MoEVT and only with a few of the

schools in the sample. Therefore, this section uses data availed through the secondary school survey

questionnaire. These data allowed a comparison to be made of teacher allocation by school location

as well as additional aspects associated with teaching such as attendance and the extent to which

schools employ and/or pay teachers from cash received from the various sources mentioned in

previous sections.

: Page 93

8.3.2.1 STUDENT-TEACHER RATIOS

The allocation of teachers in secondary schools as measured by pupil/teacher ratios varies between

urban and rural schools. As summarised in table 45 on average there are 29 students to a secondary

school teacher paid for from any source. However, when limited to teachers paid for solely by the

government, the P/T ratio rises to 41.

For schools in Dar es Salaam the P/T ratio is 29 and 32 for all teachers and government paid teachers

respectively. For schools in other urban areas these ratios become 28 and 43 respectively. In

contrast, for schools in rural areas the ratios are much higher at 34 for all teachers and 43 for

government paid teachers only.

The margin of difference in P/T ratios for all teachers and government paid teachers is smallest in

Dar es Salaam and largest for schools in rural areas. It means that schools in rural areas have a

higher proportion of teachers that are paid for by the schools cash receipts in the form of MoEVT

cash grants and parents contributions.

Table 45 - P/T ratio in secondary urban and rural schools - survey data

All teachers Average Std Dev Min Max

Dar es Salam 29 9 17 42

Other Urban 28 14 6 81

Rural 34 16 7 62

Total 29 14 6 81

Teachers govt. payroll Average Std Dev Min Max

Urban 32 12 17 50

Other Urban 43 24 14 119

Rural 43 19 8 71

Total 41 22 8 119

Community schools have far less teachers per student than government schools, both overall and in

terms of teachers on government payroll.

As evident from table 46, the lower number of teachers on government payroll per student for

community schools is partly compensated by employing teachers paid from the schools own cash

contributions. Government schools have a lower students-teacher ratio (16 for all teachers and 18

for teachers on government payroll only), i.e. from the outset a comparative low P/T ratio which

allows them to apply grants allocated to a larger extent to non-wage inputs (ref. sections presenting

composition of expenditures).

: Page 94

Table 46 - P/T ratio in secondary urban and rural schools - survey data

All teachers Average Std Dev Min Max

Community 29 14 6 81

Government 16 7 7 25

Total 29 14 6 81

Teachers govt. payroll Average Std Dev Min Max

Community 43 22 14 119

Government 18 8 8 29

Total 41 22 8 119

The significantly higher level of cash contributions available to rural and community as reflected in

table 43 and 44 above could have suggested that they have opportunity to employ more teachers to

compensate for the lower allocation. This is however only partly the case since a large share of the

contributions are tied to infrastructure investments (ref. sections below on composition of

expenditure).

There were significant variations in cash contributions between schools as well as structure of

expenditure. Some schools did used contributions for contracting teachers despite that the

allocation was indented for a different purpose.

Among the secondary schools, and in particular community schools, there are some with high level

of grants per student, high levels of parent contributions and thus resources to employ teachers

from own funds (hence a 'low' P/T ratio). There are also schools with low grant per student, lower

levels of parent contributions and thus limited resources to employ extra teachers above what is

paid from the government payroll. Finally, and as mentioned above, there are schools that apply

earmarked contributions for infrastructure and other none-wage spending for teacher salaries to

meet the demand for teachers in particular subjects. Accordingly, the level of and composition of

resources for schools vary significantly as will also be further discussed in sections on composition of

expenditure.

8.3.2.2 TEACHERS ATTENDANCE AT SCHOOLS

Similar as for primary schools, three sets of data were collected;

1. the number of teachers on government payroll allocated to the school and the number of

these teachers absent the full school year, and

2. the number of teachers present less than 50% of the year (recorded as 50% part time

teacher in our survey) and the number of teachers present more than 50% of the year

(recorded as full time teacher).

3. number of teachers on the payroll in May 2008 and the number of these teachers actually

present in the secondary schools during the same month.

This approach will, as mentioned in sections on primary schools, eventually give an estimate of full

time teaching above what is the reality.

Teacher absence was recorded in 56% of the schools in the sample with some schools as many as

70% of the teachers not present during the time for which data were recorded. Many of these

: Page 95

teachers were not on government payroll, i.e. several teachers are employed for a short period of

time depending on the availability of cash at the schools and opportunity to employ them for a

longer period of time.

In some of the schools it was observed that some of the teachers paid from the schools own

resources were in fact on the government payroll at another school. These teachers are included in

our sample of schools as teachers not on government payroll since they were not on the regular

payroll of the school in our sample (even if they are on the government payroll of another school).

They were hired to conduct extra classes or to compensate for lack of teachers on government

payroll and/or teachers with required qualifications and then sharing their teaching time between

schools (at one school paid by the government payroll while the other paid by the school cash

contributions).

Table 47 - Average number of teachers absent in percent of total number of teachers - survey data

School location Average Std Dev Min Max

Dar es Salaam 4 % 7 % 0 % 22 %

Other urban 12 % 12 % 0 % 36 %

Rural 16 % 22 % 0 % 70 %

Type of school Average Std Dev Min Max

Community 13 % 18 % 0 % 70 %

Government 18 % 23 % 0 % 48 %

Total 13 % 19 % 0 % 70 %

There are highest incidences of absence by teachers in rural schools (16%) compared to urban

schools (with 4% and 12% for Dar es Salam and other urban schools respectively). Government

schools have higher incidence of absence (18%) than Community schools (13%). In terms of teacher

absence it was predominately teachers on government payroll that did not report for reasons

unknown to the school i.e. of the teachers not present 89% were on government payroll..43

Since we do not have detailed payroll data for teachers on government payroll by each school we

are not in position to determine the exact loss due to teachers not performing at the school.

However, the cost of teachers on payroll is a more prominent issue for secondary schools than

primary schools judging from our sample and using total government PE expenditure for secondary

school teachers44 as a proxy for teacher costs the estimated cost was approximately 11.7 bln TSh (9.2

million USD) which is equivalent to 12,405 TSh (9.8 USD) per student in government and community

secondary schools. While the total cost is lower than for primary schools it is significantly higher

when measured per student compared to primary schools.

43 In one rural community school as many as many as 7 out of 10 teachers were absent for unknown reasons (illustrated by the Max figure in the table of 70%) with 5 of the teachers on government payroll.

