Document of The World Bank Report No: ICR00004211 IMPLEMENTATION COMPLETION AND RESULTS REPORT (IDA-47480) ON A CREDIT IN THE AMOUNT OF SDR 98.8 MILLION (US$ 150 MILLION EQUIVALENT) TO THE UNITED REPUBLIC OF TANZANIA FOR A SECONDARY EDUCATION DEVELOPMENT PROGRAM II APL 1 PROJECT June 27, 2017 Education Global Practice Africa CURRENCY EQUIVALENTS (Exchange Rate Effective June 21, 2017) Currency Unit = Tanzania Shillings (Tsh) US$ 1.00 = Tsh 2,237.0188 US$ 1.00 = SDR 0.7253 FISCAL YEAR July 1 – June 30 ABBREVIATIONS AND ACRONYMS ADEM Agency for the Development of Education Management APL Adaptable Program Loan CAS Country Assistance Strategy CPF Country Partnership Strategy DFID United Kingdom Department for International Development DP Development Partner DPL Development Policy Loan EMIS Education Management Information System EP4R Education Program for Results ESDP Education Sector Development Program ESMF Environment and Social Management Framework FM Financial Management FY Fiscal Year GDP Gross Domestic Product GER Gross enrollment rate GoT Government of Tanzania IDA International Development Association IO Intermediate Outcome JAST Joint Assistance Strategy for Tanzania LGAs Local government authorities M&E Monitoring and evaluation MKUKUTA National Strategy for Growth and Reduction of Poverty MoEST Ministry of Education, Science and Technology NECTA National Examination Council of Tanzania P4R Program for Results PAD Project Appraisal Document PCT Project Coordination Team PDO Project Development Objective PETS Public Expenditure Tracking Survey PMO-RALG Prime Minister’s Office-Regional Administration and Local Government QTSR Qualified teacher-student ratio RF Results Framework SEDP Secondary Education Development Program (2004-2020) ii SEDP I Secondary Education Development Program (2004-2009) SEDP II Secondary Education Development Program (2010-2015) Sida Swedish International Development Agency TIE Tanzania Institute of Education Tsh Tanzania Shillings TVET Technical and Vocation Education and Training Senior Global Practice Director: Jaime Saavedra Chanduvi Practice Manager: Sajitha Bashir Project Team Leader: Cornelia Jesse ICR Team Leader: Huma Waheed ICR Primary Authors: Huma Waheed and Brian Corry iii UNITED REPUBLIC OF TANZANIA Secondary Education Development Program II APL 1 Project CONTENTS Data Sheet A. Basic Information B. Key Dates C. Ratings Summary D. Sector and Theme Codes E. Bank Staff F. Results Framework Analysis G. Ratings of Project Performance in ISRs H. Restructuring I. Disbursement Graph 1. Project Context, Development Objectives and Design ............................................... 1 2. Key Factors Affecting Implementation and Outcomes .............................................. 6 3. Assessment of Outcomes .......................................................................................... 13 4. Assessment of Risk to Development Outcome......................................................... 25 5. Assessment of Bank and Borrower Performance ..................................................... 26 6. Lessons Learned ....................................................................................................... 29 7. Comments on Issues Raised by Borrower/Implementing Agencies/Partners .......... 30 Annex 1. Project Costs and Financing .......................................................................... 31 Annex 2. Outputs by Component ................................................................................. 32 Annex 3: Economic and Financial Analysis………………………………………… 37 Annex 4. Bank Lending and Implementation Support/Supervision Processes ............ 39 Annex 5. Summary of Borrower's ICR and/or Comments on Draft ICR ..................... 40 Annex 6. List of Supporting Documents ...................................................................... 50 MAP iv A. Basic Information Secondary Educ. Country: Tanzania Project Name: Development Program II Project ID: P114866 L/C/TF Number(s): IDA-47480 ICR Date: 06/20/2017 ICR Type: Core ICR REPUBLIC OF Lending Instrument: APL Borrower: TANZANIA Original Total XDR 98.80M Disbursed Amount: XDR 81.27M Commitment: Revised Amount: XDR 82.18M Environmental Category: B Implementing Agencies: Ministry of Education, Science and Technology Cofinanciers and Other External Partners: B. Key Dates Revised / Actual Process Date Process Original Date Date(s) Concept Review: 09/03/2009 Effectiveness: 09/08/2010 05/22/2015 Appraisal: 03/18/2010 Restructuring(s): 11/24/2015 12/20/2016 Approval: 05/27/2010 Mid-term Review: 10/15/2013 10/04/2013 Closing: 12/31/2015 12/31/2016 C. Ratings Summary C.1 Performance Rating by ICR Outcomes: Moderately Satisfactory Risk to Development Outcome: Substantial Bank Performance: Moderately Satisfactory Borrower Performance: Moderately Satisfactory C.2 Detailed Ratings of Bank and Borrower Performance (by ICR) Bank Ratings Borrower Ratings Quality at Entry: Moderately Satisfactory Government: Moderately Satisfactory Implementing Quality of Supervision: Moderately Satisfactory Moderately Satisfactory Agency/Agencies: Overall Bank Overall Borrower Moderately Satisfactory Moderately Satisfactory Performance: Performance: v C.3 Quality at Entry and Implementation Performance Indicators Implementation QAG Assessments (if Indicators Rating Performance any) Potential Problem Project Quality at Entry No None at any time (Yes/No): (QEA): Problem Project at any Quality of Supervision Yes None time (Yes/No): (QSA): DO rating before Moderately Closing/Inactive status: Satisfactory D. Sector and Theme Codes Original Actual Major Sector/Sector Education Secondary Education 100 100 Major Theme/Theme/Sub Theme Human Development and Gender Education 42 42 Access to Education 42 42 Education Financing 34 34 Science and Technology 8 8 Standards, Curriculum and Textbooks 8 8 Teachers 8 8 E. Bank Staff Positions At ICR At Approval Regional Vice President: Makhtar Diop Obiageli Katryn Ezekwesili Country Director: Bella Bird John McIntire Practice Manager: Sajitha Bashir Christopher J. Thomas Task Team Leader(s): Cornelia Jesse Arun R. Joshi ICR Team Leader: Huma Ali Waheed ICR Primary Author: Huma Ali Waheed Brian Lyn Corry F. Results Framework Analysis Project Development Objectives (from Project Appraisal Document) Improve the quality of secondary education, with a focus on underserved areas. vi Revised Project Development Objectives (as approved by original approving authority) (a) PDO Indicator(s) Original Target Actual Value Formally Values (from Achieved at Indicator Baseline Value Revised approval Completion or Target Values documents) Target Years Completion rates at the O level (disaggregated by gender) Indicator 1 : Value quantitative or 22% (M-24%; F-20%) 39% 30% 36.9% Qualitative) Date achieved 05/03/2009 06/30/2015 06/30/2016 06/30/2016 123% of the revised target and 95% of the original target was achieved by Comments (incl. % project closing. Data disaggregated by urban/rural was not reported after 2012 achievement) and this information is not currently available. Completion rates at the A level (disaggregated by gender) Indicator 2 : Value 5.4% (M-6.4%; F- quantitative or 2.6% (M-3.2%; F-2.0%) 4.1% 3.5% 4.5%) Qualitative) Date achieved 05/03/2009 06/30/2015 06/30/2016 06/30/2016 154% of the revised target and 131% of the original target was achieved by Comments (incl. % project closing. Data disaggregated by rural/urban was not reported after 2012 achievement) and is not currently available. Qualified Teacher-Student ratios (QTSR) Indicator 3 : Value quantitative or 1:59 1:48 1:25 1:20 Qualitative) Date achieved 05/03/2009 06/30/2015 06/30/2016 06/30/2016 This indicator was disaggregated by rural-urban at preparation but this Comments (incl. % information was not reported after 2012. Over 100% achieved by project achievement) closing. Proportion of public schools meeting government-approved minimum standards Indicator 4 : for infrastructure and a student-teacher ratio of 40:1 Value quantitative or 4% 37% 26% Qualitative) Date achieved 05/03/2010 06/30/2015 06/30/2016 vii Comments (incl. % 70% of target achieved achievement) Qualified Teacher-Student ratio (QTSR) for 500 hard to reach schools Indicator 5 : Value quantitative or 1:59 1:25 1:17 Qualitative) Date achieved 05/03/2010 06/30/2016 06/30/2016 This indicator was introduced after the first restructuring to measure Comments (incl. % performance in lagging rural schools. The target was substantially exceeded by achievement) project closing. Number of direct project beneficiaries (of which female) Indicator 6 : Value 1,300,000 quantitative or 0 1,728,492 (48%) 1,469,589 (50%) (48%) Qualitative) Date achieved 05/03/2010 06/30/2015 06/30/2016 06/30/2016 Comments (incl. % 113% of revised target achieved achievement) Increase in gross enrollment rate in the five lowest enrollment regions at the O Indicator 7 : level Value quantitative or 22% 30% 28% Qualitative) Date achieved 05/03/2010 06/30/2016 06/30/2016 Comments This indicator was added in May 2015. 93% of target achieved by project (incl. % closing achievement) (b) Intermediate Outcome Indicator(s) Original Target Actual Value Formally Values (from Achieved at Indicator Baseline Value Revised Target approval Completion or Values documents) Target Years Secondary schools meeting agreed minimum infrastructure standard (% Indicator 1 : increase) Value (quantitative 4% 37% 26% or Qualitative) Date achieved 05/03/2010 06/30/2015 06/30/2016 Comments 70% of target achieved viii (incl. % achievement) Hard to reach secondary schools with teacher residences Indicator 2 : Value (quantitative 0 28 367 375 or Qualitative) Date achieved 05/03/2010 06/30/2015 06/30/2016 Comments (incl. % This indicator was changed in May 2015. 102% of revised target achieved. achievement) Secondary classrooms built (cumulative) Indicator 3 : Value (quantitative 0 4800 n/a 2764 or Qualitative) Date achieved 05/03/2010 06/30/2015 06/30/2016 Comments (incl. % 57% of target achieved. achievement) Student-teacher enrollment in the science and mathematics teaching courses at Indicator 4 : teacher training colleges Value (quantitative 2,650 3000 7896 or Qualitative) Date achieved 05/03/2010 06/30/2016 06/30/2013 Comments The last enrollment number available is from 2013. Cumulatively, 12,600 were (incl. % enrolled. achievement) Newly qualified teachers taking up posts in hard to reach schools and receiving Indicator 5 : settling in allowances Value (quantitative or Qualitative) Date achieved This indicator was dropped given that settling-in allowances were not provided Comments (incl. % to teachers due to opposition. An alternate incentive (furniture for residence) achievement) was provided to attract teachers to hard-to-reach schools. Teachers trained in service (number) Indicator 6 : Value (quantitative 0 30,000 n/a 28,060 or Qualitative) Date achieved 05/03/2010 06/30/2015 06/30/2016 ix Comments (incl. % 93.5% achievement) Management training for heads of public schools (number) Indicator 7 : Value (quantitative 0 12,300 n/a 8,233 or Qualitative) Date achieved 05/03/2010 06/30/2015 06/30/2016 Comments (incl. % 67% of target achieved achievement) % of schools receiving full resources (Tsh 25,000 per student) Indicator 8 : Value (quantitative 0 100% n/a 71% or Qualitative) Date achieved 05/03/2010 06/30/2016 06/30/2016 Comments (incl. % 71% of target achieved achievement) Schools receiving allocations on a timely basis Indicator 9 : Value (quantitative 0 100% n/a 100% or Qualitative) Date achieved 06/30/2016 06/30/2016 Comments (incl. % 100% of target achieved achievement) Textbook-student ratio in mathematics, sciences and languages Indicator 10 : Value (quantitative 1:5 1:1 n/a 1:1 or Qualitative) Date achieved 05/03/2010 06/30/2015 06/30/2016 Comments (incl. % 100% of target achieved achievement) Textbooks procured by schools in mathematics, biology, chemistry, physics, Indicator 11 : English and Kiswahili (cumulative) Value (quantitative 1.7 million 6.8 million n/a 7.17 million or Qualitative) Date achieved 05/03/2010 06/30/2015 06/30/2016 Comments 105% of target achieved (incl. % x achievement) An integrated EMIS with data on secondary education completed and Indicator 12 : operational in LGAs (% of LGAs reporting through the integrated EMIS) Value (quantitative 0 100% 100% or Qualitative) Date achieved 05/03/2010 06/30/2015 06/30/2016 Comments (incl. % Target fully achieved achievement) An assessment of capacity for institutional development of TIE, ADEM, NECTE and EMAC with linkages to MoEST and PMO-RALG completed and a Indicator 13 : plan for phased institutional strengthening developed (yes/no) Value No institutional plans (quantitative Yes Yes under development or Qualitative) Date achieved 05/03/2010 06/30/2015 06/30/2016 Comments (incl. % Target fully achieved achievement) Examination pass rates in math, science and language Indicator 14 : A level (M- A level (M-64%; 83%; S-79%; A level (M-85%; S- A level (M-62%; S-79%; Value S-80%; L-96%) L-95%) O 94%; L-100%) O (quantitative L-95%) O-level (M-18%; O-level (M-21%; level (M-18%; level (M-17%; S- or Qualitative) S-52%; L-65%) S-54%; L-69%) S-52%; L- 53%; L-67%) 65%) Date achieved 05/03/2010 06/30/2015 06/30/2015 06/30/2015 Comments Original targets met for A level pass rates in all three subjects. Revised targets (incl. % met for Science and language for O level and nearly met for Math. achievement) G. Ratings of Project Performance in ISRs Date ISR Actual Disbursements No. DO IP Archived (USD millions) 1 06/27/2010 Satisfactory Satisfactory 0.00 2 03/26/2011 Satisfactory Satisfactory 43.47 3 11/13/2011 Satisfactory Moderately Satisfactory 43.47 4 06/20/2012 Moderately Satisfactory Moderately Unsatisfactory 43.47 5 12/25/2012 Moderately Satisfactory Moderately Satisfactory 70.10 6 05/17/2013 Moderately Satisfactory Moderately Satisfactory 70.10 xi 7 12/07/2013 Moderately Satisfactory Moderately Satisfactory 86.01 8 07/01/2014 Moderately Satisfactory Moderately Satisfactory 86.42 9 01/03/2015 Moderately Satisfactory Moderately Satisfactory 86.42 10 06/30/2015 Moderately Unsatisfactory Moderately Unsatisfactory 126.20 11 01/05/2016 Moderately Satisfactory Moderately Satisfactory 126.20 12 06/29/2016 Moderately Satisfactory Moderately Satisfactory 130.06 13 12/30/2016 Moderately Satisfactory Moderately Satisfactory 130.06 H. Restructuring (if any) ISR Ratings at Amount Board Restructuring Restructuring Disbursed at Reason for Restructuring & Key Approved PDO Date(s) Restructuring Changes Made Change DO IP in USD millions Changes to some key performance indicators were made to clarify language and 05/22/2015 N MS MS 126.20 targets and to bring them into closer alignment with the PDO. To take into account the impact on project targets of the Form 2 11/24/2015 N MU MU 126.20 Exam Policy of the government. Cancelation of funds allocated mainly for construction of 12/20/2016 MS MS 130.06 schools. xii I. Disbursement Profile xiii 1. Project Context, Development Objectives and Design 1.1 Context at Appraisal 1. Country Context. Tanzania’s economic performance had improved markedly over the previous decade at the time of project preparation in 2010. The country experienced annual growth rates exceeding 5 percent on account of increased mining and manufacturing activity; elimination of price controls; favorable interest and exchange rates; foreign and domestic investment inflows (including foreign aid receipts) for private and public infrastructure development; and increased tourism. 2. Improvements to tax administration and strengthening of the tax base increased government revenues. From the year 2000 to 2010 government revenues, as a percent of Gross Domestic Product, increased from 11.78 percent to 15.69 percent1. Some of the additional financial resources generated were deployed by the government to enhance social indicators. Over the same period, real GDP per capita more than doubled from US$308 to US$7082 and was accompanied by general improvements in human development indicators. 3. Efforts at alleviating income poverty, on the other hand, were not as successful. While poverty incidence decreased by 2.2 percentage points from 35.6 percent of the population in 2001 to 33.4 percent in 2007, given the population growth rate, the number of poor increased from 12.4 million in 2001 to 13.5 million in 2007, indicating that growth was not pro-poor. 4. Sector Context. At the time of Project appraisal, enrollment had increased significantly at the primary as well as post-primary levels. Access to education had largely been achieved at the primary school level. However, access to education at the secondary level was not equitable. A large share of children from low income households did not transition from primary to secondary education3. In 2006-2007, of the students who had not begun their first year of secondary school, 27 percent came from the lowest income quintile and 14 percent came from the wealthiest income quintile. At the lower secondary, there were no marked differences in enrollment based on gender. However, beginning in upper secondary, gender gaps began to appear. For the 2009-10 school year, there were two boys for each girl enrolled in Form 6 government schools. In addition, there were large rural-urban disparities. In 2007, for example, 77 percent of secondary school-age children in rural areas did not attend secondary school compared to only 17 percent in urban areas. The following were the key issues in the secondary education sector: 5. Misalignments in teacher availability and incentives for performance. There had been a large increase in the number of secondary schools which resulted in teacher shortages with an average qualified teacher-student ratio (QTSR) of 1:57. These shortages were more acute in the remote rural areas. There were no incentives for teachers to improve their performance and existing systems for accountability were inadequate. Traditional pedagogical practices and teaching to pass exams were common. 1 https://www.quandl.com/data/ODA/TZA_GGR_NGDP-Tanzania-General-Government-Revenue-of-GDP 2 http://data.worldbank.org/indicator/NY.GDP.PCAP.CD?locations=TZ 3 In Tanzania, the education cycle comprises of 6 years of primary, 4 years of lower secondary (Forms1-4 also known as O Level) and 2 years of upper secondary (Forms 5-6 also known as A Level) 1 6. Lack of adequate infrastructure. The government’s Secondary Education Development Program I (SEDP I) improved access to secondary education through a rapid increase in the number of school units, thereby achieving higher enrollment numbers. The expansion had taken place largely through a community based construction model which had led to large variations in quality. According to Ministry of Education, Science and Technology (MoEST) estimates, about 70 percent of secondary schools did not have laboratories, and about 90 percent required some form of rehabilitation, refurbishment, or completion, including addressing shortages of latrines and access to reliable sources of water. Low girls’ enrollments in the higher forms was due in part to the lack of water and toilet facilities at schools. The government had construction standards for secondary schools, which were not being strictly enforced given resource and capacity constraints. 7. Fluctuating intrasectoral financing and insufficient capitation grants. As a percent of the total education budget, the secondary education budget had been on a decline. At 12 percent, it was much lower than the regional average of 20 percent. Capitation grants (transfers to schools for teaching and learning material) were inadequate, partly due to the rapid increase in enrollments. Only Tsh13,500/student of the expected Tsh25,000/student were disbursed from 2006-09 due to rapid enrollment expansion. Levels of allocations and timeliness of grant receipts also varied widely. 8. Project Context. Government of Tanzania had developed a Secondary Education Development Program (SEDP) 2004-2020 to meet the growing parental, social and labor market demand for better educated and skilled citizens capable of stimulating and sustaining economic and social development and competing in the global economy. SEDP was designed for implementation in three phases over a 15-year period. SEDP I, implemented from 2004-2009, helped meet the early demands of secondary expansion. Under SEDP I, while plans were in place for expansion, quality, and capacity improvements, excessive demand for secondary education overshot even the high case scenario, leaving the system unable to provide adequately qualified teachers and sufficient amounts of capitation grants to schools. 