Skills Training and Youth Employability (ID: P169268) Central African Republic (AFRICA) i Executive Summary 1. The objective of the report which is in two phases, is to explore the overarching factors in the Central African Republic (CAR) that contribute to the out-of-school youth phenomenon with a view to providing recommendations for improving existing policies and programs. The report focuses on the 15- 24 years old cohort in particular because there is limited information on: (i) why youth drop out of school without acquiring basic literacy and numeracy skills; (ii) skills training provided in schools and training centers; iii) the quality and relevance of the curricula provided at technical and vocational schools and training centers; and (iv) the relevance of the skills acquired by the graduates of technical and vocational schools. 2. The larger report will be based on work done in two phases. The current report covers Phase I where the authors: (i) analyze the magnitude and nature of out-of-school youth by socio-economic characteristics; (ii) provide a diagnosis of the supply and demand side determinants of out-of-school youth; and, (iii) propose relevant policy actions for policy dialogue in the short to medium term. Phase II will be developed based on key findings from Phase I including a skills training roadmap that assesses formal/ informal labor supply and demand structures and provide actionable recommendations for implementation combining both the findings and recommendations in Phase I and II. A summary of the main findings and policy recommendations stemming from the analysis carried out in Phase I is provided below. Key Findings Prolonged conflict has severely affected the education system making out-of-school youth a product of the “broken system� 3. The 2013 crisis which was a civil war and a humanitarian crisis worsened the situation in CAR which in itself was alarming before the crisis. Insecurity and conflict have led to the collapse of state institutions, particularly the state’s capacity to provide basic service delivery and respond to the population’s urgent needs. Not surprisingly, the education sector was also profoundly affected by the crisis-- school facilities were looted or destroyed and thousands of children lost years of schooling. The crisis increased gender inequality in terms of access to education and the vulnerability of girls to exploitation and abuse was even made more acute. The sector is poorly financed, and the government has limited capacity in providing and retaining primary school teachers as 51 percent of public primary schools have no government-provided teachers. The provision of school infrastructure is also extremely weak and existing classrooms are overcrowded, with average class sizes of 95 in primary public schools and even higher in public secondary schools at 124. A large proportion of the youth population is inactive and lack numeracy and literacy skills 4. An estimated 660,000 youth (i.e. 60 %) in the CAR are not in school and a large share remain inactive and register low levels of educational attainment. Most out-of-school youth leave the education system with low levels of education and without acquiring basic foundational skills. Many youths start ii school in CAR, but the majority drop out at an early age and in fact, delayed entry could be one of the factors contributing to early dropout. In fact, 70 percent of the out-of-school youth are dropouts. The level of inactivity is generally higher among young girls, youth from poor family, and those living in urban areas. Out-of-school youth tend to be girls and those from disadvantaged social groups 5. The out of school youth population is made up disproportionately of young girls, youth living in rural areas, and those from economically disadvantaged groups. The out-of-school rate among young girls is higher, standing at 70 percent; as opposed to young boys, who are at 49 percent. Similarly, the out-of- school incidence affects youth living in rural areas (69%) more than those living in urban areas (48%). Likewise, the out-of-school incidence is relatively higher among youth from poorer households than those from wealthier households. This further reinforces the existing inequalities and hinders the poverty reduction agenda of the country. Moreover, the out-of-school rate varies largely across regions as some regions are more affected by the conflicts than others. The out-of-school rate is lower in region 7 (Bangui) standing at 35 percent; while it reaches 70 percent in region 4 (Kagas). Key determinants of out-of-school youth vary by region, gender, area of residence and wealth quintile 6. Family refusal, cost of education, distance to school and being too young are the principal reasons cited for out-of-school youth who have never been in school whereas the cost of education, failed examinations, insecurity and pregnancy are widely cited as the main reasons for dropping out of school. However, the main reasons vary greatly by region, gender, area of residence and wealth quintile. For example, insecurity is not among the principal reasons for being out-of-school, yet it tends to disproportionally affect young boys and youth from poor households and those living in region 3 (Yade) and region 6 (Haut-Oubangui). Likewise, pregnancy and marriage are among the top reasons accounting for girls’ dropout and these are also cited as one of the top reasons among youth from wealthy households. Early pregnancy is a factor affecting drop outs regardless of the socio-economic background and region for girls. 