44 Using FY2008 PE expenditure for secondary education sub-vote as reported by MoFEA.

: Page 96

8.3.2.3 TEACHER QUALIFICATIONS

Teachers in rural areas have generally lower qualifications than their urban counterparts. The higher

representation of more qualified teachers in urban areas also results in higher payroll cost in urban

areas where more of the qualified teachers tend to work. This is confirmed by the data in our sample

which show a higher share of teachers in rural and community schools are only licensed teachers

with Form VI exam as the highest level of education.

Table 48 - Distribution of teachers by qualifications (in percent of total teachers) - survey data

School location Graduate Diploma Licensed teacher Total

Dar es Salaam 15.5% 64.4% 20.1% 100.0%

Other urban 12.6% 66.7% 20.7% 100.0%

Rural 13.3% 55.2% 31.5% 100.0%

Type of school

Community 6.3% 64.3% 29.4% 100.0%

Government 50.9% 48.5% 0.6% 100.0%

Total 13.9% 62.0% 24.1% 100.0%

The government schools have the highest share of graduate teachers and very few of the schools

employ teachers with only Form VI exam, i.e. in terms of teacher qualifications urban and

government schools are able to attract the best qualified teachers.

8.3.3 GOVERNMENT GRANTS

8.3.3.1 CAPITATION GRANTS

As mentioned above the MoEVT allocates a number of grants to secondary schools. Two of these

grants are in total labelled capitation grants (grants for recurrent expenditures); a 'learning grant'

and 'school fee subsidy'. They are both allocated to the schools on the basis of the estimated

number of students the school has enrolled and which are eligible for exemptions from school fees

and paying for learning materials (i.e. the number of students in the formula do not include all

students enrolled).

In 2008 the Learning Grant was TSh 11,130 per student and the School Fee Subsidy was TSh 13,500

per student. These two grants make up what is the Capitation Grant for secondary schools. In this

case totalling TSh 24,630 per student on average.

The total allocation of the grants to the schools and the amount actually received per student

enrolled are displayed in table 49. The averages are lower than the grant amount used in the

formula since the grant is allocated for the number of students to be exempted. For 10.7% of the

schools all students were considered for exemptions.

: Page 97

Table 49 - Allocation and receipts of capitation grants per student enrolled (in TSh per student enrolled) - survey data

School location Capitation grant - MoEVT

student estimates Capitation grants allocated

per student enrolled

Capitation grants received per student

enrolled

Dar es Salaam 15,501 15,848 13,299

Other urban 17,873 18,648 18,828

Rural 17,550 18,543 18,076

Type of school

Community 17,090 17,964 17,821

Government 17,714 17,593 11,740

Grand Total 17,157 17,922 17,136

The difference between the average amount allocated and the amount disbursed and received by

the schools is due to within year change in enrolment data which differed from the estimated

number of students at the time estimating the total allocation. This is among others reflected by the

difference in per student allocation based on MoEVT estimates of students enrolled in the schools

and the actual number of students enrolled (the MoEVT data on number of students are generally

higher than the actual number of students actually enrolled in the schools).

There are two additional observations that can be made from the above table;

Schools in Dar es Salaam received less than initially allocated. The total amount of capitation

grants transferred was 16% below the initial allocation which in particular affected the urban

government schools in Dar es Salaam.

Since the capitation grants are allocated on the basis of estimated number of students to be

exempted from school fees and payment for learning materials it would have been expected

that there would be a higher per student allocation and release for rural schools than urban

schools but this is only the case for schools in Dar es Salaam45. It reflects to some extent the

result of that NBS HBS that found higher incidences of poverty and low income household in

peri-urban areas than in rural communities.

The capitation grants are allocated as a subsidy to schools to encourage enrolment of students with

low ability to pay for school fees and teaching materials, i.e. minimise the impact of household

income for access to secondary education. It would then be expected that in schools with lower cash

contributions from parents the allocation of capitation grant per student would be higher and; for

schools with higher cash contributions from parents the capitation grant per student would be

lower.

45 However, and as confirmed by NBS HBS 2007, peri-urban areas also have high incidence of low-income groups and higher than many rural communities.

: Page 98

Figure 20 - Secondary school MoEVT capitation grants and parent's cash contribution (in TSh per student enrolled) - survey data)

As reflected by the data in our sample this appears not to be the case. Only for some schools with

low levels of parent contributions per student is the MoEVT allocations higher. In most cases, and in

particular for community schools parent contributions are a major source of funding for contracting

of extra teachers in addition to teachers on government payroll i.e. parents of students exempted

from school fees also contribute with equal, and in many cases more cash, to schools than others.

8.3.3.2 DEVELOPMENT GRANTS

The development grant to schools comes in the form of contributions towards constructions and

rehabilitation of class rooms, teacher houses, libraries, school laboratories, administration buildings,

hostels, etc. At the level of the MoEVT the allocation of the development grant is based on needs

assessments i.e. those schools that are considered to be in need of a certain facility.

The average development grant per student was TSh 72,699 during FY2008. However, there are

significant variations between schools based on the extent to which they have been prioritised for

an investment in a specific facility. This variation is reflected in table 50 below.

At the school level, much more is received per student as development grant per student than

capitation grants. This observation is further reflected by the composition of expenditures at school

level as presented in the sections below. The amounts from Central Government transferred

through the sub-treasury exceeded the figures initially allocated by the MoEVT; i.e. additional

sources of funding for investment and rehabilitation of facilities were extended from others (among

others councils and RAS).

Table 50 - Allocation and receipts of development grants per student enrolled (in TSh per student enrolled)

School location Allocated Received

Dar es Salaam 15,474 13,695

Other urban 51,181 29,312

Rural 95,773 93,246

Type of school

Community 72,799 68,962

Government 70,919 1,830

Grand Total 72,699 65,382

-

10 000

20 000

30 000

40 000

50 000

60 000

70 000

80 000

- 20 000 40 000 60 000 80 000 100 000 M

oEV

T al

loac

tio

n p

er s

tud

ent

Parents cash contibution per student

: Page 99

The major share of the development grants was allocated to rural community schools. The level of

releases were as allocated for these schools. 46 For the government schools a large amount initially

allocated was not released however part of the amount was reallocated for other schools.

As will be presented in sections below however, the actual amount of development grants

transferred is substantially more than the amount actually invested in infrastructure.