9. The Bank proposed to take a long term approach to supporting the secondary education sector in Tanzania using an Adaptable Program Loan (APL)4 financing modality. The APL was proposed to be implemented in three phases over a 10-year period to support the second phase of, SEDP II. SEDP II was based on lessons learned from the implementation of SEDP I and designed to address the remaining challenges in the sector. The first phase of the APL, which this Implementation Completion Results Report (ICR) covers, was implemented over the period 2010- 2016. It aimed to provide a basic set of inputs (teachers, infrastructure, resources for teaching and learning) to secondary schools in the country to make them ‘fully functional’. 1.2 Original Project Development Objectives (PDO) and Key Indicators (as approved) 10. The overall development objective of the APL was to assist the Government of Tanzania (GoT) in implementing secondary education reforms to improve learning outcomes among secondary students. The specific PDO of the first phase of the APL was to improve the quality of secondary education with a focus on underserved areas. Underserved areas were defined at project approval as rural areas where public schools had inadequate infrastructure and unfavorable 4 APLs have since been discontinued by the Bank as a financing modality. 2 qualified teacher-student ratios (less than 1:40). The key indicators for measuring achievement of the PDO were: (a) Completion rates at the O level; (b) Completion rates at the A level; (c) Qualified teacher-student ratios (QTSR); (d) Improved examination pass rates in mathematics, sciences, and languages; (e) Proportion of public schools meeting government-approved minimum standards for infrastructure and an QTSR of 1:40; and (f) The number of direct project beneficiaries and the percentage of those beneficiaries who are female. 1.3 Revised PDO (as approved by original approving authority) and Key Indicators, and reasons/justification 11. No changes were made to the PDO. Changes to key indicators are described below. 12. In May 2015, several adjustments to the results framework (RF) were made: (i) the indicator related to examination pass rates was changed from a PDO indicator to an Intermediate Outcome (IO) indicator given the frequent changes made to the cutoff for passing examinations as well as changes to other pass/fail requirements which made it an unreliable measure of student learning. (ii) The IO indicator for availability of qualified teachers in rural areas was made a PDO indicator with a revised definition, QTSR in 500 hard-to-reach (HTR)5 schools. The rationale provided was to more clearly focus on the provision of teachers in underserved areas in line with the PDO. (iii) A new PDO indicator, the average gross enrollment ratio (GER) at the O level in the five lowest enrollment regions6, was also added to monitor progress in underserved areas in line with the second part of the PDO. Project performance at the time of the restructuring was Moderately Satisfactory. There were no significant changes to the project design or activities. 13. Further changes to the RF took place in November 2015 to take into account the impact of the Form II examination policy introduced by the government in 2012/13. This new policy led to the use of the Form 2 exam as a mechanism to tighten the flow of students to Form 3 onward which led to lowering of completion rates7 since fewer students than anticipated were being promoted to the final grades of the lower secondary and eventually upper secondary cycle. Project ratings had consequently been downgraded to Moderately Unsatisfactory. The reduction in number of pupils also impacted STR. Again, there were no major changes to the project design or activities. Changes to indicators and targets made during these two restructurings are summarized in Table 1. 14. Finally, in late 2016, the Government decided to change the approach to construction in the sector based on relatively better performance using other mechanisms (also supported by the Bank). The decision led to switch of funds allocated to the last phase of construction under SEDP II to the Education Program for Results (EP4R) project supported by the Bank and other 5 Hard to reach schools are the quartile of schools (500 in number) in underserved areas with the highest Student Teacher Ratios (STRs). 6 In 2009 (baseline year for the indicator), these were Katavi, Geita, Dar es Salaam, Tabora, and Rukwa. 7 Completion rates were defined as number of pupils/students who took the end of cycle exam of a given level of education expressed as percentage of school-age population of that grade. 3 development partners. Funds were canceled under SEDP II and added to an Additional Financing for the EP4R which was approved by the Bank’s Board in May 2017. No other changes were made. Project performance at that stage had been Moderately Satisfactory. Table 1: Changes Made to the PDO Indicators and Targets During Project Changes Indicator and Original targets Revisions in May 2015 Revisions in November 2015 PDO Indicator - Examination Made an IO indicator with revised No change pass rate in math, sciences, and targets – A level (M-83%; S-79%; languages at A and O levels – A L-95%) O level (M-18%; S-52%; level (M-64%; S-80%; L-96%) L-65%) O level (M-21%; S-54%; L- Targets were changed to reflect 69%) that expanding secondary enrollment brought in segments of the population without previous formal education and reducing pass rates. IO Indicator – Percent increase Made PDO indicator –QTSR for The targets for this indicator were in hard to reach schools with 500 hard-to-reach schools – 1:48. changed after the change in Form 2 qualified teachers – 56% This change reflects focus on the examination policy. The policy resulted provision of teachers to hard-to- in a lower number of students continuing reach schools. to Form 3 which improved the QTSR and targets were thus enhanced as well. PDO Indicator – Increase in This PDO indicator was No change gross enrollment rate in five introduced during the first lowest enrollment regions at the restructuring to measure project O-level – 30% impact on underserved areas. PDO Indicator – Completion No change The targets for this indicator were revised rate at the O-level – 39% after the introduction of the Form 2 examination policy which aimed to restrict the flow of students from Form 2 to Form 3. The target for O-level completion rate was reduced to 30%. PDO Indicator – Completion No change Similar to O level, A level completion rate at the A-level – 4.1% rates were expected to decline as a result of the new policy. The target for A level completion was revised to 3.5%. PDO Indicator – qualified No change The lower number of students passing teacher to student ratio (QTSR) Form 2 exams as a result of the new for public schools – 1:48 policy automatically translated into a decline of the QTSR. The target was revised to 1:25 to reflect this change. PDO Indicator – number of No change Given that the number of project direct project beneficiaries – beneficiaries fell as a result of the Form II 1,728,492 exam policy, the target for direct project beneficiaries was revised to 1,300,000 1.4 Main Beneficiaries, 15. The primary beneficiaries of SEDP II were secondary students with a particular focus on those in underserved areas. The project aimed to reach about 30% of the total number of students in government secondary schools in Tanzania. These students were expected to benefit from improved learning environments, additional school facilities, more and better qualified teachers and more resources for teaching learning material including textbooks. In addition, secondary government school teachers and school heads were expected to benefit from professional 4 development. Administrators of the government secondary school system benefitted from capacity building activities. In the long-term, the project sought to assist the government in building a high quality, productive and adaptive labor force for the benefit of the economy. 1.5 Original Components (as approved) 16. The project had four main components to support achievement of the PDO. 17. Component 1 – Upgrading existing schools infrastructure (Total US$270.5 million; IDA US$78.3 million). The objective of this component was to increase the number of secondary schools meeting the minimum infrastructure requirements, moving them toward becoming fully functional. This component was to provide support for rehabilitation and completion of existing structures, with a minimum, cost-effective infrastructure package. The component focused on six elements that were identified to be the most essential for enhancing the teaching and learning environment in Tanzania’s schools: classrooms, science laboratories, teacher residences (in HTR schools), latrines, water supply, and electricity supply (through the grid or solar power). These were also selected because they addressed concerns regarding teacher deployment and enrollment and attendance of girls. 18. Component 2 – Improving the equitable provision of teachers and the quality of teaching in mathematics, sciences, and languages (Total US$27.1 million; IDA US$27.1 million). This component was meant to deploy more teachers to rural areas and to introduce better teaching practices and instruction in core subjects. This component was to provide housing and a one-off settling-in allowance for new teachers in hard-to-reach schools. It also monitored and regulated the supply of teachers, provided science teaching facilities at teacher-education institutions, as well as a bridge course for new entrants to the science teacher training programs. It also focused on improving teaching delivery through attendance management, in-service teacher training and professional development, management training for administration, and a roadmap for the professionalization of secondary school teachers through performance standards and a skills upgrading credit system. This component also financed surveys monitoring teacher supply as well as impact evaluations of teacher incentives. 19. Component 3 – Ensuring adequate financing to secondary schools and improving utilization of resources (Total US$164.9 million; IDA US$37.8 million). This component was meant to ensure that financing for purchase of teaching and learning materials were disbursed on time and monitored effectively. Teaching and learning material included teacher supplies, manuals, and guides as well as student notebooks and pencils. This component also provided laboratory equipment, apparatus, and materials commensurate with student enrollment. 20. Component 4 – Providing capacity building and technical assistance to implement reforms (Total US$6.8 million; IDA US$6.8 million). This component aimed to strength the EMIS to collect secondary education data that is accurate, easy to retrieve and process, and readily available. It also planned to finance assessments of capacity of critical institutions in the secondary school sector including the Tanzania Institute of Education (TIE), Agency for Development of Education Management (ADEM), and the National Examination Council of Tanzania (NECTA)8, 8 EMAC was disbanded and its functions moved to TIE 5 and implement a phased plan for capacity development for the institutions to deliver on their mandates more effectively. 1.6 Revised Components 21. No changes were made to the project components during implementation. 1.7 Other significant changes 22. The closing date of the project was extended by one year from December 31, 2015 to December 31, 2016 at the time of the first restructuring in 2015 to allow for completion of construction and rehabilitation of school infrastructure. 2. Key Factors Affecting Implementation and Outcomes 2.1 Project Preparation, Design and Quality at Entry 23. SEDP II was designed to contribute to achieving the higher-level objectives of (i) the National Strategy for Growth and Reduction of Poverty (MKUKUTA in its Swahili acronym), (ii) the Joint Assistance Strategy for Tanzania (JAST), (iii) the Education Sector Development Program (ESDP) for 2008-2017, and (iv) the Secondary Education Development Program (SEDP), 2004-2020. It was thus well aligned with Government priorities and strategies for sector development. 24. The Bank’s support to Government of Tanzania’s SEDP was planned as a series of investments to ensure sustained improvements in the sector over a long term horizon given the upward trend in demand for secondary education in the country. The design of the project (SEDP II) benefitted from the implementation experience of the first phase of the program (SEDP I) which ended in 2008 and was supported by the Bank through a Development Policy Credit (DPC). SEDP I had succeeded in improving access to secondary education, especially for rural students, through reduction in school fees and rapid infrastructure expansion. However, it also created an urgent but sustained need to focus on quality aspects since adequate consideration was not given to teacher numbers and capitation grants, resulting in falling education quality and students’ learning outcomes. Schools had been constructed but adequate teachers had not been provided leading to student teacher ratios of 50+ in some regions. Resources to schools to meet needs for teaching and learning material were grossly inadequate and there were serious gaps in the quality of infrastructure which wasn’t conducive to learning at the secondary level. Expansion under SEDP I had been achieved through following a community supported construction model. Although this was successful in expanding the number of schools rapidly, the quality of construction was poor meaning that many schools didn’t meet minimum standards established by the government. There was a need to address these quality issues urgently. 25. The choice of an Adaptable Program Loan (APL) given past experience and the Government and Bank strategy of taking a long term approach to improving the secondary education sub-sector was sound. The choice of the instrument allowed the Bank to ensure that key policy parameters were implemented and sustainably financed. The experience of SEDP I showed that although policy actions were taken, their implementation was weak and more sustained support would be required to achieve results. A longer-term, phased implementation approach facilitated by an APL also provided an opportunity to build capacity in the government and phase in reforms systematically. Six years after project approval, there is strong evidence that this long term approach was the optimal choice. Although there have been significant improvements in the 6 sector in the last 6 years with support from SEDP II, large gaps remain in access to quality secondary education in Tanzania and there is a need for sustained support to the sector which the Government has requested despite the discontinuation of the APL instrument. The Bank has committed to providing this support through the results based Program-for-Results (P4R) instrument (currently under preparation). 26. The design of the APL drew from a strong base of independent studies, including a Public Expenditure Tracking survey (PETS) and review of human resources in the education sector carried out by UK’s DfID. It also effectively used lessons learned from the experience of SEDP I and other countries in East Africa. Some of the main lessons that SEDP II incorporated include:  Prioritizing and sequencing interventions taking into account limited capacity to implement large and complex programs. Although a comprehensive reform program for the sector was desirable, the capacity to implement, especially with several different and new stakeholders involved under a newly decentralized context, was weak9. The project proposed sequencing of interventions to first address urgent matters and setting a strong foundation before progressively strengthening capacity needed for more sophisticated reforms in later phases of the APL. This was an important design choice that was appropriate given the recent significant changes in the delivery structure with decentralization during project preparation.  Project design was simplified and focused on delivering a coherent set of interventions. SEDP II was designed to address the most urgent issues of incomplete schools, the lack of teachers, and inadequate funding for teaching and learning materials. The project aimed to bring secondary education provision up to at least a minimum standard to provide a conducive learning environment for students. To achieve this, it focused on implementing a coherent and well-coordinated system that delivers basic inputs following minimum standards and in a predictable and sustainable manner to schools. Infrastructure improvements were designed to be aligned with availability of qualified teachers and adequate provision of resources for the purchase of teaching and learning materials. This alignment of a small number of critical inputs to schools was one of the strongest aspects of project design.  The design of the project incorporated simple incentives and accountability measures to institutionalize efficient implementation of basic activities. For instance, LGAs were responsible for providing capitation grants and had the experience at the primary education level, but the finding in the draft 2009 education sector PETS showed that capitation grants actually reaching primary schools were less than what was released by the government and these funds were applied to other purposes within LGA plans. Incomplete information on the allocation and use of capitation grants constrained schools in their planning and MoEST in its monitoring and evaluation of the impact of such resources on learning. SEDP II introduced incentives to promote equitable teacher deployment, methods for monitoring teacher attendance, transparent mechanisms for the transfer and use of funds by schools, and full information disclosure and reporting. These were critical design choices that helped set a strong foundation for sector growth. 27. Although the lessons from past experience were incorporated, project design could have also benefitted from a deeper analysis of issues that affected implementation, especially with 9 Secondary education was decentralized to local governments in 2009 under SEDP I. Details in section 2.2 below. 7 respect to construction of school facilities. The project proposed to completely switch the construction modality from a community based model (which had faced quality issues) to one that was managed by the LGAs which turned out to be relatively slower and less cost effective (although quality was better). There were benefits to having communities involved in construction, which could have been preserved for the most effective approach to school construction, especially since responsibility for construction had only recently been decentralized to LGAs, implying that they did not have experience in the sector, which was an important risk to account for. Another key aspect that was not adequately addressed at the time of preparation was the coordination between MoEST and PMO-RALG/LGAs. The LGAs had a key role in implementation and their capacity was known to vary considerably. No formal arrangements were proposed to clarify roles of the two main entities involved in implementation. There were no formal commitments with respect to what each party would be responsible for delivering. This was especially challenging in the early stages of implementation and slowed progress considerably. 28. Project design could also have benefitted from deeper consultations for some of the politically challenging interventions that were proposed under the project. One example was the provision of a settling in allowance to teachers placed in hard to reach areas which faced opposition from teacher groups and could not be implemented as planned. Its feasibility for implementation could have been determined through consultations with relevant groups during preparation. 29. Key risks and mitigation measures. Despite the shortcomings described above, overall, key risks to implementation were adequately identified with appropriate mitigation measures. At the time of appraisal, the key risks identified came from inadequate financing and transition of secondary education delivery management from the MoEST to the PMO-RALG and LGAs. As a share of the overall education budget, the budget allocated to secondary education was low compared to the regional average. This indicated potentially weak support for reforms from MoFEA and MoEST. To mitigate the financing risk, the Bank agreed on an annual dated covenant for GoT contributions to the SEDP II Program. To mitigate the risk of low implementation capacity of MoEST and PMO-RALG, the project focused on key fundamentals only to avoid overwhelming system capacity. 