7. The results from the analysis show that gender, area of residence, wealth status, and supply side factors are strong predictors of youth’s school participation. First, when holding all other factors constant at the mean, being a young girl increases the incidence of being out-of-school by 17 percent; whereas, being from the wealthiest quintile or living in urban areas reduces the incidence of being out-of-school by 17 percent and 14 percent respectively. Second, supply side factors tend to affect mostly young girls and youth living in rural areas. At the national level, the shortage of classrooms increases the out-of-school incidence by 14 percent and 15 percent for the middle and top terciles distribution, respectively. However, the inadequate supply of school infrastructure tends to affect more out-of-school youth who have never attended school. There is a demand for skills but there is inadequate supply (provision of training opportunities) 8. Education and skills training are positively correlated with better employment opportunities in iii CAR. Despite the economy being dominated by the informal sector, education is not only associated with higher wage earnings, it also increases the chances of finding employment in sectors with higher returns and contract employment which offers greater stability. However, youth educational attainment tends to be low despite the benefits. Similarly, the large share of youth who have received skills training are employed but only 37 percent of them can receive some form of skills training due to the low capacity of training centers to provide formal and informal training opportunities. Policy recommendations 9. The issue of out-of-school youth in CAR is a systemic issue but at the same time there is no substitute for strengthening the quality and access to basic education system. Supply and demand side factors affect a youth’s decision of schooling at different stages. As such, remedial policies for out -of- school youth should focus both on preventive strategies through the formal education system and curative strategies for those out of school youth who will not have access to the mainstream formal education system. A more comprehensive set of policy recommendation and priority interventions are expected to be developed during Phase II but the initial following four recommendations can be made from the analysis in Phase I. Improve the school environment to attract and retain children 10. Evidence shows that adequate school infrastructure and teachers are key factors determining on- time enrollment or early dropout. As such, the provision of a conducive learning environment will reduce the incidence of out-of-school by addressing early enrollment and retention in the system. Key school level intervention includes the provision of adequate classrooms facilities, trained teachers and other materials. The incidence of out-of-school among girls, children living in rural areas and children from poor households, requires a combination of interventions including the provision of school level sanitary packages, financial support for poor households, and the promotion of girls’ friendly school environment. Provide alternative education for out of school youth who can no longer enter formal schooling 11. Second chance opportunities should be given to youth who lost the opportunity to attend the formal system. Given many youths missed at least four years of schooling during the prolonged conflicts, the government should consider prioritizing the provision of alternative education for these youth. Several alternative education modules could be designed to fit the characteristics of the youth: (i) condensation of primary education curriculum to give an opportunity to complete primary education in a fast track; (ii) programs that include remedial education plus skills training geared for a specific job or self- employment. These modules can be adjusted to suit different contexts and include extra support for youth with special needs, adolescent girls with children or other family responsibilities. Develop gender-oriented programs to empower young girls and make an impactful change 12. Develop gender-oriented programs to empower young girls and improve the livelihoods for adolescent girls. As evidence shows, adolescent girls face multiple challenges including health and economic ones which are usually associated with early pregnancy, unemployment and other social iv factors. Key suitable programs for women empowerment including life/soft skills training, entrepreneurship training; advisory services; technical training in information and communications technology; tutoring to girls in transitioning to secondary education; work experience through internships; and job placement support. Develop policy that integrates youth in the labor market by focusing on marginalized groups 13. Promote work-based training and apprenticeships by incentivizing private sector participation and involvement based on key priority sectors. Not all learners can access TVET instruction via formal established institutions in a classroom/workshop setting besides which there may not be enough opportunities at TVET institutions. Work-based learning opportunities provide another solid avenue for learners to access technical, vocational and skills both in public agencies and formal and informal private sector enterprises. In addition, apprenticeships and on-the-job training could also improve the chances of youth to find employment after completing their training programs. These programs should also focus on disadvantaged social groups-girls, youth from rural areas and poor families. v Table of Contents I. Introduction ........................................................................................................................................ 