8.3.3.3 TIMING OF DISBURSEMENTS OF GRANTS TO SCHOOLS

The timing of transfers and receipts may be part of the explanation for lower level of execution by

schools compared to the amount of grants received. It is also an indicator for efficiency of resource

flows and their utilisation. Grants disbursed and received late will have reduced impact.

On average for the schools, the first disbursement of capitation grant for FY2008 was received in

December 2007 i.e. 181 days after the beginning of the fiscal year but before the start of the next

school year. There are however, considerable variations in durations between schools with some

experiencing short delays of 43 days while some recording delays of as long as 269 days.

Table 51 - School receipt of first capitation grants allocation - Average number of days into FY2008 - survey data

School location Average Min Max

Dar es Salaam 137 100 257

Other urban 146 72 269

Rural 185 43 474

Type of school

Community 165 43 474

Government 172 100 269

Grand Total 165 43 474

On average schools receive the first tranche of the grant during November/December in the fiscal

year with average releases earlier for urban schools and later for rural schools. For one rural

community school in the sample the first tranche was released close to a month after the beginning

of the fiscal year while at the other end one school did not receive the first tranche for the FY2008

before the first quarter of the next fiscal year.

For the development grant a similar pattern of delays in disbursement and receipt are observed as

illustrated in table 52 with average delays of 225 days into FY2008 on disbursement and receipt of

this grant. In general releases of development grants take place at later date in the fiscal year than

capitation grants and with one school not receiving any of the grants allocated before the first

quarter of the next fiscal year FY2009 (end of the school year 2008).

46 The school that received the highest development grant per student was 167 kilometres from the 'urban' centre (Council head quarters). It was a foster school with a total of three classrooms and a total enrolment of 23 students. Clearly a unique situation in a rural part of the country.

: Page 100

Table 52 - School receipt of first development grants allocation - Average number of days into FY2008 - survey data

School location Average Min Max

Dar es Salaam 230 137 290

Other urban 191 109 302

Rural 245 29 476

Type of school

Community 221 29 476

Government 302 302 302

Grand Total 225 29 476

Unlike for the primary schools we did not have access to actual release dates from the sub-treasuries

for reasons mentioned earlier. Accordingly it could not be determined if the lead time for school

bank accounts to be credited was due to late release from MoEVT, the sub-treasury or due to

exceptional long banking days (or a combination of them). However, the findings from the analysis

of primary school data suggest that there significant delays in crediting the school bank accounts by

some of the bank branches, i.e. significant internal transaction delays and in case of primary schools

even lower amounts credited the school account than what was instructed by the councils according

to their payment advise.

8.3.4 SCHOOL LEVEL EXPENDITURES

Secondary schools spend the grants and other contributions on a number of cost items which are

accounted for according to the Government regular chart of accounts as reflected in the table 53

below. The data are presented separately for community and government schools since the latter

were also a pay-station for payment of salaries to their teachers while the former do not have

payroll data.

The major cost item for all community schools is expenditure for infrastructure (new classrooms,

staff houses, library, laboratory, hostels for boarding schools, etc.). The cost however varies

significantly between schools in respect of infrastructure investments since not all were awarded

resources in the form of development (infrastructure) grants by MoEVT in FY2008. Investments in

government schools are in most cases managed with direct payments to contractors from MoEVT

and thus not recorded as expenditures in school cash books and ledgers.

Another significant cost item is payment to teachers (both on government payroll and other

contracted teachers). On average rural schools and in particular rural community schools spend a

significant share of their revenue to employ extra teachers over and above the teachers employed

on government payroll (like for allowances to teachers for extra classes and other contributions to

them). This is further confirms the results presented in sections above presenting data on allocation

and distribution of teachers.

: Page 101

Table 53 - Secondary schools - Structure of expenditure (in TSh per student) - survey data

School location

Teachers on gov. payroll

Teachers/ others not

on gov. payroll

Textbooks/ teach. tools

Exercise books, exam. paper, chalks,

pencils

School meals

Construction/ rehab.

Desk/ chairs/ tables

and repair

Other

Dar es Salaam 2,084 1,713 10,227 4,687 1,733 8,587 938 6,304

Other urban 3,501 2,764 7,673 3,077 3,473 15,907 918 2,897

Rural 3,220 6,906 6,754 3,664 16,070 12,254 1,942 6,211

Type of school

Community 2,947 8,825 7,383 3,755 3,721 13,479 1,558 5,137

Government 3,858 2,402 11,850 3,547 59,197 2,904 132 6,698

Total 3,036 5,490 7,821 3,735 9,160 12,442 1,418 5,290

School meals are a significant cost item for the average rural government schools while urban and

government schools allocate more to textbooks and teaching tools. Other teaching and examination

materials are more or less equally distributed.

Comparing the expenditure data with grants and other contributions received it is evident that part

of the development grants have not been fully executed (infrastructure investments account for 19%

of grants transferred) of which part of the amount remains as balances in school accounts but parts

of the grants have also been advanced by the schools for non-investment purposes i.e. the

development grant earmarked for infrastructure investments have been used by the schools for

other purposes, among other to hire additional teachers.

8.4 SCHOOL LEVEL RESOURCES AND PERFORMANCE

Tanzania’s expansion in secondary education over the past six years is impressive. However, the

ultimate question is whether the expansion in secondary education is matched with a

commensurate increase in resource allocation. More importantly whether such expansion is

implemented without undue sacrifices on quality of secondary education as measured by students’

performance. In this section we will present data on school performance first and foremost

measured by pass rates.

Out of the 75 sampled secondary schools, 24 had classes up to Form IV with examination results. 72

of the 75 schools had Form II examination results. Tables 54 and 55 summarise student examination

performance by urbanisation as well as government versus community secondary schools.

For 2008, Form II pass rates were on average 74% for all schools in the sample and Form IV

examination results displayed an aggregate pass rate of 84% on average for the 24 schools.

There are notable variations in performance between rural and urban secondary schools. While

schools in Dar es Salaam had an average pass rate of 91%, other urban schools had an average pass

rate of 69% and rural schools on average recorded a pass rate of 72%.

: Page 102

Table 54 - School Form II and Form IV pass rates (in percent) - survey data

School location Average Pass Form II Average Pass Form IV

Dar es Salaam 91% 93%

Other urban 69% 94%

Rural 72% 79%

Type of school

Community 73% 82%

Government 90% 96%

Total 74% 84%

The difference in examination performance is significant between government and community

schools. While government owned schools have a 90% pass rate on average, community owned

schools have an average of 71%.