2.2 Implementation 30. The project was approved by the Board on May 27, 2010 and declared effective September 8, 2010. The project closed after six years on December 31, 2016. Comprehensive arrangements had been put in place to implement project activities. Project implementation was overseen by MoEST in close collaboration with PMO-RALG. The project was designed to use already existing structures and mechanisms. The Education Sector Development Committee (ESDC), a body with representatives from basic education, TVET, and higher education, monitored the progress and provided a forum to receive updates from MoEST. The Director of Secondary Education, MoEST was responsible for project coordination and management of the project in conjunction with counterparts in PMO-RALG and other government departments. Together, PMO-RALG and MoEST, formed a project coordination team (PCT) to bring together relevant central and LGA- level officials in the areas of procurement and financial management, engineering, and M&E to ensure that different agencies of the government worked together in implementing project activities. Although there were some shortcomings with respect to clarity of roles and responsibilities of the different implementing partners, during the early years of project implementation, overall arrangements for implementation and monitoring were adequate. 8 31. The project was implemented in a context of major shifts in education delivery mechanisms. Secondary education had been decentralized to local governments10 in 2009, just before project approval. The decentralization led to several gaps in the capacity of the institutions at the district and school level which had not been adequately prepared for the additional responsibilities. The project helped fill these gaps by creating an enhanced and comprehensive structure for secondary education delivery through various interventions: (i) management training of the heads of schools; (ii) training of the heads of schools and education officers to ensure better utilization of the capitation grants; (iii) capacity building of the education officers and other relevant LGA staff such as the DSEOs, District Procurement Officers (DPOs); District Financial Officers (DFO); (iv) better coordination mechanisms between PMO-RALG and MoEST; (v) a fully functional EMIS for better monitoring; and (vi) Financial Management; Procurement and Safeguard training of the project staff and other relevant officers. The main project activities (construction, capitation grants) could not fully benefit from these efforts since they were still under process during implementation which is evident in the slow implementation progress more than half way through project implementation. Their impact began showing closer to the end of the project period when implementation picked up significantly and the majority of construction was undertaken and capitation grant delivery improved. These activities have, however, helped set a strong foundation for future interventions in the sector that can benefit from stronger systems for higher quality delivery. 32. Given the early issues with capacity of LGAs to manage interventions in the secondary education sector and other administrative reasons, several elements of the project were delayed by over a year in the initial stages. A project coordinator came on board more than six months after approval. Full MoEST and PMO-RALG teams were not appointed until August 2011, significantly affecting implementation timelines and coordination between ministries. This led to a delay in procurement for construction, capitation grant transfers, and training of teachers in the early stages. 33. The disbursement of counterpart funds also affected project implementation. After the first successful year of implementation of capitation grants, sufficient resources were not provided to LGAs to meet the target of 25,000 Tsh per student. Over the subsequent years, disbursements levels were volatile and did not reach schools at predictable intervals. This led to significant challenges for the schools as they were unable to plan in advance for procurement of teaching and learning materials. The delays and volatility was partially due to unpredictable amounts transferred from MoF to LGAs as well as delays in transfers from LGAs onward to schools. In some cases, LGAs did not transfer funds for several months. 34. Implementation was also impacted by the slow contracting process which acutely affected the civil works planned under the project. By the time of the Mid-Term Review in October of 2013, there had been minimal progress on construction. Main issues, mainly to do with capacity of LGAs and coordination between LGAs and MoEST were resolved soon after the MTR after which 10 Local governments were given the mandate to build school infrastructures and improve teachers’ living conditions, including providing regular salaries, housing, safe water, and reliable communication and provision of capitation grants. 9 construction picked up pace. In the three years after the MTR, construction in 792 schools was completed and the target for teacher residences was far exceeded. 35. A major development during implementation was the introduction of the Education Program for Results (previously known as Big Results Now in Education) supported by the Bank and other development partners. The EP4R provided support to primary and lower secondary education in Tanzania. SEDP was able to effectively utilize synergies with the EP4R. As an example, efforts in teacher development were coordinated for larger coverage for teacher training efforts. The EP4R, which follows a results based financing modality, has been successful because it promotes alignment of incentives between the local authorities and MoEST. By tying financing to performance, it has created accountability for results which has trickled down to multiple levels including local authorities. The approach is thus especially effective at delivering results in the context of decentralization. Some of the challenges SEDP faced in getting LGAs to deliver on the two main aspects of construction and capitation grant delivery have been resolved successfully through this mechanism. The complementary efforts of the EP4R also played a major role in accelerating progress in construction and capitation grant delivery under SEDP II. Given this more effective modality for getting results in the education sector, and more widely, there is strong justification for moving progressively toward results based financing modalities. In this context of better performance under a P4R modality in the education sector and others in Tanzania11, the Government requested moving funds allocated for the third phase of construction under SEDP II to the EP4R12. This course correction to pursue a more promising approach to achieving results was a sound choice and will lead to more efficient and effective achievement of results on the whole for the sector. 36. Throughout implementation, SEDP II maintained focus on building institutions and strengthening their capacity to undertake reforms and manage sector expansion effectively. A teacher service commission was developed and is awaiting approval. The establishment of this commission will institutionalize and promote sustainability of the teacher management efforts begun under the project related to teacher allocation and development. The Tanzania Institute of Education (TIE) was strengthened which has enabled it to more effectively fulfill its mandate related to curriculum development for secondary education. It has already worked on the development of a research instrument for needs assessment for the review of secondary education curriculum. The result of the assessment is being used for ongoing development of a new curriculum for secondary education. Also, TIE was able to review and develop 30 textbook titles for lower-secondary education and 12 for higher-secondary education. This enabled quick achievement for availability of prescribed textbooks for secondary education under the single textbook policy. 37. NECTA has successfully enhanced its capacity in ensuring quality of national examination assessment. NECTA trained its staff on SQL technology for database management, created a decentralized data center and conducted professional development on question items creation for 11 The follow on project for the Secondary Education Sector (under preparation) also proposes to follow a P4R financing modality. 12 The $23 million (approx.) canceled under SEDP II have already been approved by the Bank Board as Additional Financing for the EP4R. 10 320 secondary examination setters and 20 NECTA examination officers. NECTA also acquired ICT facilities which helped the institution to increase its efficiency in the marking of exams, data capturing during the marking session and processing and analysis of examinations results on time. Reliable information in student learning outcome is key to sustainable improvements in the sector. The strengthening of NECTA has also provided the secondary education sector with information that will allow it to successfully pursue a results based approach to improving sector outcomes. 2.3 Monitoring and Evaluation (M&E) Design, Implementation and Utilization 38. The results framework for this project was centered on a core set of indicators that measured changes in secondary school quality and equity. It also measured intermediate indicators to assess progress in infrastructure development, teacher deployment, availability of resources for the purchase of teaching and learning materials, and institutional capacity improvements. 39. The project emphasized using existing mechanism and strengthening them for improving overall M&E is the sector. Existing EMIS monitoring instruments were used to collect data on indicators listed in the results framework. The project supported the expansion and improvement of the existing EMIS, with the support of UNESCO, to better capture and help aggregate and analyze information related to secondary schools, and incorporate new indicators, and to conduct impact evaluation as needed. Two new sets of instruments were developed, piloted, and integrated into the EMIS: one collecting information on the use of resources for the purchase of teaching and learning materials and one for collecting and reporting information on teacher attendance. 40. Although this was a sound approach, some critical data to capture the project’s impact on achieving outcomes was not available. For instance, the EMIS did not disaggregate data on rural/urban and this was not available throughout implementation and at closing. Also, there were no mechanisms for collecting outcome data other than QTSR on the 500 hard to reach schools to assess overall impact on outcomes of these schools. 41. An impact evaluation was designed under the project to rigorously evaluate impact of performance-based incentives for teachers and schools. Its main objective was to guide the design and implementation of performance-based non-monetary incentives for teachers in the secondary education sector. The evaluation was completed by project end and found that incentives for teachers lead to modest improvements in student achievement across different subjects. In contrast, incentivizing students only does not produce observable learning gains13. An important finding of the evaluation that can hold lessons for future engagements in the sector is that public school teachers responded to incentives by focusing primarily on mid to low ability students. 42. Deficiencies with project indicators, such as the use of pass rates to measure student learning outcomes, were fixed during implementation. Data management, however, faced significant challenges. There were delays in recruiting consultants and once they were brought on board their management was problematic. The consultants, rather than working alongside the Government and contributing to their abilities and facilities, collected data from the LGAs directly and stored the data apart from Government data on a different system. The consultant then became 13 Filmer, Habyarimana and Sabarwal, ‘Students or teachers? The effects of giving and taking away incentive in public and private schools in Tanzania’, April 2017 (Draft) 11 an intermediary between the Government and its own data leading to delays in reporting and deficits in internal capacity. Over the course of implementation, several of these ongoing issues were addressed, however, missed opportunities due to delays negatively affected the project. 43. Utilization of data has been high in the sector and continues to improve. Data was used to target interventions on lagging schools and identifying vulnerable populations that required prioritized support and was a key objective of the project. Regional comparisons are also being used by the MoEST in planning provision of resources. Implementation of interventions under the project were also integrally linked to use of data, for instance for the provision of capitation grants. 2.4 Safeguard and Fiduciary Compliance 44. Safeguards. The project was assigned Environmental Assessment Risk Category B due to envisaged construction and rehabilitation of facilities such as laboratories, offices, and teaching blocks within compounds of existing educational institutions. The project triggered only one environmental safeguard policy, Environmental Assessment (OP/BP 4.01). Since all construction and rehabilitation activities were carried out within the footprints of the existing school premises, the project did not involve land acquisition issues, thus Involuntary Resettlement Policy (OP/BP .12) was not triggered and Resettlement Policy Framework was not prepared. 45. It was anticipated that under the construction and rehabilitation activities in Component 1 the potential adverse environmental and social impacts would include noise, dust, soil and water pollution from construction wastes, and traffic problems. To address these impacts and the fact designs and selection all projects sites were not completed prior project appraisal the Government prepared an Environmental and Social Management Framework (ESMF) to guide site specific Environmental and Social Assessment (ESIA) consistent with Tanzanian environmental laws as well as OP/BP 4.01. The implementation of the ESMF was the responsibility of the LGAs and MoEST and PMO-RALG were responsible for supervision. Adherence to environmental safeguards was rated as satisfactory during each Implementation Support Mission. During the two restructurings of the project the safeguard policies remained unchanged. 46. Financial management (FM): Overall, the FM arrangements of the project were noted to be satisfactory and adequate to support and manage the implementation of the project. There had been tremendous improvement in financial management and resolution of issues identified during the implementation. The project uses country systems in all aspects of financial management. Budgets are prepared by the project team and consolidated with the ministry budget for approval by Parliament by the Directorate of Planning. Funds are transferred directly to school bank accounts and monthly reports are prepared by each school and submitted to the LGA for consolidation. The accounting function is mainstreamed. Internal audit undertakes the review of the entire Ministry activities including the project activities. The external audit is covered by the Controller and Auditor General as mandated by the law. There were some issues with respect to eligible expenditures but these have been resolved. All audit recommendations of the previous years had been implemented by project closure. 2.5 Post-completion Operation/Next Phase 47. SEDP II benefitted from synergies with the EP4R which became effective while SEDP II was still under implementation. The EP4R, through a focus on many of the same results areas as SEDP II such as provision of capitation grants and teacher professional development, is helping 12 promote sustainability of many of the interventions started under it. Further, given the success of the results based approach to achieving significant gains in education outcomes in Tanzania, the government has requested follow-on support using a Program for Results (PforR) instrument to accelerate progress in the sector. The follow-on project intends to build on the institutional foundations set up under SEDP II and deepen the systems approach to achieving sustainable results in the sector. This is also in line with the original approach of SEDP II which was planned as the first phase of a longer term program of support to the sector. 3. Assessment of Outcomes 48. Overall, SEDP II performed well and contributed substantially to the most urgent priorities of the secondary education sector in Tanzania. Five of the seven PDO indicators exceeded the end target with total project beneficiaries at nearly 1.47 million, exceeding the revised end target by almost 170,000. Substantial contributions were made in terms of (i) improving the quality of secondary education resulting in improved completion at both the O and A levels; (ii) increasing participation in secondary education in underserved areas; (iii) improving the learning environment in secondary schools with better teachers, infrastructure and resources; (iv) increasing parity in terms of gender enrollment in secondary schools; (v) addressing the shortage of teachers in secondary schools in underserved areas (vi) increasing the capacity of government bodies to implement large scale projects such as the EP4R; and (vii) establishing a well-functioning and reliable Education Management Information System (EMIS) to inform monitoring and planning of reforms. 3.1 Relevance of Objectives, Design and Implementation 49. Relevance of the objectives of SEDP II is High. The objectives of SEDP II were highly relevant at the time of project approval and continue to be so today. Investments in secondary education remain crucial for improving the employability of workers, the productivity of enterprises, and the inclusiveness of economic growth.14. Tanzania has a large young population and has not yet been able to adequately capture the demographic dividend. Thirty-four percent of the working age population is in the 15-24 age range (ILFS 2014). However, only 15.7 percent has secondary education; and, only around 3 percent have some tertiary education. Consequently, there is considerable unmet demand for skilled workers in Tanzania’s growing economy. In the 2015 Tanzania Small Enterprise Survey, 54 percent of responding firms stated that attracting an adequately educated workforce was at least a moderate problem, with 14 percent stating that this was a major or very severe obstacle. 50. The project is also in line with the current education sector priorities of the Government of Tanzania. The project remains highly relevant to the secondary education portion of the Education Sector Development Plan (ESDP) and supports activities aligned with its main themes, including support for increased access and equity, improved learning environment of secondary schools through capitation grants and supply of teaching and learning materials; and improvements to 14The introduction to “The Right Skills for the Job?” by World Bank (2013), references literature that empirically supports the value of investing in education to develop human capital and on the contribution of education to growth and development. These include: Vandenbussche, Aghion, and Meghir 2006; Aghion 2008; Helpman, 2004; Hanushek and Kimko 2000; Krueger and Lindahl 2000; Hanushek and Woessmann 2007). 