3 II. Country context................................................................................................................................... 6 III. Key Sector Performance and Challenges ............................................................................................ 8 IV. Out-of-School Youth Characteristics ................................................................................................. 11 V. Determinants of Out-of-School Youth .............................................................................................. 17 VI. Conclusion and Policy Recommendations ........................................................................................ 24 References .................................................................................................................................................. 29 Annexes ....................................................................................................................................................... 31 1 Table of Figures Figure 1: Number of Out-of-School Children and Youths by Age Group .................................................... 11 Figure 2: Youth School Work Status ........................................................................................................... 12 Figure 3: Youth Educational Attainment by Socioeconomic Group ........................................................... 13 Figure 4: Educational Attainment and Opportunities of Employment in Wage and Productive sector .... 13 Figure 5: Typology of Out-of-School Children and Youth by Age ............................................................... 14 Figure 6: School Status of the Youth Population by Socioeconomic Group ............................................... 15 Figure 7: Youth School Attendance and Dropout Pattern by Socioeconomic Group ................................. 16 Figure 8: Educational Attainment of Out-of-School Youth Who Attended School in the Past................... 16 Figure 9: Reasons for Never Been to School, Youth Age 15-24 .................................................................. 17 Figure 10: Reasons for Dropout from School, Youth Age 15-24 ................................................................. 18 Figure 11: Household Perception on Key Recent Socioeconomic Indicators ............................................. 19 Figure 12: Correlation of Youth School Attendance and Household Perception on Key Recent Socioeconomic Indicators ........................................................................................................................... 19 Figure 13: Supply Side Factors: Effect of Teachers and School Infrastructures on Youth School Status ... 20 Figure 14: Marginal Effect of Determinants of Out-of-School Youth, Age 15-24 ....................................... 21 2 I. Introduction 14. The Central African Republic (CAR) is one of the most fragile countries in the world, which has recently begun to come out of the latest in a series of internal conflicts . Despite a wealth of natural resources such as uranium, crude oil, gold, diamonds, cobalt, lumber, wildlife and hydropower, as well as significant quantities of arable land, CAR remains one of the poorest countries in the world with enormous development challenges. With a Human Development Index (HDI) of 0.36 and a GDP per capita of USD 335 (constant 2010) in 2017, CAR is ranked 188 out of 189 countries on the HDI and second to last on GDP per capita (229th out of 230 countries). Recent estimates, based on observed trends in GDP, suggest that the poverty rate, measured at the international poverty line of US$1.90 per day in 2011 (purchasing- power parity terms), has surged to more than 75 percent in 2016. Naturally, poverty remains a major concern for the Central African Government1. Approximately 62 percent of the population live below the national poverty line, and the poverty rate currently stands at 69 percent in rural areas and 50 percent in urban areas. 15. The 2013 crisis which was a civil war and a humanitarian crisis undid much of the economic progress achieved between 2008 and 2012. Insecurity and conflict have led to the devastation of the country’s economy and the collapse of state institutions. The government controls a limited number of mining areas. The agriculture sector, which more than 75 percent of the population relies on, was especially affected by the crisis.2. Insecurity and population displacement led to a 58 percent decline in crop production, increased food prices and an elevated risk of malnutrition. The state institutions which provide basic services also suffered from the conflicts. Currently, the country is still in an extreme fragile economic state and is facing strong structural constraints. GDP growth declined to -37 percent in 2013 from 4 percent in 2012; and inflation surged to 25 percent in 2014 while it was 5.8 percent in 20123. Although signs of recovery have been observed since 2014 in terms of GDP growth and reduced inflation rate, the deteriorating security environment is a clear hindrance to ensuring sustained economic recovery. 