For Form IV results the same pattern is observed. With an overall average pass rate of 84%, the

average for government schools is 96% while community schools have a pass rate of 82%.

Urban government schools are performing better as measured by Form II and IV pass rates while

rural community schools have lower pass rates. It is in rural community schools that a major share of

the expansion in enrolment has taken place during the last years.

Pass rates are not correlated with overall P/T ratios. However it is correlated with P/T ratios of

teachers on government payroll. It suggest that the return on teachers on government payroll is

higher than for teachers employed directly by the school. Combining this result with data on teacher

qualifications clearly indicate that schools with teachers with higher levels of education display

higher pass rates than schools with teachers of lower qualifications. The former is a typical feature of

government and urban schools, the latter a typical feature of rural and community schools.

Figure 21 - P/T ratio (government payroll) and school pass rate Form II - survey data47

47 Schools with expenditures on investments in new administration and other buildings have not been included since the impact on larger infrastructure investments are assumed to have impact after the investment has been completed.

-

20

40

60

80

100

120

- 0,200 0,400 0,600 0,800 1,000 1,200

Pass rate per school

P/T

go

vern

men

t te

ach

ers

: Page 103

Government efforts to employ more, and more qualified, teachers on account of teachers

contracted by the schools themselves paid by cash grants and parents' contributions will increase

school performance as measured by pass rates. It suggest that the large expansion of community

schools to a large extent served by teachers employed and paid for by the schools themselves has

have resulted in an overall decline in school performance as measured by pass rates. It suggest that

there is a need to put significantly more emphasis on recruitment of teachers with higher

qualifications in community schools the improve overall performance (and likely quality of education

when using pass rate as a proxy for quality of education).

Our data suggest that non-wage spending per student is correlated with the examination

performance of a school (measured by pass rates) as illustrated in the figure 22. It is also linked to

the Gross Point Average (GPA)48. The more total non-wage spending per student the higher the

average O Level GPA for a school.

The magnitude of the input from contract teachers is correlated with schools non-wage resources

from which they are paid. It means that more non-wage finance for schools allow them to contract

more teachers and retain them for a longer period of time. However, and as suggested by the above

findings, it would yield an even higher return measured by pass rates if the government recruited

more qualified teachers and deployed them to community schools rather than the schools using

government cash grants and other contributions to directly employ teachers. Another approach, as

also applied in many other countries, is to set standards for teacher education levels serving in a

school (in some countries even specific minimum qualifications linked to subject being taught). It

means that teachers directly employed by the schools need to meet higher qualification standards

than what appears to be the case today.

Figure 22 - Expenditure per student (in TSh per student) and school pass rate Form II - survey

data49

48 Gross Point Average is used by NECTA as an index of results across A and O level exams in different subjects.

49 Schools with expenditures on investments in new administration and other buildings have not been included since the impact on larger infrastructure investments are assumed to have impact after the investment has been completed.

-

10 000

20 000

30 000

40 000

50 000

60 000

70 000

80 000

90 000

100 000

- 0,200 0,400 0,600 0,800 1,000 1,200

Exp

en

dit

ure

pe

r st

ud

en

t

Pass rate per school

: Page 104

The classifications of schools by rural versus urban, level of poverty in a region of council and other

aggregate indicators are not taking into account the diversity of communities in which schools are

located. Location is assumed not only to affect a school overall resource input and use (ability to

employ and retain teachers) but also the performance of students and may serve as an indicator of

poverty levels and overall environment for the level of all services. As with primary education the

actual location of a school measured by its distance to an 'urban' location appears to be of

significance for school performance measured by pass rates.

In figure 23 we show distance to council headquarters and Form II pass rates for schools located in

'rural districts'. Evidently, location of schools and subsequently location of communities in which the

students live have impact on their performance.

Figure 23 - Expenditure per student (in TSh per student) and school pass rate Form II - survey

data50

Combining the above results show that rural schools in 'rural districts' with low pass rates have

higher P/T ratios for teachers on government payroll and the pass rates are lower the lower level of

resources they have available for spending, not least for contracting of teachers to compensate for

lower P/T ratios of government teachers. It suggests, similar to primary schools that there is a need

to address teacher allocation disparities and their level of qualifications. One option is to provide

added incentives for secondary school teachers to serve in 'rural' schools in 'rural' districts, or, higher

level grant allocations per student to be able to contract an even higher number of teachers from

these additional resources. In addition, applying and enforcing minimum qualification standards of

teachers will likely yield significant returns and if also applied in rural community schools it will yiled

significant returns as measured by higher pass rates.

8.5 SECONDARY EDUCATION AND GENDER

The gender distribution of enrolment in schools is more biased for secondary schools than for

primary schools51. On average 42% of the students enrolled in secondary schools are girls in our

50 Schools with expenditures on investments in new administration and other buildings have not been included since the impact on larger infrastructure investments are assumed to have impact after the investment has been completed.

-

20

40

60

80

100

120

140

- 0,200 0,400 0,600 0,800 1,000 1,200

Dis

tanc

e fr

om

co

unic

ilH

Q

Pass rate per school

: Page 105

sample with a lower share for rural schools. However, there are significant variations between the

schools in our sample regardless of location. Some schools have lower than 20% share of girls (rural

schools) while none have more than 54%, i.e. there is a significant challenge to enrol more girls in

secondary education.

Table 55 - Percent share of girls enrolled in community secondary schools - survey data

School location Average Std Dev Min Max

Dar es Salaam 44 % 7 % 27 % 51 %

Other urban 45 % 5 % 32 % 50 %

Rural 41 % 9 % 18 % 54 %

Total 42 % 8 % 18 % 54 %

The gender distribution of teachers is notably different for secondary schools compared to primary

schools in our sample with an overall average of 26% female teachers. This low share of female

teachers concerns first and foremost rural and community schools and some have only male

teachers serving (both teachers on government payroll and teachers contracted by schools).

Table 56 - Average number of teachers absent in percent of total number of teachers - survey data

School location Average Std Dev Min Max

Dar es Salaam 48 % 20 % 22 % 78 %

Other urban 35 % 23 % 0 % 95 %

Rural 17 % 16 % 0 % 59 %

Type of school Average Std Dev Min Max

Community 25 % 21 % 0 % 95 %

Government 44 % 23 % 30 % 78 %

Total 26 % 22 % 0 % 95 %

P/T ratios, capitation grants, development grants, other contributions and expenditure per student

were analysed from the perspective of gender. Generally there is no notable correlation between

the proportions of girls in schools or those of teachers in schools with the school level inputs.