13 teacher motivation and qualification. The Government of Tanzania is in the process of finalizing its third Secondary Education Development Program (SEDP III) with three major priority areas: improving quality and relevance; enhancing access and equity; and improving management efficiency and good governance which continue to be in line with the objectives of SEDP II. 51. The objectives of SEDP II also continue to be closely aligned with World Bank Group’s Country Assistance Strategy (CAS) 2012-201515 and the Country Partnership Framework (CPF) 2018-2022 that is currently under preparation. Access to and quality of education were a part of key outcomes laid out in the 2012-2015 CAS which emphasized four strategic objectives: (i) promote inclusive and sustainable private sector-led growth; (ii) build infrastructure and deliver services; (iii) strengthen human capital and safety nets; and (iv) promote accountability and governance. The operation contributed to elements (ii); (iii); and (iv) of this strategy. The CAS also emphasized public service delivery as a critical area of focus for Tanzania’s on-going effort in reducing poverty and promoting shared prosperity. The 2018-2022 CPF proposes three main focus areas: (i) diversify growth and enhance productivity; (ii) boost human capital and social inclusion; and (iii) make institutions more efficient and accountable which lie at the intersection of the government’s development objectives, the country’s main development constraints as highlighted by the Systematic Country Diagnostic (SCD) and the Bank’s comparative advantage and capacity. The objectives of SEDP II are highly relevant to all three proposed areas under the new CPF. 52. The project’s focus on underserved areas16 is also highly relevant given the large disparities across regions in terms of availability of resources and learning outcomes. Completion rates at A- level in rural areas was only 1% compared to 6.1% in urban areas at the time of project approval. Further, the impact of the acute shortage of qualified teachers was compounded due to inequitable distribution. For example: for the 2009/10 school year Kigoma Rural district had only 48 science teachers, out of a total of 459 teachers (10.4 percent), while in Kibaha district in the Pwani region, 48 percent of secondary teachers had a science qualification. Many schools in rural areas had no qualified science teachers at all. Due to the inefficient distribution, QTSR in rural at 1:77 was twice that of urban at 1:34. 53. No changes to the PDO were made during project implementation so the rating for relevance of objectives was high both before and after the restructuring. 54. Relevance of Design improved from Substantial to High during implementation. The main issues that the project aimed to address to improve sector outcomes also remain relevant. The focus on continuous learning by teachers and school heads rests on a large body of research that shows that teacher quality is the main school-based predictor of student achievement17 and that 15 Report No. 60269-TZ, May 9, 2011 16 Underserved areas were defined as rural areas, where public schools have inadequate infrastructure and unfavorable student- qualified teacher ratios. For the APL 1 period, unfavorable qualified teacher-student ratio was defined as less than 1:40. 17 E. A. Hanushek and S. G. Rivkin. 2010. "Generalizations about using value-added measures of teacher quality." The American Economic Review, 100(2), 267-271; S. G. Rivkin, E. A. Hanushek, and J. F. Kain. 2005. "Teachers, schools, and academic achievement." Econometrica, 73(2), 417-458; B. Nye, S. Konstantopoulos, and L. V. Hedges. 2004. "How large are teacher effects?" Educational evaluation and policy analysis. 26(3), 237-257; RJ.E. 14 strengthening teacher’s professional development can improve learning outcomes 18 . Construction/refurbishment of secondary schools also continues to be highly relevant, especially in light of the additional capacity requirements associated with the Free Basic Education Policy19. 55. Project design addressed the key priorities of the secondary education to improve education quality. The design was informed by lessons learnt from previous sector experience and was guided by the analytical work and stakeholder consultations conducted by the World Bank and the Government (as discussed in section 2 above). Project design was cognizant of the capacity constraints of the sector, especially in light of the recent decentralization of secondary education management to local authorities. The choice to take a phased, long term approach to sector development was also sound and continues to be relevant as evidenced from the continuation of support to the sector. The design choice to focus on a few key interventions to make schools ‘fully functional’ while making investments to strengthen main education institutions was a sound approach to sector improvement and has set a strong foundation for accelerating sector improvement despite the large remaining agenda in the sector. The focus on strong M&E systems, which included impact evaluation to test ways of influencing teacher and student behavior for improved student learning in addition to strengthening monitoring of the sector were also strong design aspects that will help further strengthen sector performance. 56. There were minor deficiencies in design, related mainly to underestimation of the effort required for effective coordination between multiple stakeholders in a decentralized context. This led to initial delays (in construction and capitation grants) as the modalities of working between the national and sub-national levels were being worked out. The project could have benefitted from stronger incentives for better coordination between the different implementing arms. It could also have benefitted from more detailed consultations with relevant stakeholders. The subcomponent on teacher incentives through settling-in allowances faced challenges due to lack of adequate mitigation measures to address the political risk of implementing the program which led to the need to identify alternate measures to ensure equitable distribution of teachers. Lastly, the results framework did not collect and report adequate information to evaluate performance of a key element of the PDO, performance in underserved areas. Completion rates disaggregated by urban/rural were not collected and hence there was limited information on quality improvements in the key focus area of the project. 57. Minor adjustments to the project design were carried out in May 2015 at the time of the first formal restructuring. First, teacher settling-in allowances were replaced with furniture for the teacher residences constructed under the project which were politically neutral but still served as an incentive for teachers to relocate to hard-to-reach schools. Second, issues of capacity at the sub- national levels were resolved by instituting technical teams at the PMO-RALG to assist the LGAs in resolving issues that arose and coordination issues between the ministries were addressed Rockoff.2004 ." The impact of individual teachers on student achievement: Evidence from panel data." The American Economic Review, 94(2), 247-252. 18 World Bank 2013. What matters most for teacher policies: a framework paper. Washington, D.C.: The World Bank. 19 The Free Basic Education Policy introduced in 2015 makes basic education free and compulsory for all children in Tanzania with removal of school fees. This has resulted in large increases in enrollment at the primary level. 15 through concerted efforts including high level inter-ministerial meetings. Third, the issue of delays in capitation grants due to the multiple levels involved in the decentralized context were resolved by directly transferring funds from treasury to schools which greatly improved timeliness of transfers to schools from 2015 onwards. This was also promoted through coordination of incentives through the EP4R. The project added an indicator to measure performance in the lowest enrollment regions, however, this indicator did not provide complete information on quality in rural, underserved areas as one of the lowest enrollment regions was Dar es Salaam which is almost completely urban. Relevance of project design is thus rated substantial prior to the first restructuring and upgraded to high after the restructuring given that main design issues were mostly corrected. 58. Relevance of implementation is Substantial. The main factor influencing implementation quality was the weak capacity, especially at the local authority level and inadequate financing to meet capitation grant requirements as described in the sections above. This hampered implementation in the early phases of implementation with limited progress in both construction and provision of capitation grants. Efforts, however, were made throughout implementation to enhance capacity of the implementing agencies to improve the pace of implementation. This was done through a mix of solutions such as (i) continuous training of staff; (ii) contracting external expertise when required; and (iii) introduction of improved processes. For example, gaps in capacity at the LGA level were addressed through setting up of a technical secretariat at the PMO-RALG to provide support to the LGAs and increase the efficiency of the school rehabilitation related activities20. Lessons learnt from the first phase of school rehabilitation were incorporated in the subsequent phases to improve implementation performance which resulted in a significant improvement in the pace of civil works with 1,259 classrooms built in the last year of implementation (year 6). The continuous reflection on implementation performance and incorporation of measures to improve pace of implementation led to significant improvements in the second half of implementation. 59. Project activities under all components were completed successfully with the exception of the school construction component21. However, the decision to cancel implementation of the last phase of construction was a sound choice for the sector as a whole. There was recognition that results based financing is more effective at aligning incentives at the national and local levels for more effective and cost efficient implementation of key sector activities for producing results. The government and the Bank, following a portfolio approach, made a decision to divert resources to this relatively improved approach that was better suited to the operational environment of the country. Results based financing has also met with success in other sector in Tanzania and there is an increasing shift towards its use in the country. This decision led to cancelation of funds under SEDP II, but given the bigger picture, was the optimal choice. 60. Project performance with respect to relevance of implementation is rated substantial both before and after the restructurings. 20 Secretariat comprised of technical staff such as Quantity Surveyor; Architect and Engineer 21 792 schools were upgraded against a target of 1200. 16 3.2 Achievement of Project Development Objectives 61. The Project Development Objective was to improve the quality of secondary education with a focus on underserved areas. Performance in achieving the PDO was measured using seven indicators of which targets were surpassed for five22 by project end. While the project had a few shortcomings, the government was able to implement the majority of planned activities resulting in improved quality of secondary education in Tanzania, especially in underserved areas. The project contributed to increased participation and completion (main indicators for measuring quality) primarily by improving the learning environment in secondary schools through (i) more qualified teachers and their improved availability; (ii) better infrastructure; and (iii) increased resources and teaching and learning material. Finally, the overall education delivery system was enhanced for monitoring and delivering effective reforms. The following paragraphs evaluate performance against the two main parts of the PDO, (i) improvements in quality of secondary education; and (ii) improvements in underserved areas. A discussion on factors that contributed to the performance follows after that. 62. PDO 1: Improvements in quality: In Tanzania, formal secondary education consists of two sequential cycles; the first cycle is a four-year program (Forms 1-4) of Ordinary Level (O- Level) secondary education and the second a two-year program (Forms 5-6) of Advanced Level (A-Level) secondary education. Completion rates at these two levels of secondary education were used as the primary indicators for measuring quality improvements by the project. Completion rates are an appropriate indicator to use to measure changes in the performance of the system in providing a quality of education to its students. The project used the UNESCO definition for completion and calculated rates using NECTA and Census 2012 data, which the ICR team has also done for the last year. 63. The project also intended to use improvements in pass rates in core subjects (math, sciences and languages) at both these levels as a measure of improved quality. However, given the changes in the methodology for calculating 23 the pass rates, it does not provide a reliable means for determining improvements in student learning. First, the cutoff score for passing was changed annually making year to year comparisons inappropriate. Second, the pass/fail formula was changed from a flexible grade range adjusted annually to a fixed grade range requirement. In 2013- 2014, an expanded minimum grade requirement was added and in 2014/2015, the government changed their method for calculating grades. Because of all of these changes, the ICR does not use this indicator to evaluate project performance in achieving its PDO, although it was measured and reported by the project as an intermediate outcome indicator. 22 Targets for (i) completion rate at O level; (ii) completion rate at A level; (iii) QTSR; (iv) QTSR for hard to reach schools; and (v) project beneficiaries were achieved. Target for gross enrollment rate in lowest enrollment regions was nearly achieved and the proportion of public schools with minimum standards for infrastructure was not achieved. 23 While the difficulty of the examination questions varies little between years, the minimum score required to pass each examination changed annually. A given score may be sufficient to pass one year but not the next, making comparisons between pass rates between years inappropriate. 17 64. Completion rates at the O level increased significantly from 22% in 2009 to 37%24 by project closing in 2016 achieving the revised target (30%) and almost reaching the original target (39%). Both male and female completion rates increased dramatically over this time period. The package of critical inputs the project provided, specifically provision of qualified teachers, teacher development, provision of textbooks and funds for other teaching learning material in addition to improved infrastructure are expected to have translated into better learning outcomes of students with more reaching Form 4 (see paras 70-72). 65. These achievements are substantial, especially in view of the introduction of a new examination policy in 2012 which impacted the flow of students in the secondary school cycle. Completion rates had been increasing before the introduction of the policy (see table 2). Under the policy, the Ministry of Education of Tanzania set a performance standard in its Form 2 examination in an effort to improve pass rate performance in the Form 4 examination. The policy curtailed the flow of students to the next grade and ultimately impacted the number of students who would complete the secondary education cycle. As a result of this policy, promotion rates from Form 2 to Form 3, which were historically at the 90% level, declined sharply to 56.8% during the first year of the policy. In the following year, the promotion rate bounced back to 74.3 percent given that a relatively large number of students repeated Form 2 in 201325, increasing the size of that cohort after which it has leveled out. The policy further dictates that a student can only take the Form 2 exam twice, meaning that a large number of students would exit the system at that point. The policy impacted O level completion rates in 2014 when there was a drop in completion rates which had been increasing before then. Performance of this main project indicator was High before the restructuring which was undertaken to account for the policy changes. The project end target of 39% had already been achieved before the policy was introduced. Table 2: O Level Completion Rates  2009  2010  2011  2012  2013  2014  2015  2016  22  39.4  39.2  46.4  41.2  26.8  37.8  36.9  Source: World Bank calculations using examination registration data and Tanzania census data26. 66. Targets for both O Level and A Level completion were lowered once the impact of the policy became clear to 30% and 3.5% respectively. Nevertheless, completion rates at the O level far exceeded the revised target of 30% and nearly reached the original target of 39%. The improvements in completion despite measures to control student flow were quite remarkable and indicate that improvements in learning were larger than the project had estimated with a larger number of students meeting higher standards of learning than was expected at the time of project approval. The achievement after the restructuring is thus rated High for this PDO indicator. 24 Source: World Bank calculations using NECTA test registrant and 2012 census data for population projections. 25 The number of repeaters in Form 2 in 2013 was 71,301 compared to 5,263 in 2012. Data from Tanzania Basic Education Statistics (BEST) report. 26 The formula used by the government for calculations differs slightly. They use enrollment in Form 4 as the numerator vs number of test registrants. The completion rate reported in Government documents is thus slightly higher (at 38.6% in 2016) The ICR team used the formula used by SEDP from the beginning for comparison across years. 18 67. At the A level, completion increased to 5.4% from 3% in 2009, exceeding both the original target of 4.1% and the revised target of 3.5%. The completion rate for girls more than doubled from 2% in 2009 to 4.5% in 2016. Table 3: A Level Completion Rates  2009  2010  2011  2012  2013  2014  2015  2016  3.0  4.6  4.9  5.3  4.6  4.3  4.8  5.4  68. The cohort which reduced in size as a result of the new Form 2 policy reached Form 6 in 2016. This, however, did not seem to have an impact on A Level completion. A possible reason could be that students had traditionally been filtered in the O Level examination (Form 4) in any case, so the Form 2 exam policy would not have impacted the students who would have reached Form 6 (A Level). The quality improvements, however, are evident in the improvement in overall completion rate at this level. Again, given that the project was on track to achieve its original target before the restructuring and was able to achieve it after the restructuring, project performance in this area was High. 69. Overall performance of this PDO aspect is rated High given that targets were reached both before and after the restructuring for both O and A level for this primary indicator measuring improvements in student outcomes27. The comprehensive package of interventions provided by the project contributed to these achievements. First, the quality of teachers, which is the single most important factor for improving student performance was improved substantially. A total of 28,060 teachers were provided professional development under the project 28 . The in-service training of teachers under the project was targeting teachers hired in previous years who did not have adequate qualifications. 20,601 teachers were trained in-service on subject knowledge and 8,318 teachers trained on pedagogical methods through the Student Teacher Enrichment Program (STEP). This accounts for 56% of government secondary teachers in Tanzania in 2012 and is a substantial contribution to improving teacher quality. Second, recruitment of diploma-holding and graduate teachers as a result of increased enrolment in teachers’ colleges and universities in Tanzania has increased dramatically as a result of SEDP II. 12,600 were enrolled in pre-service diploma math and science programs under the project, of whom 3,991 have graduated. In 2012, 57% of the teaching force in government secondary schools was qualified. By 2016, this number has increased to 95% as a result of hiring of qualified teachers and intensive training of teachers to improve their qualifications under SEDP and EP4R. To promote sustainability of these efforts, the project supported the establishment of a semi-autonomous Teachers’ Service Commission to manage teacher affairs and a Teachers’ Professional Board to set professional standards and accreditation of pre-service teacher training programs.29This institutionalization of teacher training is expected to promote sustained provision of qualified teachers for the sector in the future. 27 In the ICR team’s opinion, the other PDO indicators (QTSR, infrastructure quality) were more intermediate indicators that measures performance of output rather than outcomes. 28 The total of number of teachers trained over the period of 2010 – 2016 was 34,200 if those funded by the Education Program for Results are included. 29 The Commission is currently awaiting approval. 