16. As of 2018, the Central African Republic has an estimated total population of 4.9 million and is characterized by a large youth cohort; with about 60 percent of the population below the age of 25 . CAR is a landlocked country, not densely populated (8 people per square kilometer), with a land area of about 620,000 square kilometers. Sixty percent of the population lives in rural areas. Bangui the capital city, is the largest urban center with an estimated population of 800,000 people (40 percent of the total urban population). CAR also have an estimated annual growth rate of 0.7 percent, which is among the lowest in the region and below the SSA average (2.7 percent). 1. An estimated 660,000 youth aged 15–24 years are out of school of which a large proportion remains economically inactive. Out-of-school youth are unlikely to have a productive future without the benefits of adequate training and credentials, and in most cases, work and life skills. The large share of out-of-school children is also a serious concern for CAR’s human capital accumulation, inclusiveness and 1 This is based on the latest year of the last nationally representative household survey (2008) 2 Central African Republic Consolidation and Economic Recovery Development Program Development Policy Operation, 2018 3 CAR Economic Update, 2019 3 global competitiveness. A quarter of children aged 6-18 years in CAR are out-of-school, of which the majority has never been in school. There has been no significant improvement in the rate of out-of-school between 2010 and 2018, and if this chronic issue is not adequately addressed, the low education profile of the labor force will persist and will negatively impact the current economic outcomes which is driven by skills global competition. Within the next decade when this cohort (6-18 years) becomes the core of the labor market, the workforce will likely be composed of workers with low skills who have not completed primary education, which in turn increases the likelihood of working in the informal and weak productivity sectors. The lack of work and life skills will impair these youth’s ability to find jobs in desirable occupations, resulting in low and unstable incomes while exposing them to potentially long periods of unemployment. 17. This report provides a comprehensive review of the state of out-of-school youth in CAR, the determinants of school and job outcomes for the youth, and the policy initiatives in place to support these youth. The objective of the report is to explore the overarching factors that contribute to the out- of-school youth phenomenon and make recommendations for improving existing policies and programs. The report focus on the 15-24 years old cohort for the following reasons: (i) more than half of the youth population leaves the school system without completing primary education, and there is limited information on the reasons explaining school dropouts which contribute to the tendency of the youth leaving the school system without acquiring basic literacy and numeracy skills; (ii) there is limited information on skills training provided in schools or training centers to determine whether the high level of youth inactivity is mainly a supply side issue; (iii) given that the shortage of technical skills is widespread in the country, there is also limited information or data on the quality and relevance of the curricula provided at technical and vocational schools and training centers for curricula revision; (iv) although there is some evidence showing a clear shortage of high-skill youth, there is very limited information available on the labor market side about the expected or required skills from the graduates of technical and vocational schools or training centers and; (v) although several studies have been conducted to inform the reform of TVET, no policies have yet been implemented. 18. The overall report is structured in two phases. This report covers Phase I. Phase II will be developed once Phase I is approved and the team is given the go ahead. Phase I aims to: (i) analyze the magnitude and nature of out-of-school youth by socio-economic characteristics, which include a breakdown of out-of-school by socio-economic characteristics and type such as those who have never been in school and delayed entry; and (ii) show a diagnosis of the supply and demand side determinants of out-of-school youth to help identify key reasons explaining the out-of-school incidence for each socio- economic group and propose relevant policy actions. The key research questions answered in Phase I are as follows: (i) what is the magnitude/incidence of out of school youth in CAR? (ii) Does the prevalence of out-of-school youth vary by social economic group? (iii) what are the key supply and demand factors that affect youth schooling decisions? And (iv) Is there any evidence that the Central African’s labor market signals demand for skills? 19. Phase II would: (i) simulate the costs associated with addressing issues of out-school youth under different skills development mechanisms; (ii) analyze the educational attainment of the youth and adult population to evaluate the adequacy of the stock of human capital; (iii) assess the nature of the labor market in signaling skills to determine if there is skills mismatch; (iv) review the relevance and flexibility 4 of the curriculum offered in technical institutions to prepare young individuals for entrepreneurship and other skills; (v) review the adequacy of technical institutions (both formal and informal including private providers) to provide the youth with necessary skills in a fragile, conflict and violence contexts; (vi) develop a skills and training strategy/framework and a roadmap on how to implement the strategy/framework in the medium term based on lessons learned from fragile, conflict and violence afflicted countries. 