However, as presented in sections above, P/T ratios and resources per student are linked to school

location and so is gender distribution in terms of girls and female teachers. While level of non-wage

inputs may not change the gender bias, anecdotal evidence suggests that employment of female

teachers has impact on girls enrolment in schools. Combined with results presented in sections

above, it suggest that recruitment of higher qualified female teachers in rural and community

schools will have a significant impact on improvement in pass rates and enrolment of girls in

secondary education.

51 In our sample all government schools were either girls or boys schools and subsequently the focus has been on community schools judging from the data in our sample.

: Page 106

ANNEX I - TERMS OF REFERENCE

CONSULTANT

TERMS OF REFERENCE

FOR

TANZANIA EDUCATION PUBLIC EXPENDITURE TRACKING (PETS)

TANZANIA (2008/9)

BACKGROUND

In recent years, key education policy reforms in Tanzania have ensured sustainable

progress towards quality education for all and the achievement of MKUKUTA52

education goals. Impressive quantitative progress has indeed been registered in

both primary and secondary schools, as a result of the inception of the Primary

Education Development Programme (PEDP) in the Secondary Education

Development Programme (SEDP) in 2002 and 2004 respectively. The Net Primary

Enrolment Rate (NER) has increased dramatically from 65.5% in 2001 to 97.2% in

2008. In Secondary education the NER has also increased from 6.3% in 2003 to

23.5% in 2008.

The Ministry of Education and Vocational Training (MOEVT) has also undertaken a

comprehensive process and developed an Education sector development

programme (ESDP) 2008-2017 which covers all levels of education. This is the

guiding document for the education sector. The operationalisation of the ESDP,

including prioritization and costing is currently in process for 2008/09 to 2010/11.

The funding for the sector is largely provided through the GoT budget, with an

average of 18% of the national budget provided to the sector. This reflects strong

political will to the education sector.

Understanding how the funds flow through the system is an integral part of the

Ministry’s planning and budgeting purposes. To that end, Public expenditure

Reviews have been conducted for 2002, 2003, 2004, and 2005, with a Public

expenditure tracking survey (PETS) undertaken in 2004. Despite the fact that

mechanisms for control such as internal and external auditing have been taking

place, there is no formal assessment of public expenditure tracking across all levels

of education that has since been undertaken. It was proposed during the Education

52 National Strategy for Growth and Poverty Reduction , in Kiswahili

: Page 107

High Level Meeting in December 2007 that further information on financial

management was required for the sector and that the best method for this would be

a Public Expenditure Tracking Survey (PETS).

As public expenditure reviews all function with different scopes, it is useful to clarify

this study’s objectives. It is intended to conduct a PETS which will seek to identify,

quantify and explain issues in the flow of resources budgeted for service delivery in

the education sector by examining flows from one part of government to another

down to the level of the school and focused on the single fiscal year of 2007/08.

Discussions around the methodology showed the need to review previous studies and

recommendations from PERs (Public Expenditure Reviews) and the 2004 PETS in the

education sector. As a result, it was decided to conduct the study in two phases.

Phase one to be a literature review to determine progress made on the previous

recommendations. Also clarification of objectives and key issues for survey design.

Phase two to include survey design, data collection and analysis.

The 2008/9 PETS should precede and feed into future public expenditure reviews in

education. The 2008/9 PETS will focus on 2007/8 financial data.

OBJECTIVES

The overarching objective of the study is to assist the MoEVT to improve the quality

of education service delivery in Tanzania. The PETS compares resource allocation

plans (budget) with actual expenditure of public funds with these direct objectives:

To Identify, quantify and explain flows of resources, including actual allocation of

resources in primary and secondary education – this includes an analysis of the rate

at which allocated public resources for education pass through different levels of

government and service delivery, including an estimation of the amount of outflow at

each level.

To examine financial information flow at primary and secondary education levels of

the sector with a view to provide a diagnosis of the public financial management

challenges in the flow of funds.

To examine the financial monitoring system by developing an accurate mapping of

the official channels for public resources to primary and secondary schools for the

relevant survey period (2007/8).

To examine efficiency and effectiveness of resource usage and accountability within

primary and secondary education

To explore challenges across primary and secondary education and provide

recommendations for improvement.

TASKS

: Page 108

The key tasks of the consultant will be to:

1. Deliver all reports as per the expected outputs in this TOR

2. Complete all activities indicated in tasks 1 through 4 that are detailed within annex 2 of this TOR. These include:

a. Efficiency of fund flows

b. Equity of fund flows

c. Monitoring results and

d. Policy implications

LITERATURE REVIEW AND METHODOLOGY

The study will be conducted in two phases.

PHASE ONE:

Phase One is a literature review that will include, but not be limited to the PERs from 2002,

2003.2004, 2005, and the PETS from 2004. It will also be useful to review the work

conducted at LGA level with USAID support to ensure that capacity is built across levels

down to the local level, during the PETS process and thereafter.

The consultants will undertake a literature review of relevant official, civil society and

academic literature relating to progress made against the recommendations of the 2004

PETS. The literature review should also take into consideration other relevant

documentation such as the revised household budget survey (2008) and the “Views of the

people” papers from 2008, as well as other studies done by Universities in Tanzania, and

development partners.

It is also proposed in phase one to conduct a joint field study visit to review the lessons

learnt on PETS that have been conducted in neighbouring countries of Zambia and Uganda.

A visit to the two countries would provide a comparative background and valuable lessons to

MOEVT staff on methodology, inclusiveness of relevant stakeholders, and dissemination

strategies.

PHASE TWO:

Phase Two will be undertaken once the literature review is completed. It will include the

development of instruments, data collection and analysis, culminating in a final PETS report.

The PETS will use standard methodology for studies of this type. Generally, the approach to

this study will include:

Charting the resource allocations and budget flows

Sampling of schools, districts and regions;

Selection and training of enumerators;

Design and pretesting of quantitative and qualitative survey instruments;

: Page 109

Conduct quantitative and qualitative assessments and discussions at primary and secondary

levels, across all levels of government and with key stakeholders, using various methods to

triangulate findings;

Collate and analyze data;

Write report:

Conduct at least two stakeholders’ workshops to get their inputs

Incorporate stakeholders comments;

Produce drafts for review in the ESDP fora, beginning from RACEF level

Incorporate comments from the fora on the recommendations made before drafts are

finalized

Disseminate study findings.