19 70. Second, the project advanced the availability of teaching learning materials by ensuring (i) one book per student in Mathematics, Sciences and Languages as opposed to one book per five students in 2010; and (ii) 71% of the schools received funds to cover the cost of teaching and learning materials in schools. This is an impressive achievement considering the lack of timely transfers in adequate amounts in the past. Since December 2015 an amount of over Tsh12,000 per student per month 30 has been transferred to secondary schools. Further, going forward the government aims to improve the availability of resources for the schools under the Free Basic Education Policy. This outcome has been achieved through the complementary efforts of the Education P4R. SEDP II also ensured that there are clear guidelines for the use of the resources for the schools, and trained the Head of schools and relevant education officers in the effective use of resources. 71. Third, school infrastructure quality has also been improved substantially with support from the project. The proportion of public schools meeting the government approved standards and student teacher ratio of 1:40 reached 26% in 2016 compared to a baseline of only 4%. The annual school census 2016 shows that a total of 8,790 classrooms were constructed between 2010-2016 under the overall SEDP II. Of these, 2,634 were built using project funds. This has gone a long way in making sure adequate school facilities including laboratories and teacher residences are available for fully functional secondary schools, however, a large gap remains that needs further investments. 72. PDO 2: Improvements in underserved areas: One of the key objectives of the project was to improve outcomes in underserved areas which were rural areas that generally had lower teacher availability and resources than urban areas. Targeted interventions to improve the outcomes of these schools were built into the design of the project to address the large disparities focusing mainly on ensuring teacher availability for which teacher residences were built and furnished as an incentive. The project exceeded its target for number of residences provided in hard to reach areas by a large margin. As a result, QTSR in the 500 hard to reach schools, which were the worst off in the rural areas, at 1:17, improved more than it did at the national level (1:20) which can be attributed directly to the project since it was the only one focusing on equitable access of resources to these schools. No other partners provided support to the sector during SEDP II. It was the government’s main program for intervention in the sector and was supported solely by the Bank. Complementary support from the EP4R for lower secondary (Forms 1-4) began in 2015. 73. Data for the most rural areas of the country corroborates this trend of improved resources at secondary schools in these underserved areas. In the five regions with the highest proportion of rural population, STR ratios improved from an average of over 46:1 in 2010 for the 5 regions to 17:1 in 2016. Although improvements in student classroom ratios were not as dramatic, the trend is positive with average ratios improving from over 39:1 in 2013 to 37:1 in 2016. This data does not include the last phase of classrooms built under the project which would have contributed further to improving the availability of classrooms. 30 The remainder is retained at the central level for textbook procurement, which is now undertaken at the central level to benefit from competitive prices. 20 Table 4: Provision of resources in rural areas Rural  population  Student/Teacher  Students/Classroom     %  2010  2013  2016  2013  2016  Kagera  91  43  29  17  38  37  Manyara  86  50  24  15  39  37  Simiyu  93  49  28  18  44  39  Singida  88  47  25  17  33  35  Tabora  88  42  28  16  42  37  Avg    46  26.8  17  39  37  Source: Rural population data from 2012 census, education data from BEST 2013 and 2016. 74. On the outcome level, completion data disaggregated by rural/urban is only available until 2012. Between 2009 and 2012, there were substantial improvements in completion rates of rural areas, which improved from 21.1% (2009) to 38.8% (2012). Given that about 70% of Tanzania’s population lives in rural areas (2012 census), it is reasonable to expect that any large changes in the overall completion rate improvements would be driven by improvements in rural areas. However, given that disaggregated data was not collected by the government for the remainder of the project, information to substantiate this is limited. 75. In the absence of data on completion rates in rural areas, the project monitored gross enrollment rates at the O level in the five lowest enrollment regions to measure improvements in lagging areas. The interventions under the project were not targeted at these five regions so attribution is weak at best. However, improvements were recorded with enrollment increasing from the baseline of 22% in 2009 to 30.3% in 2016 meetings its target of 30%. All regions registered notable improvements with the exception of Tabora. Table 5: Gross Enrollment Rates in the lowest enrollment regions Region 2009 (baseline) 2016 (endline) Change Dar es Salaam 25 39.5 +14.5 Geita 20 32.5 +12.5 Katavi 16 25.5 +9.5 Rukwa 25 29.7 +4.7 Tabora 25 24.3 -0.7 Average 22 30.3 +8.3 76. Overall performance on this aspect of the PDO both before and after the restructuring was Substantial based on the data that is available to demonstrate improvements in rural areas. 77. Lastly, capacity of the education delivery institutions was strengthened to monitor and deliver reforms: The project developed GoT’s capacity for data collection, cleaning and 21 management for the Education Management Information System (EMIS) and helped develop improvement plans for the institutions managing curriculum, school management, student assessment, and textbooks. The project has financed eight technical assistance works of which four supported EMIS. The Tanzania Institute of Education (TIE) accomplished all institutional capacity building activities which enabled it to more effectively fulfill its mandate related to curriculum development for secondary education. This includes the development of a research instrument for needs assessment for the review of secondary education curriculum. The result of the assessment is being used for ongoing development of a new curriculum for secondary education. Also, TIE was able to review and develop 30 textbook titles for lower-secondary education and 12 for higher- secondary education. This enabled quick achievement for availability of prescribed textbooks for secondary education under the single textbook policy. Similarly, NECTA has successfully enhanced its capacity in ensuring quality of national examination assessment. NECTA trained its staff on SQL technology for database management, created a decentralized data center and conducted professional development on question items creation for 320 secondary examination setters and 20 NECTA examination officers. NECTA also acquired ICT facilities which helped the institution to increase its efficiency in the marking of exams, data capturing during the marking session and processing and analysis of examinations results on time. Finally, capacity development of ADEM was also conducted. 78. Establishment of a well-functioning and reliable Education Management Information System (EMIS) is a key feature of the project as it helped improve monitoring of the education sector in Tanzania resulting in targeted policies. This, along with the improved capacity of the institutions played an important role in the effective delivery of major education sector projects such as Education Program for Results and has laid the foundation for future engagement in the sector. 3.3 Efficiency 79. Efficiency of the SEDP II is rated as Substantial. There are extensive private and social returns to investment in secondary education which validate public sector intervention for its provision. In terms of actual implementation, the cost analysis showed mixed results. While the approach to provision of infrastructure to schools in this project was not optimal, significant cost efficiency was attained in some subcomponents such as the construction of teacher residences. 80. Economic Analysis: The economic analysis examines the returns to the project in terms of expected improvements in secondary level completion rates. The two main outcomes expected to be affected by the project interventions are higher completion rates at the O level and A level. The rationale of the project is that investments in secondary education have a positive impact on labour market outcomes. 81. The baseline completion rates of the project were estimated in 2010 at 22 and 2.6 percent for O- and A-level respectively. Rates improved during the project and by its end they achieved 38.6 and 5.4 percent in each level, an average annual increase of 10 and 13 percent. We estimate the gains in terms of the additional enrolment in contrast with what had occurred if the annual increase would have been half of what was observed. We assume that additional enrollment will be observed during the operation period of the program (7 years), and the benefits will stop right after the program with completion rates remaining at the same value observed the last year of the program operation. Other parameters used for the estimations are the lifetime earnings premium 22 of additional schooling (based on private returns) and labor force participation and employment rates. A 6 percent discount rate on lifetime earnings was used to reflect inflation in Tanzania. On these assumptions, the Net Present Value (NPV) of the project is estimated at US$53.9 million, the Benefits to Cost Ratio (BCR) is 1.5, and the Internal Rate of Return (IRR) is 5 percent, indicating that the project was an acceptable investment. 82. Sensitivity analysis assessed the impact of the program under two different scenarios. In Scenario 2 an increase in completion rates scenario it is assumed that the annual growth of the completion rates doubles. Scenario 3 maintains the completion rate achieved but uses a lower discount rate. Under scenario 2, NPV was US$142 million and the BCR was 2.4. Under scenario 3, NPV rises to 67 million and the BCR to 1.6. Table 6: Cost-benefit and Sensitivity analysis (in (USD millions) Scenario Assumptions NPV of net benefits Benefit-to-cost IRR (USD millions) Ratio (BCR) (%) (1) Main Estimation Goals achieved (Discount rate 6%) 53,944,348 1.5 5 (2) Increase Higher completion rates (Discount rate 141,868,725 2.4 14 Completion Rates 6%) (3) Lower Discount Goals achieved (Discount rate 3%) 67,068,474 1.6 7 Rate Source: WB staff calculations 83. There are additional social benefits of secondary schooling in Tanzania that justify public sector intervention. The GoT has an important role to play in secondary education provision because (i) secondary education attainment leads to behavioral changes that improve societal and workforce health, (ii) secondary education affordability in Tanzania is impeded by imperfect capital markets and liquidity constraints, (iii) educational attainment decisions do not fully reflect the intergenerational benefits accruing from the pursuit of good-quality secondary education, and (iv) there is a persistence of inequality of access to secondary education opportunities in Tanzania. 84. Cost Analysis: SEDP II shifted implementation responsibilities from school communities upwards to the LGAs. This was done because it was found that the quality of construction through the communities in SEDP I did not meet the minimum standards. Under the SEDP II model, the average unit cost per classroom, is about Tsh 26 million equivalent to US$ 14,200 in 2016, far above the initial estimation of Tsh 18 million in 2011. However, inflation during this period was high. Taking this into account, the initial estimate of Tsh 18 million is equivalent to US$ 11,450 in 2011 and to US$12,575 in 2016. Then the actual unit cost of the sample is only 13% higher than the initial estimate. This translates into a unit cost of construction of US$213 per m2. In addition, delays in the implementation of construction works also added to the cost of the works. 85. Considerable cost efficiency was achieved under the construction of teacher residences sub-component. The original design of the project planned a single two-bedroom house for the teachers’ residence. However, during the actual implementation it was decided that it would be much more cost effective to build low-cost multi-unit teachers residences (typically 6 unit residences). The unit cost of the flat was just US$ 14,000 compared to the estimated US$ 37,000 per house. 23 3.4 Justification of Overall Outcome Rating Rating: Moderately Satisfactory 86. Project relevance was substantial with high relevance of objectives and substantial relevance of design and implementation. Project performance in achieving the PDO of improved quality was High and Substantial with respect to improvements in underserved rural areas. Overall efficacy is rated High. The project led to an overall improvement in learning environment in secondary schools with improved teacher quality; efficient distribution of teachers; better infrastructure; and improved resources and teaching learning materials. It also strengthened the capacity of education sector institutions which will help improve the monitoring and implementation of reforms. Efficiency of outcomes was Substantial. 87. The assessment above attempts to evaluate performance both before and after restructuring of the PDO indicators. However, the split rating methodology was not applied to calculate overall scores and rating given that 100% of the revised amount was disbursed by the time of the restructuring of the PDO targets and under the methodology the weight for the period prior to restructuring would be 100%. This would not be appropriate given that 100% of the funds were not utilized at the time of the restructuring. In fact, some were refunded to the Bank to enable cancellation. Second, there were no major changes in project design or activities that drastically impacted project performance - PDO targets were adjusted in response to a change in government policy and project performance mainly remained the same. A single rating of Moderately Satisfactory based on Substantial Relevance, High Efficacy and Substantial Efficiency is thus given. 3.5 Overarching Themes, Other Outcomes and Impacts (a) Poverty Impacts, Gender Aspects, and Social Development 88. Investments in secondary education have compounded impact on poverty reduction through its effect on the overall economic growth of the country and through private returns of additional years of schooling. As mentioned above, recent analysis conducted by UNICEF shows that completion of the secondary education by a household head in Tanzania reduces the probability of being in a low income category by half (at 10%) compared to completion of primary education at 20%. 31 89. The latest poverty statistics suggest that 38 percent of Tanzania’s population lives below US$1.25 a day. This population benefited disproportionately from this operation given: (i) poor education quality affects the poor disproportionately; addressing quality had direct impact on equity; (ii) it is the poor who are especially harmed by poor service delivery and hence suffer from poor outcomes, hence improving education service delivery benefited them more; and (iii) the project focus on improving financial and human resources for schools in poor areas. For example, the project had a particular focus on social inclusion and emphasized increase in quality of poorer; hard to reach schools. Incentives through teacher residences were created to ensure that there is an adequate supply of qualified teachers in these areas resulting improvements in QTSR from 1:59 to 1:17. 31 UNICEF: The investment case for Education 24 90. The project aimed to address the needs of students in underserved areas, which includes addressing the needs of girls, female teachers, students with disabilities, teachers deployed to hard to reach schools, as well as hygienic school environments. The four project components addressed these needs through the provision of additional infrastructure and adequate sanitary equipment and water, which benefited all students but especially increased girls enrollment in secondary schools. By ensuring adherence to minimum infrastructure requirements, the project provided handicapped-accessible classrooms to enable the physically disabled to continue their education. Thus social inclusion and creating gender parity in attainment of education outcomes was an integral part of the project. While the gender imbalance still exists in terms of completion rates at the A-level, the completion rates for girls at O-level have increased from 20% to 33.3%. Similarly, the completion rates at A-level have increased from 2% to 4.5%. About 50% of the direct beneficiaries of the project were female. In addition, proper living arrangements also meant a safer environment for female teachers, thus making a job offer in hard to reach schools more attractive and helping create a more gender-balanced teaching force. 91. Lastly, by strengthening the EMIS and collecting data on student achievement performance, the Ministry is able to gather more comprehensive data regarding the student population with information on the most vulnerable students in the education system. This has helped the Government to create more targeted and effective programs. (b) Institutional Change/Strengthening 92. The project helped strengthen the capacity of the Tanzania Institute of Education (TIE), the Agency for Development of Education Management (ADEM), and the National Examination Council of Tanzania (NECTA)32 to deliver quality education and improve their linkages with the MoEST and PMO-RALG as described in the sections above. (c) Other Unintended Outcomes and Impacts (positive or negative) 93. Continuous engagement in the education sector of Tanzania helped create better coordination and synergies between various stakeholders and donors. This engagement helped leverage resources during the preparation of the Education P4R where a significant portion is funded by UK Department for International Development and Swedish International Development Cooperation Agency with USD 100 million and 30 million contribution respectively. Lessons learnt from the project have also informed the Education P4R. 4. Assessment of Risk to Development Outcome Rating: Substantial 94. The Government of Tanzania is committed to transforming the education sector. This is evident from the recently approved Free Basic Education Policy which includes abolition of fees and other costs at lower secondary education as well as all direct costs in primary education33. In addition, the government has announced its intention to make lower secondary compulsory beginning in 2021. Under this revised policy, examinations at the end of primary and in Form Two will no longer decide promotion. The effect of these combined policies is expected to lead to a 32EMAC was disbanded and its functions moved to TIE 33 While there is no fees under the policy, parents retain responsibility for purchase of exercise books, pens/pencils, sports uniforms, and other costs 25 significant increase in school enrolment, particularly in lower secondary, by 2025. Under these ambitious targets for basic education, the primary school population would rise from 8.30 million in 2015 to 11.53 million in 202534. Lower secondary enrolment would rise from 1.65 million in 2015 to 6.29 million in 2025. This increase in enrolment necessitates the need to increase school infrastructure over the coming years. However, there is a risk to quality of the provision of education while the government is focusing on increasing participation. 95. Other aspects of the project such as teacher training; teacher distribution and monitoring and release of capitation grants will continue to be addressed through the on-going Education P4R. The World Bank is also in the process of designing a follow-on project which will help the government build on the results of this project. In addition, it will help address the quality issues that may arise as a result of the massive increase in enrolment through the FBEP. 96. The project has managed to create sustainable structures that will retain the quality and supply of qualified teachers that would help maintain the results on learning outcomes. These include: (i) institutionalization of in-service teacher training; (ii) establishment of Teacher Service Commission is underway; (iii) capacity building of education delivery institutions such as NECTA; TIE and ADEM; and (iv) better monitoring of education outcomes through the EMIS. The strengthening of these institutions has improved the capacity of the sector to better manage the large influx of students which will help mitigate some of the risk to erosion in quality of outcomes at the secondary level as a results of the Free Basic Education Policy. 