20. The analysis in this report uses various data sources, including household surveys, administrative data and other quantitative and qualitative data. In particular, the data comes from: (i) household surveys which include three series of the Indicator Cluster Surveys (MICSs 2000, 2006, and 2010) and the 2016 and 2018 Enquête Nationale sur les Monographies Communales (National Survey for Communal Monographies); (ii) administrative data including the Ministry of Primary and Secondary’s budget, human resources input, and Education Management Information System(EMIS); (iii) national test and the 2006 Program for the Analysis of Education Systems [PASEC] ‘light’ results; (iv) stocktaking of existing relevant data which includes the latest information from development partners as well as tasks undertaken by other units of the Bank such as Poverty, Health Nutrition and Population, Social Protection and Labor GP, etc.; (v) qualitative data collected on youth in Bimbo district4 and through the preparation and organization of experts’ opinions questionnaires and several consultative meetings and focus group discussions with participation from local experts, and government officials; and (vii) desk assessment of policies and procedures, and other documents including the evolution of curriculum at the technical school level. 21. The rest of the report is organized as follows. After the introduction, Section II presents an overview of key education sector performance indicators. Section III provides an overview of the state of out-of-school youth, including the magnitude and overarching characteristics of the out-of-school youth. Section IV explores both demand and supply determinants of out-of-school youth including youth schooling decisions at different transitions stages (primary to lower secondary to upper secondary). Lastly, Section V outlines the main findings and policy recommendations. The annex sections provide detailed empirical analysis to support the findings of Phase I as well as methodological notes. In addition, preliminary assessments of Phase II analysis are included in the annex to provide some insights on Phase I’s policy recommendations. These preliminary assessments include a review of policies and programs in place in fragile contexts, linking intervention tools to the underlying causes of dropping out, such as household constraints, youth behavior, and supply-side problems as well as a mapping of the TVET curriculum and some observations from focus group discussions on constraints of skills development and demand in CAR. 4 Bimbo district is in Ombella-Mpoko prefecture in Region 1 5 II. Country context Macroeconomic Context 22. The Central African Republic (CAR) is an extremely fragile country in central Africa that faces enormous development challenges. Despite a wealth of natural resources such as uranium, crude oil, gold, diamonds, cobalt, lumber, wildlife and hydropower, as well as significant quantities of arable land, CAR is one of the three poorest countries in the world. With a Human Development Index (HDI) of 0.36 and a GDP per capita of USD 335 (constant 2010) in 2017, CAR is positioning it at 188 out of 189 countries and territories on HDI and second to last on GDP per capita (229th out of 230 countries). Recent estimates, based on observed trends in GDP, suggest that the poverty rate, measured at the international poverty line of US$1.90 per day in 2011 (purchasing-power parity terms), has surged to more than 75 percent in 2016. With approximately 62 percent of the population living below the national poverty line, with 69 percent poverty rate in rural areas and 50 percent in urban areas, poverty remains a major concern for the Central African Government5. 23. The 2013 crisis which was a civil war and a humanitarian crisis undid much of the economic progress achieved between 2008-2012. Insecurity and conflict have led to the devastation of the country’s economy and the collapse of state institutions. The government controls a limited number of mining areas; and the agriculture sector, which more than 75 percent of the population relies on, and was especially affected by the crisis6. Insecurity and population displacement led to a 58 percent decline in crop production, increased food prices and an elevated risk of malnutrition. The state institutions which provide basic services also suffered from the conflicts. Currently, the country is still in an extreme fragile economic state and is facing strong structural constraints. GDP growth declined to -36 percent in 2013 from 4 percent in 2012; and inflation surged to 12 percent in 2014 while it was 2.6 percent in 2012. As shown in Figure 1, CAR relatively performed well in relation to other Fragile and conflict affected countries, lower middle-income countries as well as SSA countries after the crisis in 2013. Although signs of recovery have been observed since 2014 in terms of GDP growth and reduced inflation rate, the deteriorating security environment is a clear hindrance to ensuring sustained economic recovery. 