SURVEY INSTRUMENTS

The main instruments that will be used to conduct the PETS will focus on school, LGA,

Regional and Central Government. These will be developed by the consultants in line with

those at MOFEA, and during the inception period. It will be the responsibility of the

consultant to provide sample instruments for review by the RACEF TWG before they are

utilized.

The main areas covered in the instruments at primary school level will be flow of funds to

school level and other sources of support, and at secondary level it will also include staffing

levels and personnel management. As regards Government officials, the focus will be on

grants to LGAs and schools, recurrent and development expenditure, and other education

subsidies in cash or in kind.

All survey instruments and training materials that are required to successfully complete the

survey portion of the PETS should be completed in English but will also require translation

into Swahili as that will be the medium of training and dissemination of the survey

instruments. Copies must be provided to the sub-committee for review and approval before

proceeding to utilisation during the fieldwork.

Before the PETS can start to collect data, a survey strategy needs to be developed and

tested by the consultants and submitted to the sub-committee for approval. All Instruments

should be tested and refined amongst small numbers of the relevant types of respondents to

suit varying ranges of respondents. This should be combined with the enumerator training

exercise.

SAMPLE AND SAMPLING TECHNIQUES

The selection of the schools should be conducted using nationally accepted statistical

procedures and stratified random sampling.

The study will cover 10 regions including Dar es Salaam. Those ten regions will be chosen

to provide a sample of 3 regions each from the top, the middle and the bottom of the national

poverty profile, as set out in the 2008 Human Development report, as well as Dar es

Salaam. Further, in each region, three districts will be included. In each district 15 schools

: Page 110

will be visited: 10 Primary and 5 Secondary. These will be split into 40% urban and 60%

rural schools. The total number of schools to be covered in the survey will be 450. In

addition, centrally, three ministries (PMO-RALGA, MOFEA and MOEVT) will be covered.

For the study to be useful as a planning and budgeting tool within the sector, a stratified

samples needs to be developed which will allow the survey results to have equitable

representation across the country and statistical significance.

The PETS will cover primary and secondary education levels. The sources of funding

tracked will be public on-budget and off-budget funds.

With these factors in mind, the finalization of the school selection requires it to be

undertaken and approved by the sub-committee of the RACEF TWG. This must be

completed two weeks prior to the pre-testing of the instruments.

DATA COLLECTION, PRESENTATION AND ANALYSIS

The data collected should be both qualitative and quantitative in nature and will be reflected

in the development of the survey instruments.

The two lead consultants will manage the implementation of the data collection of the PETS.

They will work with the national firm of experienced enumerators to implement the survey.

The makeup of the consultant team for this assignment is detailed below.

Enumerators will need to be trained using the instruments for different respondent groups.

They will need to know about the purpose of the instruments as well as simply the

conventions for filling them in. Some instruments will involve qualitative discussions. The

field test of the instruments should be part of the enumerator training exercise. Some

modifications to the instruments might be made, including in consultation with enumerators,

after the field test.

Ground needs to be laid for survey implementation. Timetabling will be important to ensure

that instruments are all talking about the same period and are not affected by changes in

conditions during the sampling period. An efficient implementation will demand that

introductory letters are sent to all participating ministries, LGAs and schools, so no-one is

surprised when enumerators turn up. Consultants, along with MOEVT counterparts, would

also be responsible for conducting spot checks throughout the enumeration process.

Data entry and cleaning requires experience and resources and should result in a data set

that is accessible with an accompanying “data users’ guide”. The final output of this is a

database that can be used for future research purposes.

All parts of the survey design and implementation will be peer reviewed through the sub-

committee of the RACEF and will include support from NBS and/or other external experts.

DISSEMINATION

The continued engagement of all partners (GoT, DPs, CSOs, and relevant stakeholders) in

the process of financial tracking system studies such as the PETS, will be key to the

successful dissemination that leads to the meaningful sector policy reforms. Special

attention will be paid to the close involvement of all relevant stakeholders.

Recommendations coming out of the PETS as well as those from the stakeholders are to be

discussed internally and will form the basis of the Ministry’s action plan.

: Page 111

EXPECTED OUTPUTS

The consultants will jointly deliver an inception report, mapping, database, literature review

report and final PETS report. These outputs will be delivered as follows:

An inception report that further clarifies the objectives for both Phase one and Phase two. It will also provide details on the proposed sampling methodology, sample size and/or survey instruments for approval of the sub-committee, chaired by the MoEVT. The inception report will also include a plan of action for carrying out the tasks, and a detailed budget and time requirements for the makeup of the team members. All of these details shall be included in the inception report and be provided to the sub-committee within 2 weeks of the commencement of the consultant’s contract. This report should not exceed 10 pages.

An accurate mapping of the official channels for public resources to primary and secondary schools in light of restructuring within the MoEVT.

A database on public expenditures in primary and secondary education will include all the data that has been collected and the documentation of the sources of these data that have been collected and used during the study. This final database will be submitted on CD ROM disk at the end of the consultancy period.

A draft and final report for Phase One (literature review) of no more than 20 pages (excluding annexes) that includes:

a concise executive summary

Full documentation of the analyses made and list of documents reviewed

Recommendations on how to incorporate findings and gaps in the literature to enable smooth progression into Phase Two of the study

Proposed revision of the TORs for Phase Two of PETS to be presented to the sub-committee for finalisation

A draft and final report of no more than 50 pages (excluding annexes) that incorporates the findings and recommendations of Phases One and Two that includes:

a concise executive summary.

a description of the methodology (including the process followed, included but not limited to sampling procedure, sampling lists, training materials for enumerators/interviewer manual, completed surveys, and clean dataset

full documentation of the analyses made; fully covering all direct objectives of the PETS including:

an exploration of the relationship between different pieces of data to examine the hypotheses underlying the survey, with a narrative to discuss the implications of the findings

a discussion of the qualitative findings about the PFM challenges which might be preventing allocated resources reaching schools. It should link this discussion with findings on other drivers of variation in resource intensity, quality and outcomes across different schools.

descriptive statistical analysis of the major data areas in the instruments, including basic statistical data to understand averages, medians and variability between different administrative areas

: Page 112

detailed analysis which will explore the issue of the flow of funds including efficiency of management of resources, background on the processes of budgeting, administration, institutional context, etc. that affect the flow of funds

descriptive analysis of adherence to financial guidelines across all levels (national, regional and local authorities)

providing policy recommendations and conclusions to address planning and budgeting challenges in the sector.