5. Assessment of Bank and Borrower Performance 5.1 Bank Performance (a) Bank Performance in Ensuring Quality at Entry Rating: Moderately Satisfactory 97. The project objectives and design were closely aligned with the CAS 2012-15 and the secondary education portion of the Education Sector Development Plan (ESDP). It supported activities aligned with the themes of the ESDP Program, including support to increased access and equity, improved learning environment through capitation grants and supply of teaching and learning materials; and improvements to teacher motivation and qualification. The components of the project were coherent, comprehensive and supported the PDO. The design was informed by lessons learnt from previous sector engagements and was guided by the analytical work and stakeholder consultations conducted by the World Bank and the Government. The team identified the risks and incorporated design features to mitigate them with the view to promote sustainability of project outcomes. Finally, a comprehensive results framework was designed to monitor the progress of the project. There were some deficiencies with respect to consultations with relevant stakeholders to check feasibility of some politically difficult reforms and an overestimation of capacity of the local governments to implement the construction components. There were some weaknesses in the results framework as well. Specifically, (i) pass rates were included as an outcome indicator even though they weren’t a reliable measure given the annual changes in criteria for passing; (ii) data sources and collection mechanisms were not thought through (rural/urban completion data is not available); (iii) adequate indicators to measure outcomes of the 500 HTR 34 World Bank projections 26 schools to which some project activities were targeted were not included in the RF. Overall performance is thus rated Moderately Satisfactory at entry. (b) Quality of Supervision Rating: Moderately Satisfactory 98. The Bank team worked closely with the Government to identify bottlenecks and find solutions that would lead to sustainable results. The overall Bank team, some of whom were based in Dar es Salaam, was able to build a strong working relationship with the Government and other development partners active in Tanzania at the time. When implementation challenges arose, the Bank team worked with the Government to devise appropriate solutions. The timeliness of actions was not optimal. The first restructuring of the project was processed more than 18 months after the midterm review although issues had been highlighted at the time of the MTR. But since the restructuring did not include any major changes to design or activities, project implementation continued while the restructuring was being processed. Project supervision benefitted from technical experts in key areas who worked closely with the implementing agencies to build their capacity in key operational areas, including financial management, procurement and safeguards. Continued supervision and engagement of Education specialists with the PMO-RALG and MoEST staff facilitated the project’s achievements with regards to teacher training; effective data collection through the EMIS and training of the Heads of School. 99. ISRs and aide-memoires were detailed and provided management with a thorough assessment of the project’s challenges and achievements. These reports also laid out specific recommendations to guide implementation of the project and to address challenges related to FM, procurement and M&E that arose during the course of the project’s implementation. The supervision teams consistently reported on financial management and procurement progress during supervision missions and worked with the MoEST and PMO-RALG to build their capacity in these areas. The Bank team also played a proactive role in ensuring the project design remained in alignment with the priorities of the government through the restructurings. By the project’s closing date, a majority of the envisaged outputs and outcomes had been achieved, if not exceeded. (c) Justification of Rating for Overall Bank Performance Rating: Moderately Satisfactory 100. Overall the Bank team was effective at building strong working relationships with the counterpart to effectively implement the project. It was also effective at identifying bottlenecks to implementation and working with the implementing agencies to resolve this or finding alternate solutions to achieve outcomes. Issues with the quality at entry (described above) were eventually resolved. There were issues, however, with the timeliness of the response. Overall Bank performance is thus rated as Moderately Satisfactory. 5.2 Borrower Performance (a) Government Performance Rating: Moderately Satisfactory 101. The Government expressed and demonstrated its high level of commitment to the achievement of the PDO which was well-aligned with its own strategic agenda for transforming the education sector in Tanzania. MoEST worked closely with the Bank team to design the project and to determine those elements for which Bank support would be most effective and appropriate. 27 The Government’s commitment to the project objectives was also ensured through the Letter of Sector Policy. 102. However, due to fiscal constraints, the release of funds by the Government had been unpredictable. By June 2015, the Tanzanian Government had distributed only 44 percent of the committed funds for capitation grants, whereas IDA’s contribution was almost double that number, reaching about 88 percent of its committed funds. While the government had put up the agreed resources for upgrading infrastructure, the volatility in availability of funds led to delays in implementation progress. There were also issues with procurement; and challenges associated with the M&E system and financial management. Further, there was lack of coordination mechanisms between the government departments especially at the LGA level which resulted in consistent delays in implementation. 103. Implementation progress picked up pace in the latter phase of the project and counterpart funding became more regular and predictable, especially in terms of transfer of resources to schools. Further, in FY 2016-17, the government increased its allocation for capitation grants which the average capitation grant expected to increase to over Tsh 25,000. Similarly, procurement and financial management issues were also addressed. Finally, monitoring and evaluation of the project was also improved through better functioning of the EMIS system. These achievements were critical to the project’s effective implementation. The Government remains committed to the objectives supported under this project and intends to build on the results of the program through a follow-on results based project to improve secondary education outcomes. (b) Implementing Agency or Agencies Performance Rating: Moderately Satisfactory 104. MoEST was responsible for the overall coordination and management of SEDP II, which was to be implemented in partnership with PMO-RALG. The project team also comprised of relevant central and LGA-level officials in the areas of procurement and financial management, engineering, and M&E to ensure that different agencies of the government coordinate to implement project activities. However, there were delays in appointment of full teams for the project by MoEST and PMO-RALG and issues in communication and clarity in terms of responsibilities between the two agencies. Further, the capacity of the project staff was weak in terms of procurement; monitoring and evaluation and financial management. There were consistent delays in submission of IFRs, progress reports and other relevant documents. The problem was more prevalent at the LGA level which hindered timely execution of project activities. 105. Eventually and with the Bank’s support, mechanisms were adopted to ensure adequate completion of project activities. This included: (i) enhancing the capacity of project staff for adequate reporting on program activities; (ii) issuing guidelines to LGAs for effective delivery and training the respective staff in monitoring and reporting of resources; (iii) improved coordination between relevant government departments; (iv) improving accuracy and reliability of EMIS data for monitoring project results; (v) safeguards, financial management and procurement training to relevant staff to ensure compliance with project standards; and (vi) wherever necessary alternative measures were adopted for smooth implementation for example: it was decided to directly transfer capitation grants to secondary schools from the central level and not through LGAs as previously, to ensure regular and timely flow of funds to schools. The MoEST’s efforts in ensuring improvements in the quality and distribution of teachers was critical in achieving the PDO targets. 28 106. While, the implementation progress of the project was slow at the beginning and it faced considerable challenges, mechanisms were adopted by the implementing agencies to ensure significant project activities were implemented by project closing and substantive progress made towards achievement of the PDO. In light of the above, implementing agencies’ performance is rated as Moderately Satisfactory. (c) Justification of Rating for Overall Borrower Performance Rating: Moderately Satisfactory 107. Overall Borrower’s performance is rated as Moderately Satisfactory. 6. Lessons Learned 108. Results based approach can be more conducive to gains in a decentralized context. The parallel implementation of SEDP II and the EP4R has highlighted the benefits of following a results based approach in a decentralized service delivery context. The PforR instrument has promoted an alignment of incentives between the national and local levels with strong emphasis on accountability which is producing results. Compared with the traditional investment project financing approach, which struggles with effective coordination between national and local level implementing partners, the results based approach creates an optimal environment for focus on achieving results. Under the EP4R, it has shifted the focus of dialogue between the MoEST and LGAs from inputs/outputs to that of results achievement and is cultivating a culture of performance based management. 109. A long term approach to service delivery improvement is well-suited for quality conscious expansion of services. SEDP II helped demonstrate that a phased approach that puts quality of provision at the center of sector expansion, especially in a capacity constrained environment, can produce impressive results. Under SEDP I, the sector had been expanded rapidly, but at the cost of quality. SEDP II’s focus on ensuring that schools met minimum quality standards and had the essential package of resources (qualified teachers, textbooks, teaching material resources) to be fully functional and provide quality education to students has helped improve efficiency through higher completion. 110. The early implementation experience of the project presents a strong lesson for ensuring readiness for implementation at approval. The project was slow to take off (by almost a year) due to administrative reasons such as hiring of key staff. Ensuring these are in place during preparation and before a project is negotiated can promote a strong start to implementation when the project becomes effective. 111. In cases where there are multiple implementing partners, it is critical to formalize implementing arrangements with clear accountabilities at the outset. There were coordination issues and lack of clarity in terms of tasks between MoEST and PMO-RALG which led to initial delays but were resolved fairly effectively through formal coordination mechanisms. The effort and time required to coordinate when there are several implementing partners with varied capacities is quite substantial and can facilitate implementation if addressed early on. 112. Consultations with relevant stakeholders are critical during the design phase. The project had to re-design a proposed incentive to attract teachers to hard-to-reach areas due to 29 political opposition which took considerable time and effort during implementation. Consultations with relevant parties and wide communication of objectives and activities of a proposed program during project preparation can help strengthen the design of appropriate interventions and mitigate the risk of resistance to implementation. 7. Comments on Issues Raised by Borrower/Implementing Agencies/Partners (a) Borrower/implementing agencies The Borrower prepared a detailed completion report which has been summarized in Annex 5. 30 Annex 1. Project Costs and Financing (a) Project Cost by Component (in USD Million equivalent) Actual/Latest Appraisal Estimate Percentage of Estimate (USD Components (USD millions) Appraisal millions) (of which IDA) (IDA utilization) (of which IDA) 469.3 215.5 Total Baseline Cost 46% (150) (119.1) Component 1: Upgrading existing 270.5 71.3 26% schools (192.2) (57) Component 2: Teacher distribution 27.1 17.6 65% and quality of teaching (27.1) (17.6) Component 3: Ensuring adequate 164.9 119.9 financing to secondary schools and 73% (37.8) (37.8) improving utilization of resources Component 4: Capacity building and 6.8 6.7 99% technical assistance for reforms (6.8) (6.7) Cancelled 23 Total Financing Required 150.00 119.1 79% (b) Financing Appraisal Actual/Latest Type of Percentage of Source of Funds Estimate Estimate Cofinancing Appraisal (USD millions) (USD millions) Borrower 319.30 96.4 30% International Development Association 150.00 119.1 79% (IDA) 31 Annex 2. Outputs by Component 1. Several aspects of Tanzanian secondary education were addressed under SEDP II. During APL I the focus was to initiate a transformation of Tanzania’s secondary schools by increasing the number of fully functional schools to provide a better quality learning environment and to ultimately help achieve the long-term program objective of improving student learning outcomes. SEDP II aimed to support the Tanzanian government’s long-term goal of reforming secondary education to improve learning outcomes among secondary students. There were four components and several outcome indicators associated with achieving this objective, these are discussed below. 2. There were three formal restructurings that took place, though no major changes were made to the project deign or activities. First, in May 2015, a new PDO indicator was introduced, one IO indicator was made a PDO indicator, and one PDO indicator was made an IO indicator. Second, in November 2015 several PDO indicators were assigned new targets. Discussions in this annex will use the final indicators and targets after the restructurings. Finally, in December 2016 some project funds, mainly related to school construction activities, were cancelled. 3. Component 1 – Upgrading existing school infrastructure (Total $270.5 million; IDA US$78.3). The original objective of this component was to increase the number of secondary schools that met the minimum infrastructure requirements. It aimed to do this by supporting rehabilitation and completion of existing school structures. The government had over 3,200 government-supported community schools. Many of these schools were missing some of the necessary infrastructure to operate effectively. The project would finance construction, supervision of works, and associated furnishing of (i) classrooms, (ii) science laboratories, (iii) teachers residences, (iv) latrines, (v) water supply, and (vi) electric supply (through the grid or solar power). The project expected that, on average, a package with teacher housing would cost approximately US$170,400, while a package without teacher housing would cost approximately US$133,400. 4. The selection of which schools would be rehabilitated was to be carried out by MoEST in collaboration with PMO-RALG with a focus on rural public schools, using objective criteria, specifically over-crowdedness (as measured by the student-classroom ratio) and student-teacher ratios. A needs assessment of the beneficiary schools was conducted by the LGAs to finalize the identification of infrastructure requirements and to develop appropriate procurement packages. Existing designs developed by MoEST were used for classrooms, latrines, boreholes, and electrical connectivity. For the science laboratories the project would finance multipurpose laboratories until additional, subject specific laboratories could be built in the future. The project was originally meant to include single-unit teacher residences, however, in May 2015 it was changed to fund multi-unit teacher housing. 5. There were three IO indicators to measure progress under this component. (i) Percentage of secondary schools meeting agreed minimum infrastructure standards. This indicator achieved 26% against a target of 37%. Progress under this component was meant to take place in three phases. There were severe delays in phases one and two and the Government and World Bank agreed not to continue with construction of phase three due to persistent delays and underperformance of school construction despite a 12-month extension. Overall, 792 schools were upgraded against a target of 1,200. (ii) “Hard to reach” secondary schools with new standard or multi-unit teacher residences. This target was surpassed due to a change from providing single- 32 unit to multi-unit residences. Overall, 1,290 teacher units were provided compared to a target of 500 standard houses because of the lower per-unit cost of multi-unit residences. (iii) Secondary classrooms built. The total number of classrooms built under SEDP II was 2,764 against a target of 4,800. This was due, again, to the significant delays experienced in procurement, design, and disbursement under Component 1. 6. Component 2 – Improving the equitable provision of teachers and the quality of teaching in mathematics, sciences, and languages (Total US$27.1 million; IDA US$27.1 million). This component was included to increase deployment of teachers to rural areas; to improve instruction in mathematics, sciences, and languages; and to improve teaching practices. In order to reach these objectives there were three subcomponents and associated IO indicators. 7. Subcomponent 2.1 – Improving equity in the distribution of teachers. In order to meet the government’s goal of improving distribution of teachers this component sought to create new teaching posts in schools with a qualified teacher-student ratio higher than 40:1. Originally, they sought to do this by using a combination of settling-in allowances and provision of housing in “hard to reach” schools. However, during implementation the settling-in allowance proved politically challenging as concerns were raised over civil servants who would not be receiving similar incentives. After the first restructuring the funds allocated for settling-in allowances for 7,500 teachers under APL I were instead used to purchase furnishings for new teacher residences. 8. Subcomponent 2.2 – Enhancing quality of instruction in mathematics, sciences, and languages. This subcomponent sought to enhance the quality of instruction by focusing efforts on teacher distribution and training. MoEST, PMO-RALG, and the LGAs were to monitor the supply of and requirements for secondary school teachers. They planned to collect data annually on student enrollment, teacher numbers, teacher attrition, and teacher training. This was meant to provide information yearly for teacher training facility requirements as well as teacher distribution. 9. This data was collected consistently during implementation. However, over the course of the project the questionnaire did not always have all information required by to make it a useful decision-making tool. For example, if a teacher was absent there was no differentiation between teachers who were on leave and those who were active in school. Many of these issues were resolved over the course of implementation and information about national teacher supply was available. 10. This subcomponent also financed enhanced science-teaching facilities at teacher-education institutions. This was done in order to allow teacher-trainees with hands-on experience with secondary school science practical experience by providing science laboratories identical or very similar to those in secondary schools. These materials were delivered to each of the eight teacher training colleges for use in teacher training. 