Participation in all stakeholders meetings during the course of the assignment including the first dissemination workshops, to ensure that there is technical support in the presentation of the results to all key stakeholders.

Dissemination entails publicising the diagnostic and analytical reports but also involves making the data and the data users’ guide available to the academic and research community, and the general public for further analytical work as they see fit.

CONSULTING TEAM COMPOSITION

There will be a need for two consultants to work in cooperation to cover the full scope of this

assignment, one international and one national. A lead consultant will have ultimate

responsibility for the fulfillment of the entirety of this Terms of Reference, including the

expected outputs listed below. S/he will be expected to work with the co-lead for both Phase

One and Two and complete the assignment within a total of up to 50 days for the

international and 80 days for the national consultant over a 9 month period. Within the

overall timeframe, the Consultants will demonstrate that the work will be substantially

completed and a first draft report will be drafted by end of June 2009. Following any

comments, a revised final report would be delivered by the end of July 2009.

SUPERVISION AND REPORTING

The two consultants will work with and supervise the enumerator team, that will be

contracted and selected by the GoT in line with procurement procedures. However, the

consultants will ultimately be responsible for all the work produced for the PETS.

The supervision of the consultant will be by the sub-committee of the MOEVT, which

includes representatives of the Government departments, DPs, and CSOs. The finalization

of the whole team will need to be completed during the inception period once the survey

instruments are finalized and agreed.

The overall supervision of this PETS, including lead consultants, will be directly from the

Permanent Secretary of the MOEVT. On the technical side, the consultants will report to a

small sub-committee within the Resource Allocation Cost Efficiency and Financing (RACEF)

Technical working group (TWG). The members of the sub-committee will include, but not be

limited to, MDAs, DPs, and NSAs. These will include MOFEA (tracking unit), MOEVT, PMO-

RALGA, NBS, MoCDGC, DPs, and CSO. The sub-committee is tasked with making

recommendations to the RACEF on the following issues:

TOR for the study;

: Page 113

Supporting logistics throughout the process;

Evaluating consultant CVs (and firms if required) in line with GoT procurement procedures and systems (if applicable);

Suggesting the consultants (or firm if required) in line with GoT procurement procedures and systems (if applicable);

Coordinating stakeholders meetings in conjunction with MOEVT staff responsible for these duties on a regular basis;

Monitoring and providing quality assurance through overall supervision of consultants during the process;

Being the first point of contact with consultants;

Meet with lead consultants every 2 weeks and as needed throughout the process;

Review, discuss, comment on and provide recommendations on the inception, draft and final reports;

Agree the timeframe with the consultants and work with consultants to remain on track throughout the process.

TIMETABLE

This timetable is an estimate that the consultants will review during the inception period and

provide a final timetable within the inception report.

Proposed Dates (2008/09)

Activities Estimated Time required

By July Agreement of concept note completed

By Nov 7 Finalization of terms of reference for consultant at next RACEF

January 2009 Joint Regional lesson learning trip to Zambia or other neighbouring country with relevant experience with DP and MOEVT

3 days

Nov/Dec Identification of consultants 1 month

December Literature review and phase one report completed 1 month

January 2009 Stakeholder meeting to introduce PETS process ½ day

February 2009 Survey design and instrument development (finalize instrument, sample selection)

1 month

February 2009 Revision of instruments, Design data entry interface, and Field test

2 weeks

Feb/March 2009 Training of supervisors Training of enumerators and pre-testing

2 weeks

Feb/March 2009 Revision of instruments after pre-testing 1 week

March/April Fieldwork 6 weeks

May Complete data entry 2 weeks

May Complete data cleaning (survey firm contract ends) 2 weeks

May/June Data analysis 4 weeks

By end June Zero draft for reference group discussions 4 weeks

By end August Final Draft report for review 4 weeks

: Page 114

Proposed Dates (2008/09)

Activities Estimated Time required

July 09 Stakeholder review meeting of final draft 1 day

By mid August 09 Final delivery of report by consultant to MOEVT -

End Sept 09 Dissemination/ stakeholder meeting to launch report 2 days

October 09 Presentation at joint ESR and Action Plan review and approval for implementation by MOEVT

1 day

BUDGET

This PETS will be financed jointly between the GoT and the Development Partners. The two

lead consultants and the lesson learning trip will be financed by the DPs externally from

Government. The remaining costs, with the exception of the stakeholder’s and

Dissemination meetings, which will be financed by the GoT, will be financed by the DPs, with

funds paid directly to the MOFEA as a ring fenced activity.

CONSULTANT QUALIFICATIONS

The consultants will, depending on the consultancy for which they are contracted, possess

the following:

Qualifications:

Masters degree or higher in education finance or public financial management or economics

Experience:

Experience in conducting PERs and PETS

Knowledge of and relevant experience of public sector management

Knowledge of and relevant experience of the education sector, particularly primary and secondary education in Tanzania is an added advantage

Experience of quantitative research design, implementation and analysis; in particular survey design and implementation in areas where data sources are not or only scarcely available

Sub-Saharan African experience

Excellent analysis and report writing skills

Swahili fluency preferred (necessary for the lead national consultant)

DETAILED SCOPE OF WORK

Further to and expanding upon the objectives and outputs detailed above, these specific

tasks will also be required by the consultants to complete this TOR in full, as follows:

Task 1:

Address all of the questions raised below through both Phase One (in the literature review)

and/or Phase Two through the quantitative and qualitative surveys:

: Page 115

Efficiency of fund flows: Do primary and secondary schools receive the full amount of public

resources that are allocated in the budget and on time? If not, where do the resources go, do

they leak out of the system, or do they stay at higher administration levels? What factors

prevent schools from receiving the full allocated amount, or receiving allocated resources

late? How do these fund flow patterns differ by category (such as salaries, allowances,

school grants, recurrent and capital expenditure, etc)? How are the decision-making

processes at each level affected by these patterns? What are the notable changes from the

previous PETS (2004) findings?

Equity of fund flows: How does the pattern of resource flows differ between regions, districts,

and schools? What are the contributing factors of the differences? How equitable is that

delivery? What is the awareness of resource allocations across decision makers in the

regions and districts?