11. Recognizing the mismatch between the demand for teachers in mathematics and science courses and the number of suitably qualified applicants, this subcomponent provided a bridging course for student teachers to boost their subject knowledge in mathematics and sciences. The bridge course was to be provided at each of the eight teacher-training colleges for up to 5,120 teacher-candidates or 1,280 per year. During implementation there were problems related to stringent entry requirements. The first and second cohorts had less than 300 candidates each. 33 Changes were made to allow Form IV students who passed two science examinations to enroll in a specialized track in Forms V and VI to acquire a one-year teaching diploma. 12. Subcomponent 3 – Improving the delivery of quality teaching and learning. This subcomponent sought to support management and training of teachers as well as contribute to the professionalization of secondary school teaching. MoEST planned to enhance its existing attendance management measures by preparing guidelines on teacher absence. Heads of public schools would also be required to monitor and record teacher attendance and to report monthly to district offices. There were some bottlenecks surrounding reporting of teachers with poor attendance. The attendance management improvements were eventually combined under heads of school trainings. These trainings were meant to develop leadership skills, mentoring abilities, integrate the enhanced teacher monitoring systems, improve expenditure reporting, and enhance awareness of key cross-cutting issues. This target was surpassed and over the course of the project 8,233 heads of schools were trained. 13. Continuous professional development of existing teachers was also supported in the form of in-service refresher courses that were to be taught during the school holidays. SEDP II provided for training for up to 30,000 secondary teachers of mathematics, biology, chemistry, physics, and English. Three cycles were planned to occur over three consecutive years with all 30,000 teachers completing all three cycles. Over the course of the project there were 28,91935 teachers trained. 14. SEDP II also supported the preparation of a roadmap for the professionalization of secondary school teachers through performance standards and a skills upgrading credit system. This faced significant delays in development due to delays in approval by the government. However, during the final year of implementation a semi-autonomous Teachers’ Service Commission was established to manage teacher affairs and a Teachers’ Professional Board was established to set professional standards and accreditation of pre-service teacher training programs. These institutions were awaiting parliamentary approval at project closing. 15. Component 3 – Ensuring adequate financing to secondary schools and improving the utilization of resources (Total US$164.9 million; IDA US$37.8 million). This component sought to ensure that adequate financing was provided to public secondary schools, disbursed timely, and monitored effectively. This was to be achieved through the provision of goods, technical assistance, and capacity building support. 16. Subcomponent 3.1 – Resources to schools. A capitation grant of TSH 25,000 was to be provided for the purchase of teaching and learning materials, including the equivalent of US$5 per student for textbooks. In order to allow government time to implement budget reallocations to achieve agreed-upon levels of secondary education financing under SEDP II, Bank financing of this component was front-loaded and was to decline over the years. By fiscal year 2010/11 more than 50% of IDA financing for this subcomponent was used. This was followed by a period of 35 This includes 20,601 trained in-service on subject knowledge by SEDP II and 8,318 teachers trained on pedagogical methods through Student Teacher Enrichment Program funded partly by SEDP II and partly by Education Program for Results. 34 unpredictable grant disbursement at inconsistent levels. This made it difficult for heads of schools to plan purchases. This, combined with issues surrounding textbook recruitment, led to two significant changes to capitation grant disbursement. 17. Initially, funds were to be released from the Ministry of Finance to the LGAs for onward transfer to local schools. Through this system, however, there were significant delays in disbursements with grants being held by the LGAs sometimes for months before onward transfer to schools. This made it extremely difficult for schools to plan purchases and anticipate incoming funds and led to such behaviors as hoarding textbooks due to a lack of certainty of when the next transfer would be received. In 2015 this was addressed by doing transfers directly to the schools from the Ministry of Finance. 18. According to the PAD the TSH 25,000 per student per year was to be transferred to the school for local procurement of teaching and learning materials. However, because each school was negotiating textbook prices there were often varying prices across schools for the same textbook. To address this issue, approximately half the capitation grant was given to PMO-RALG for textbook procurement and the rest was sent along to the schools. It is unclear, however, if this proved more a more effective strategy due to limitations in distribution networks. 19. As discussed earlier, capitation grant transfer levels were volatile and done at unpredictable intervals. By the end of the project the percent of schools receiving full resources (TSH 25,000 per student per year) was 71% against a target of 100%. The percent of schools that were receiving allocations on a timely basis, however, was 100%. Despite the challenges surrounding capitation grant transfers by project closing they were able to meet the targets for textbook procurement. The target textbook to student ratio was 1:1 and that target was met. There were 2 million mathematics, biology, chemistry, physics, English, or Kiswahili textbooks purchased against a target of 1.4 million. 20. Subcomponent 3.2 – Building capacity for effective grant management. This subcomponent sought to improve the capacity of the grant-managing stakeholders to manage the resources allocated under Component 3.1. It sought to do so by disseminating guidelines for grant implementation, tracking grant transmission, and monitoring of grant use. This subcomponent faced significant challenges, the effects of which were felt as described above. 21. Component 4 – Providing capacity building and technical assistance to implement current and future reforms (Total US$6.8 million; IDA US$6.8 million). The objective of this component was to strengthen institutional capacity for educational management. APL I was to provide support for the expansion and continued automation of the existing EMIS across the country. There was also a comprehensive capacity assessment to be conducted of TIE, ADEM, NECTA, and EMAC – key education institutions in Tanzania. After the assessment was completed, capacity building activities were to be undertaken in concordance with the assessment taken under APL I as well as other capacity assessments conducted previously. 22. EMIS expansion faced significant delays. By the time of the mid-term review in November 2013 only three LGAs (all in Dar es Salaam) were reporting their data electronically. In an effort to increase adoption of digital data entry by the LGAs the duties were transferred from MoEST to PMO-RALG. While doing this resulted in the successful introduction of digital data entry at the 35 LGA level, the way it was done introduced greater road blocks than even the previous paper-based system. PMO-RALG hired a consultant to collect data. This consultant developed his own software system and stored the school census data privately on his own servers. Although this satisfied the demand for digital data entry, it made the government dependent on the consultant for the data collection methods, software, and even data access. 23. The capacity assessment conducted under this component identified key areas of competency as well as measures that can be taken by each institution to strengthen their capacity. The comprehensive assessment was completed in May 2015 and recommendations were shared with relevant institutions. Many of the recommendations were implemented in the relevant institutions but the late completion of the comprehensive assessment as well as the lack of a clear action plan for all institutions resulted in uneven adoption of recommendations. 36 Annex 3: Economic and Financial Analysis 1. Analysis undertaken at the time of appraisal revealed investments in secondary education generate positive labor market outcomes. Estimates from Tanzania’s household budget surveys in 2000–01 and 2007 are that the private returns to secondary education at 13.5 percent are much higher than the returns to primary education at 5 percent. In tertiary education, private rates of return were 15 percent in 2007. On the margin, public spending on secondary education provides the greatest return. This is because enrollment in post-primary education is still low and labor market opportunities for primary school graduates remain quite limited. Recent analysis conducted by UNICEF shows that completion of the secondary education by household head reduces the probability of being in low income category by half (at 10%) compared to completion of primary education at 20%. 36The economic analysis examines the returns to the project in terms of expected improvements in secondary level completion rates. The two main outcomes expected to be affected by the project interventions are higher completion rates at O-levels and A-level. The rationale of the project is that investments in secondary education have a positive impact on labour market outcomes. 2. The baseline completion rates of the project were estimated in 2010 at 22 and 2.6 percent for O- and A-level respectively. Rates improved during the project and by its end they achieved 38.6 and 5.4 percent in each level, an average annual increase of 10 and 13 percent. We estimate the gains in terms of the additional enrolment in contrast with what had occurred if the annual increase would have been half of what was observed. We assume that additional enrollment will be observed during the operation period of the program (7 years), and the benefits will stop right after the program with completion rates remaining at the same value observed the last year of the program operation. Other parameters used for the estimations are the lifetime earnings premium of additional schooling (based on private returns) and labor force participation and employment rates. A 6 percent discount rate on lifetime earnings was used to reflect inflation in Tanzania. On these assumptions, the Net Present Value (NPV) of the project is estimated at US$53.9 million, the Benefits to Cost Ratio (BCR) is 1.5, and the Internal Rate of Return (IRR) is 5 percent, indicating that the project was an acceptable investment. 3. Sensitivity analysis assessed the impact of the program under two different scenarios. In Scenario 2 an increase in completion rates scenario it is assumed that the annual growth of the completion rates doubles. Scenario 3 maintains the completion rate achieved but uses a lower discount rate. Under scenario 2, NPV was US$142 million and the BCR was 2.4. Under scenario 3, NPV rises to 67 million and the BCR to 1.6. 4. General Parameters Unit of Year Source Measure 36 UNICEF: The investment case for Education 37 Exchange Rate 2,237.31 2016 http://www.xe.com/currencyconverter/ TZS to USD Inflation Rate 6% 2014 IMF, WEO, Serie PCPIPCH, Average Consumer Prices 5. Parameters to estimate benefits Unit of Year Source Measure Monthly wage for individuals 234262 2014 ILFS Analytical Report, with no education Table 10.1 Labor Force Participation Rate 86.7% ILFS Analytical Report Unemployment Rate 89.7% ILFS Analytical Report Returns to Schooling % 2014 Own’s estimates based using (Details in TZA ILFS 2014 Table XX) 6. Parameters to estimate costs Unit of Year Source Measure Total Cost of the Project US$ 120 2016 WB million 7. Regression Parameters and wage effects due to level of education Value Experience 0.08 Experience2 -0.001 Complete Primary 0.16 Complete Form 1 0.29 Complete Form 2 0.23 Complete Form 3 0.8 Complete Lower Secondary 0.8 Complete Form 5 0.35 Complete Upper Secondary 1.2 38 Annex 4. Bank Lending and Implementation Support/Supervision Processes (a) Task Team members Responsibility/ Names Title Unit Specialty Lending Supervision/ICR Sukhdeep Brar Senior Education Specialist GEDDR Bella Diallo Sr Financial Management Specia GGO25 Jane A. N. Kibbassa Senior Environmental Specialis GEN01 Gisbert Joseph Kinyero Senior Procurement Specialist GGO01 Anne Muuna Kisumo Program Assistant GFM03 Rehema Mercy Mashayo E T Temporary AFCE1 Franco Russo Senior Education Specialist GED02 Mercy Mataro Sabai Sr Financial Management Specia GGO31 Helen Z. Shahriari Sr Social Scientist GSU05 Stefanie Stallmeister Manager OPSVP 39 Annex 5. Summary of Borrower's ICR and/or Comments on Draft ICR The Borrower prepared a detailed completion report for the project to capture achievements, challenges faced and main lessons learned. The following is a summary of the main points raised in the Borrower’s ICR. Context at Appraisal 8. SEDP II was developed within the overall framework of the National Strategy for Growth and Reducation of Poverty (MKUKUTA) and the Education Sector Development Program (ESDP). These two documents outlined the long-term vision and outcome targets for the education sector in Tanzania and SEDP II was designed for the government to achieve these targets. The overall objective of SEDP II was to improve the learning outcomes among the secondary education students in keeping with Tanzania’s Education and Training Policy as well as the National Development Vision 2025. 9. At project appraisal the flow through the Tanzanian education system was very inefficient with a transition rate from primary to secondary education of 51.6 percent in 2008. Subjects taught were also not entirely relevant to developing priority economic sectors and schools were under- resourced and teachers were often poorly trained. 10. The World Bank assisted with SEDP II to address some of the challenges in the secondary education system. Institutional challenges and unmet needs were substantial and if they were not addressed they would put at risk the sustainability of the achievements of past interventions. Secondary education was also seen as crucial for successful development in other key sectors. The Bank’s involvement in the secondary education subsectors was an important complement to financing in other key sectors. The World Bank also has a comparative advantage in the education sector in Tanzania and has accumulated extensive experience through support of other education projects. Project Development Objectives (PDO) and Key Indicators 11. The APL PDO was to assist the government in implementing secondary education reforms to improve learning outcomes among secondary students. This PDO was meant to be achieved through three APLs with the outcome indicators: completion rates at the O level, completion rates at the A level, and student-qualified teacher ratios. The PDO of SEDP II APL I was to improve quality of secondary education, with a focus on underserved areas. 12. The results framework for monitoring and evaluating SEDP II was developed with a core set of indicators that measure changes in secondary school quality and equity measures. Intermediate indicators were also included to assess progress in infrastructure development, teacher deployment, availability of resources for the purchase of teaching and learning materials, and institutional capacity improvements. These were all to monitor progress toward achieving the PDO. Revised PDO 13. Some PDO and IO indicators were changed during two restructurings. Agreed changes made during the restructurings: 40 a) Changing one PDO indicator related to pass rates for mathematics into an IO indicator and updating its target values; b) Changing one existing IO indicator related to qualified teaches in hard-to-reach schools to a PDO indicator and redefining its measurement; c) Adding a PDO Indicator to monitor enrollment in underserved areas: d) Redefining two IO indicators for (i) hard-to reach secondary schools with a new teacher residence and (ii) management training for heads of public schools and updating their target values; e) Removing one IO indicator newly qualified teachers taking up posts in hard-to-reach schools and receiving a settling-in allowance that is no longer relevant to implementation; and f) Extending the project closing date by twelve months from December 31, 2015 to December 31, 2016. Main Beneficiaries 14. The primary target groups of the project were people and organizations working in the education sector in Tanzania. Key education institutions were also supported including TIE, ADEM, and NECTA to deliver quality education and to improve their linkages with MoEST and PO-RALG. The project also made more science and language teachers available. It increased the capacity of MoEST and PO-RALG through building management capacity at the central and state levels. Project Preparation, Design, and Assessment of Risks 15. The project benefited from in-depth and fairly consultative preparatory processes supported by data on the sector available at the time. Implementation and evaluation of SEDP I informed the preparation of SEDP II. Stakeholder involvement at project preparation was strong but this was primarily at the central levels of MoEST, PO-RALG, and in the national oversight agencies. It did not involve extensive collaboration at the school, LGA, and regional level. 16. Implementing actors of SEDP II acknowledged the objectives, components, and organizational modalities were fairly relevant and realistic. The major criticism raised by several stakeholders were regarding the top-down process that was used to identify needs or issues that needed to be addressed. 17. Assessment of risks were well reflected at project appraisal. However, there were serious challenges that were not identified and therefore mitigation strategies were not identified. These include bureaucratic hurdles with government and World Bank procurement procedures which caused conflicts and substantial delays. The government also failed to make sufficient and timely financial contributions. Timelines that were agreed at project appraisal did not accurately allocate time needed for preparatory activities like procurement, carrying out condition surveys, and completing needs assessments which resulted in delays. Implementation 18. Overall implementation of this project was moderately satisfactory. Below is a summary of key facilitating and constraining factors identified during desk review and consultations with stakeholders: 41 Facilitating Factors:  Establishment of the national program coordination Unit has brought the link between the two Ministries-MOEST and PO-RALG.  Close monitoring according to World Bank and Government requirements. In particular, the project required periodic reporting which ensured accountability and responsibility.  Good adaptable leadership especially the Principal Secretary (PS). The flexibility and adaptability of the current (new) leadership has made it possible to make quick decisions for important issues affecting the project.  The instrumental support from the World Bank particularly the World Bank Education specialist has been enormous and helped to resolve some issues which required extra support.  The adaptable SEDP II team has played key role in pushing the SEDP II agenda forward. Constraining Factors:  In some cases, there were issues not well communicated by the Bank to the MOEST e.g. the construction specialists were deployed to SEDP II projects with his TORs were not well known to the Ministry.  Duration of pre-construction activities were not adequately considered (e.g. drawings, procurement, etc.) so projects had to start in 2012 instead of 2010. As a result, this caused delays as the duration set for construction became very short.  Weak Education Management Information System (EMIS) negatively affected planning for teacher distribution and resources allocations (including the scrapped settling in allowance and professional courses allocation).  Government has not been able to honor its contribution and as a result some activities could not be implemented as planned  There were general delays in procurement partly due to World Bank no objections policy, re-tendering due to perceived low bid offers by the bidders at the council level and other reasons. In some cases, there has been a conflict in applying either World Bank or Government procedures. M&E Design, Implementation, and Utilization 19. The design of the M&E systems were built around the project results matrix which included several PDO and IO indicators. The PDO and IO indicators were sufficient, relevant, and reasonably linked. The design of the M&E system did not allow for real-time updates and reporting. There was also no money for M&E provided at the LGA level. The design assumed that a lot of M&E system development had already been in place which was not the case. The project also suffered from the lack of a dedicated M&E officer. 20. Data collection was done through the EMIS at the district level. It was not until 2015/16 that this data was being collected reliably. The effectiveness of EMIS is limited by lack of adequate funds for monitoring, supervision, and data auditing. There is also an inability to have direct data entry into the system at the school level. The M&E systems are being used at a moderate level. Evidence of utilization is found through the development of the National Five-year Plan, development of SEDP III documents, and the development of internal MoEST, PO-RALG, and LGA plans. 42 Safeguards and Fiduciary Compliance 21. The project adequately addressed the needs of girls, female teachers, and students with disabilities. At appraisal SEDP II was classified as a low risk intervention with regards to environmental safeguards. The government failed to complete an environmental audit of the first phase of 264 schools. An environmental management review was done afterwards which identified areas for improvement. A major lesson from SEDP II environmental and social safeguards was the proper environmental and social audits need to be done early for any construction project. 22. Procurement under SEDP II were time consuming and faced many delays. Reasons for this included restraints on use of funding by the World Bank and high turnover with leaders at MoEST and PO-RALG resulting in loss of institutional knowledge. Requirements regarding contract approvals also caused delays in procurement. 23. SEDP II benefited from fairly stable staffing on the financial team allowing for up-to-date audits and timely submission of financial reports. However, ineligible expenditures by LGAs caused FM issues. Lessons drawn here show that projects involving more than one ministry are susceptible to fiduciary challenges and proper mechanisms must be put in place to deal with them. 24. Slow, inadequate, and unpredictable release of government funds contributed to significant delays. Assessment of Outcomes 25. All four components were complementary and directly linked to the PDO. The government notes significant progress with IO indicators that is positively correlated with progress noted with PDO indicators. The project contributed to increased enrollment, completion rates, pass rates, and to improved availability of qualified science teachers. It also improved the quality of the teaching and learning environment as well as the strengthening of secondary education regulatory agencies. 26. The project was able to exceed targets for six out of seven PDO indicators. The indicator that did not achieve its target was the proportion of public schools meeting government approved standards and student teacher ratio of 40:1. Although it was not achieved there was substantial progress made from 4% of schools at appraisal to 26% at project closing. Component One: Upgrading Existing School Infrastructure  A total of 264 schools were upgraded during phase one and 528 during phase two. This increased the number of secondary schools meeting agreed minimum infrastructure standards. It also increased numbers of hard to reach schools with new standard or multi- unit teacher residences. There was also an increase in the number of secondary classrooms built.  These construction activities afforded greater access to laboratories, increased teacher motivation through staff housing, increased access to electricity, availability of clean and safe drinking water at schools, reduced congestion in classrooms, and anecdotal evidence of increased interest in science among students.  This component would have benefited at design stage from clear understanding of how government funding would be allocated since it comprised 79% of the funding under component one. 43 Component Two: In-service Training of Teachers  This component was quite successful and had substantial positive impacts on delivery of teaching and learning in Tanzania.  The success of implementation of this component was reflected in IO indicators of pass rates, teachers trained, and management from the heads of schools. Component Three: Ensuring Adequate Financing (Provision of Capitation Grants)  Stakeholders consulted during evaluation applauded the inclusion of capitation grants in program appraisal. However, uneven financial commitment from the government resulted in a general decline in per-student provision of funds during project implementation.  Capitation grants contributed to the ability of schools to purchase teaching and learning materials. The ability of schools to procure materials led to improvements in completion and pass rates in science subjects.  Through the first several years of implementation problems with delayed release of capitation funds from the LGAs hindered ability to procure materials. These problems were resolved during the final year of implementation. Component Four: Capacity Building for Sustained Reforms  An integrated EMIS with data on secondary education was completed and operational in the LGAs and a capacity assessment for educational institutional development were both completed satisfactorily during implementation.  This component contributed toward improved availability of statistics; improved strategies for capacity building at relevant educational institutions; improved environment for quality teaching; and, improved information and knowledge base for future planning and decision making.  The project would have benefitted from early completion of the institutional capacity assessments. Data entry systems should have also been implemented at the school level where possible. This would have increased data entry efficiency and quality. The project also would have benefitted from set outcome indicators to assess TIE, NECTA, and ADEM. 27. Based on analysis of views of various stakeholders at all levels there was a consensus of the overall success of the project in meeting PDO and IO indicators. Factors contributing to this success include:  The improvement to project coordination with the appointment of a national project coordinator.  Close monitoring and evaluation especially during the last three years of the project.  The pro-active and supportive leadership demonstrated by both the PO-RALG and MOEST Permanent Secretaries.  The constructive support provided by the World Bank (led by the Educational Specialist) which enabled timely and effective resolution of crucial issues.  The swift adaptability of SEDP II teams despite major turnover in their composition. Factors identified by stakeholders which constrained progress were: 44  The tendency of some World Bank commissioned consultants to merely focus on what was not working and embarking on a blame game instead of offering constructive feedback.  Failure to adequately factor in pre-construction activities in the planning and to manage them effectively for timely delivery of construction objectives.  Existence of a very weak EMIS at the onset and for four years of project implementation which, due to poor/inaccurate data, resulted into misalignment between teacher provision needs, teacher availability, and resources. The lack of timely and accurate data also weakened the quality of strategic decisions. However, these problems were generally overcome with the modernization of EMIS through SEDP II support.  Absence of explicit and sustainable factoring in of community involvement in SEPD II design. SEDP II – unlike to SEDP I, did not heavily and overtly leverage community involvement particularly in the upgrading and construction of infrastructure.  Failure of the Government to fully honor its financial contribution in SEDP II in a direct manner. The risks associated with the 71% expected Government contribution for Component 1 were not adequate assessed at the design stage and suitable mitigation measures worked out.  Huge delays especially with Component 1 due to cumbersome procurement procedures and inability to foresee the time implications of the various preparatory activities like condition surveys. There was not adequate assessment of the extent to which the obtaining public procurement regime was a risk to the project and workable strategies formulated and implemented. 28. The overall rating for performance of SEDP II is moderately satisfactory. The project was able to upgrade the infrastructure of 792 schools across the country. It completed in-service training of 9,903 science, mathematics, and language teachers. The EMIS system was upgraded and expanded to accurately and reliably inform monitoring and planning. Student performance in science and language subjects improved. There was increased enrollment of pre-service math and science teachers to address shortages in underserved areas. There was also increased capacity in key agencies in the secondary education sector. Assessment of Risk to Development Outcome 29. Risk to development outcome is rated moderate. The government is committed to transforming the education sector. It has been improving the recruitment and training of secondary school teachers with a focus on sciences to nurture industrialization and human development. The government has institutionalized in-service teacher training resulting in regular in-service trainings. The Teacher Service Commission will strengthen and align teacher policies, procedures, and systems. The government is promoting teacher professionalization through the teacher professional board. Capitation grants are being mainstreamed in the government’s budget. Reforms introduced under SEDP II of the regulatory agencies will continue to be implemented with positive impacts on quality in the sector. Development outcomes face the risk of the availability of adequate and sustainable government funding to support necessary investment in secondary education. Assessment of Bank and Borrower Performance 45 Bank Performance Ensuring Quality - Satisfactory 30. The World Bank team conducted technical discussions with management and staff of MoEST, PMO-RALG, TIE, ADEM, NECTA, and EMAC. These meetings were to review implementation progress on key project activities, trouble shoot, and help streamline implementation mechanisms. The Bank maintained sound monitoring arrangements according to Bank and government requirements. Support was given by the World Bank team in operationalizing the results framework monitoring. Quality of Supervision – Satisfactory 31. The World Bank maintained close supervision making sure that fiduciary compliance was achieved by SEDP II. It conducted training for relevant government agencies. Supervision was timely, regular, and technically sound with supervision team reachable by key stakeholders. Government Performance – Satisfactory 32. The government demonstrated its commitment to SEDP II through implementation of Big Results Now (BRN) which was run concurrently with SEDP II. BRN allowed for timely transfers and adequate resources for the overall project. Implementing Agencies Performance – Satisfactory 33. MoEST and PMO-RALG were primarily responsible for the implementation of SEDP II. Toward the end of implementation many of the issues had been resolved. However, it took MoEST over six months to appoint a project coordinator and assign a team for SEDP II. There was also a lack of coordination between MoEST and PMO-RALG. Lessons Learned Project Design, Implementation, and M&E  The major criticism raised by several stakeholders especially at the level of regulatory agencies, schools and LGAs is that the identification of needs or issues to be addressed was rather top down as opposed to being a thorough and inclusive bottom up process as well.  Correct setting and understanding and management of the Government contribution to a project is crucial for success. Failure can result in delays and poor performance  Community participation, involvement and contribution in the education sector can be quite strategic if well appreciated and considered. SEDP II did not do well on this aspect.  A project M&E system should comprehensive and superior – providing “real-time” project monitoring and enabling wide ranging formative and summative evaluations to assure effective continuous steering and strategic outcome monitoring and evaluation. SEDP II was strong at results monitoring but very weak in real time monitoring. It is no wonder it suffered so many operational delays and challenges.  Proper and vigilant concerns with managing safeguard and fiduciary compliance issues should start right from the design stage through closer. Unfortunately, it was erroneously assumed that SEDP II would have negligible social-cultural environmental issues and challenges and the first phase of 264 schools was undertaken without having ESMPs and ESIA prepared. 46  Current public procurement regime is not quite suitable for infrastructure construction at the school level. It causes a lot of delays. The special circumstances of schools need to be studied so that PO-RALG working with the relevant authorities can come up with more befitting procurement procedures so as to speed up procurement processes for projects aimed at schools.  The way Government funds projects need be well captured in reflect in project management. There was misconception regarding SEPD II funding by the Government. As explained elsewhere in this report, the government did not fund directly. Therefore, it contribution to SEDP II was actually much higher that reflected in the progress reports.  Real time EMIS is central to supporting project monitoring in the sector. To make it very functional, there is need to quickly and gradually move to getting data to be entered in real time at the school level using computers and tablets.  Effective overall national project coordination supported by a strong and dynamic national project coordinator is crucial for project success.  Close monitoring and evaluation of a project with a focus on both continuous progress monitoring and action and strategic outcome monitoring and evaluation ensures high likelihood of project success.  The pro-active and supportive leadership demonstrated by both the PO-RALG and MOEST Permanent Secretaries.  Existence of strong constructive support by World Bank contributes to project success. SEDP II benefitted a lot from this factor. Component-level  Failure to factor in and manage pre-construction activities in the overall project schedule caused undue delays and problems.  The risks caused by the public procurement process and procedures were not properly analyzed priori and strategies to prevent delay worked out. There could have been more success had the project used vetted local contractors under force account modalities.  Involvement of direct beneficiaries of the project at all stages could give better results.  Needs assessment should involve the beneficiary schools to select the projects based on their prioritized requirements.  Harmonization of different projects implemented in the country to achieve similar objectives would ensure synergies and complementarities in the resource use.  The multi-unit teacher houses are more beneficial than the traditional one teacher house.  Outsourced expertise ensures quality of output but it can be expensive and delaying in the current procurement regime.  Under situation of slow funding from the Government – it is better to start with components that can finalized fully for the money that is going to be available for certainty.  Procurement is a major source of delays if not well factored in the planning of a project.  In the future, the ministry (MoEST) should make use of specialized environmental staff to handle safeguard issues instead of assigning the role to other staff with insufficient expertise on safeguards. 47  The demand of science, mathematics, sciences and English teachers remains a crucial challenge - due to the relatively few teachers available and the huge demand for such teachers.  The budgeting for capitation grant should be in USD to meet the constantly fluctuation of Tsh against USD.  Government adherence to set financial obligations is has been a major challenge which affected implementation.  Since education transformation is costly, the SEDP II resources were rather spread too thinly to make a huge impact. If we want to have quality education, then we have to invest more. If you have good/nice schools’ students will love going to school even study harder. In relation to this, good/conducive environment attracts teachers and students too. Furthermore, teachers who regularly attend in-service training seem to be better motivated than the rest. Teachers housing key in attracting and motivating teachers. When the living, learning and teaching environment (teachers, laboratories, chemicals.) coupled with in- service training performance in science subjects improves a lot.  The WB should think about funding the Education Development Institutions such as ADEM, TIE, NECTA etc. after receiving a comprehensive needs assessment from the relevant institutions. This will reduce delays in disbursement of funds as well as meeting the critical needs of the institutions.  The demand for e-learning is imperative and there is a need to have special component of e-learning the future projects. This will increase efficiency, access and create more space for teaching.  Capacity building should be sustainable meaning that it should be not done as one off event. For example, the curriculum skills of teachers must be constantly be refreshed to meet the demand of teaching and learning in Tanzania. Project Management  Clarification of roles between MoEST and PO-RALG in the implementation of SEDP II was highly required to avoid unnecessary duplication of functions.  Participation by all stakeholders is important for maximum success. Thus, a comprehensive orientation and re-orientation of stakeholders on project components and issues is crucial.  Community involvement in the project is crucial at all stages. There should be created an atmosphere under which the surrounding communities will develop a sense of ownership over the schools in their localities.  To a large extent SEDP has been relatively centralized in decision making and planning with minimum involvement of the communities actually own the schools and therefore have to play big role in sustainability.  Cooperation between different actors in the sector improved matters. As noted by REO, Dar “The involvement of engineers, PMU staff, teachers and parents’ cooperation made realization of targets was easy. Before SEDP II, the saying used to be – ‘that is Education Officer’s job.’”  Prioritizing and phasing project activities is crucial in implementation success. That is why SEDP II was successful. 48  Timely availability of resources can affect the outcome of a project.  Delays in a project can be reduced by ensuring that money is directly disbursed to schools to ensure closer supervision and other logistics which could facilitate implementation  Capacity assessment and building focused on key implementing actors should have been conducted early in the project in order prevent implementation challenges. Sustainability  Sustainability of SEDP II outcomes with regard to pass rates at O and A levels; O-level completion rates; improvement to school infrastructures; availability of qualified teachers in Science subjects, Mathematics and English; timely availability of school resources; and availability of disability friendly school environment and facilities. However, sustaining the positive results of SEDP II calls for more dedicated efforts on the part of the Government and stakeholders in the education value chain. Unmet Needs  There are other crucial unmet needs including: school feeding; dormitories for girls; training for both arts and science teachers; construction of school administration blocks for teachers and management staff; fencing of schools; ICT systems to enhance computer- mediated teaching and learning; teacher housing; classrooms both O and A levels considering the need for aligning the pipeline from primary – lower secondary – upper secondary & higher education; teacher incentives and remuneration; water supply and health services to schools; school transport; modern library with textbook availability; ICTs facilities for teaching and admin purposes  School boards should be more involved in identifying the needs of school, monitoring and evaluation and improve procurement procedures. 49 Annex 6. List of Supporting Documents 50 51