Monitoring Results: How is financial data collected across all levels of government? Where

are the points in the statistical chain (data collection, reporting, analysis and policy

development) of financial data that require strengthening? How effective is gender

disaggregation in resource allocation across all levels of government for both primary and

secondary education? How are those resources utilized? What information is available at the

school-level to show what resources have been spent on and what outputs have been

delivered as a result? How reliable is this information? What gaps are there in this

information? How could this information be strengthened? Do schools make an attempt to

compare their original budgets with actual patterns of expenditure - and do they compare the

expected outputs with what was actually delivered? Is this information made available or

presented to the community? How effectively is the school-level information aggregated to

provide a district and national picture?

Policy implications: What policy measures should be recommended to (a) ensure that public

funds flow efficiently and effectively through central – regional – LGA – school levels; (b)

ensure school level allocations are efficient, equitable and provide inclusiveness and

accountability to communities and lead to improved learning achievements; (c) identify

possible efficiency gains within the education system for increasing resource allocation in the

context of constrained resources and (d) ensure that the restructured MOEVT receives

relevant guidance in its devolution process of management and finances of secondary

education to the LGAs. These points should be analyzed while taking into account the

findings of the previous PETS conducted in 2004.

Task 2:

Ensure that all of the following issues are addressed within the final PETS

report:

Budget execution behaviour at different stages in the budgetary process (e.g. Allocations, releases,

transfers, receipts, expenditure, utilization); for flows of funds and also of materials delivered in –kind

to the school level: for salaries and staff.

Delays in flows of resources (financial and non-financial) which reduce operational efficiency

Efficiency of funding flows – reviewing where resources flow, if the flow reaching schools is less than

budget allocations and reasoning why

: Page 116

Variation between regions, districts and schools in resources budgeted and delivered, which may

have implications for equity, efficiency, and effectiveness of expenditure

Operation of fiscally decentralized government systems

Strengths and weaknesses of government systems and procedures for recording, reporting,

monitoring and accountability at different levels for both financial and non-financial resources.

Task 3:

Ensure that the underlying key hypotheses for investigation during the course of the PETS

include:

A. Incremental PFM improvements mean a larger share of allocated resources for education

reaches schools than in 2004 or before and “leakage” has been reduced. Therefore look at:

What percentage of resources allocated by central government is released to LGAs?

What percentage of resources for primary (secondary) education, released to LGAs, is

transferred to schools by those LGAs?

What percentage of transfers to schools is used for educational purposes in classrooms?

B. Despite reduced leakage, human resources and non-salary resources are allocated very

unevenly across primary schools with poorer, rural schools least resourced. Therefore look

at:

Does the pupil: teacher ratio vary significantly across primary schools correlated with

location and population characteristics? – rural/urban, average income;

Does the intensity of non-salary resources per enrolee vary significantly across primary

schools and is correlated with location and population characteristics? – rural/urban, average

income;

Is the intensity of capital expenditures in primary education correlated with location and

population characteristics? – rural/urban, average income;

Are parental and community contributions significant in primary/secondary schools and are

they correlated with the intensity of public resources?

C. In primary education, near universal enrolment was achieved in the 2001-2005 period so

now, incremental resources should feed into quality. Variation in resource intensity causes

variation in quality and in schooling outcomes, so quality and outcomes are worst in poorer,

rural schools. Therefore look at:

Are measures of primary schooling quality and schooling outcomes (dropout, completion,

exam passes, other tests, teachers’ presence/absenteeism?) correlated with resource

intensity?

Are measures of primary schooling quality and schooling outcomes correlated with resource

intensity conditional on location and population characteristics?

D. Secondary schooling is expanding fast but with uneven resource intensity and with

established schools more subsidised than new schools. Therefore look at:

Are non-capital parental and community contributions significant in secondary schools and

are they correlated with the intensity of public resources, and the age of the school?

: Page 117

Are capital contributions from parents and the community in secondary schools inversely

correlated with the distribution of public resources for running secondary schools?

Task 4:

Ensure that during the data collection and analysis process, the following basic steps are

followed:

A. Work closely with the responsible sector ministries and the Ministry of Finance to collect

and analyze data and formulate recommendations and action plan.

B. Liaise with a range of stakeholders including NGOs/CSOs and donors in the education

sector, local government institutions, the private sector, and the selected communities that

are part of the sampling frame.

Design survey instrument(s) to ensure that data is collected systematically and follow-up

studies can be undertaken relatively easy.

Ensure that translation of all materials (training and survey instruments) is completed at least

one week before the data collection begins as has been adequately pre-tested to ensure

that the translation accurately captures what the data is supposed to produce.

Assemble a team of highly qualified enumerators to implement the data collection. To

ensure the quality of the enumerators sufficient training will be provided to them with

sufficient funds and days available for such training, development of enumerator training

materials and sufficient field-testing to evaluate the quality of the enumerators’ capabilities;

Appoint supervisor(s) that will monitor the data quality by amongst others monitoring the

performance of the enumerators in the data collection process.

: Page 118

ANNEX I I - PRIMARY SCHOOL SAMPLE

Region/council Number of schools in the sample

Arusha 30

Arusha Tc 7

Karatu 9

Monduli 7

Ngorongoro 7

Dar es Salaam 30

Ilala 10

Kinondoni 10

Temeke 10

Kagera 48

Bukoba Tc 6

Karagwe 16

Muleba 16

Ngara 10

Kigoma 53

Kasulu 18

Kibondo 10

Kigoma Rural 18

Kigoma TC 7

Lindi 35

Kilwa 10

Lindi Rural 16

Nachingwea 9

Mbeya 44

Chunya 10

Ileje 9

Mbeya TC 8

Rungwe 17

Singida 43

Iramba 14

Manyoni 8

Singida Rural 14

Singida Tc 7

Grand Total 283

: Page 119

ANNEX III - SECONDARY SCHOOL SAMPLE

Region/council Number of schools in sample

Arusha 8

Arusha Tc 3

Karatu 3

Monduli 2

Dar es Salaam 11

Ilala 5

Kinondoni 3

Temeke 3

Kagera 11

Bukoba Tc 3

Karagwe 3

Muleba 2

Ngara 3

Kigoma 12

Kasulu 5

Kibondo 3

Kigoma Rural 1

Kigoma Tc 3

Lindi 9

Kilwa 3

Lindi Rural 2

Lindi Tc 3

Nachingwea 1

Mbeya 11

Chunya 1

Ileje 2

Mbeya Tc 3

Rungwe 5

Singida 13

Iramba 2

Manyoni 1

Singida Rural 7

Singida Tc 3

Grand Total 75