24. The recent peace agreement is a positive step towards ending the ongoing crisis and will require substantial resources to be implemented7. The Khartoum peace agreement signed in February between the Government and 14 armed groups reinforces the prospects for further improvement of the political and security situation. Unlike the seven peace agreements previously signed since the outbreak of the crisis in 2013, this one is the first resulting from long and direct discussions between the parties in conflict. Both the Government and the armed groups have made significant commitment to be implemented under the peace agreement. The Government has committed to adopt a new decentralization law, a new law on political parties and legislation of land ownership. Furthermore, it will ensure an equitable 5 This is based on the latest year of the last nationally representative household survey (2008) 6 Central African Republic Consolidation and Economic Recovery Development Program Development Policy Operation, 2018 7 The Consolidation and Social Inclusion Development Program Project Appraisal Document 2019 6 representation in the Civil Service and institutions as well as accelerate the establishment of a Truth, Justice, Reparation and Reconciliation Committee. The armed groups have committed to respect the territorial integrity, constitutional order, democratically elected institutions and the unity of the State. They have also committed to cease hostilities, allow for the deployment of the State and use pacific means by the creation of political parties to make their claims. The implementation of the agreement will take time and require a substantial amount of resources. 25. Progress following the peace agreement has been slow to materialize and support from the international community, beyond the coordinated CEMAC rescue package, remains important for the CAR8. The conflict remains largely unresolved, leaving an impoverished and traumatized population exposed to further exploitation by armed groups. The level of internally displaced persons, refugees and persons in need of humanitarian aid is very high. The United Nations (UN) refugee agency reported in October 2018 an increase of almost 60 percent since 2017 of internally displaced persons (IDP), bringing the number of IDP to 648,000 at levels comparable to those observed during the 2013 crisis. The number of refugees also increased by almost 20 percent in the same period. For a country of 4.7 million inhabitants, this level of displacement is a major constraint to economic growth, macroeconomic stability, and poverty reduction. In addition, much of the security is provided by international security forces led by the Multidimensional Integrated Stabilization Mission in CAR (MINUSCA) while private security companies are increasing their presence. Various UN agencies, international organizations, bilateral donors, and more than 100 international Non-Governmental Organizations (NGO) offer emergency relief and humanitarian assistance. Table 1. Key macroeconomic indicators 201 201 201 2010 1 2012 2013 4 2015 2016 7 GDP (billions of, current LCU) 827 854 889 563 569 596 623 65 GDP (millions, current US$) 2.0 2.2 2.2 1.6 1.7 1.6 1.8 q.9 GDP growth (annual %) 3.0 3.3 4.1 (36.7) 1.0 4.8 4.5 4.3 GDP per capita growth (annual %) 2.0 2.7 .8 (36.8) 0.7 4.1 3.4 2.9 Inflation, consumer prices (annual %) 1.5 1.3 5.8 1.5 25.3 37.1 Fertility rate, total (births per woman) 5.2 5.2 5.1 5.1 5.0 4.9 4.9 Population growth (annual %) 1.0 0.6 0.3 0.2 0.3 0.7 1.1 1.4 Source: CAR Economic Update 8 The Consolidation and Social Inclusion Development Program Project Appraisal Document 2019 7 III. Key Sector Performance and Challenges Key current education sector indicators9 2. Key education sector performance indicators remain weak with girls, children living in rural areas and those from poor families being the most disadvantaged. Key of the indicators include: (i) low Sustainable Development Goals (SDGs) indicators, for example, primary completion rate stands at 64 percent, and the gender parity index is currently at 82 percent at the primary level and 70 percent at the secondary level; (ii) huge disparities in access to education across different socioeconomic backgrounds; (iii) high out-of-school rates as about a third of primary school age children are out-of-school, which disproportionately affect some regions, namely region 3 (North) and region 7 (Bangui); (iv) limited post- schooling opportunities for the youth coupled with the fact that most of the working age population has no formal education; and (v) learning outcomes based on the latest available standardized assessment (2006 PASEC ‘light’ for grade 5) are low. Key sector challenges are described below. Access 3. There is a deep access crisis particularly at the post-primary levels. At the primary level, access tends to be relatively better than the other levels of education, even though the 2013 conflict disrupted the enrollment trends. In 2016, the gross enrollment rate (GER) at the primary level was 92 percent, which was lower than its 2010 value (109%) because of the conflict. As of 2018, the GER has improved slightly and stands at 114 percent; however, given about one third (30% or 228,000 children) of primary school age children are out-of-school, access at the primary level remains low, and only 64 percent complete primary education. However, enrollment rates in preschool were extremely low even before the start of the 2013 crisis and remain low. For example, at the preschool level, the GER stands currently at only 8 percent with about 67 percent of enrollment being provided by private institutions. (Annex Figure A.2). 4. Access at the post-primary is extremely limited as the gross enrollment ratio stand at 36 percent at the lower secondary level, 20 percent at the upper secondary level and only 2 percent at the tertiary level. In general, the access rate of secondary school age children is disproportionally lower, and it has been more affected than the primary level. For example, at the lower secondary, the GER could have increased from 30 percent in 2010 to 50 percent in 2016 at the national level. Boys were the most affected, with a 41 percent reduction of the GER– this could be related to their role in armed groups or militias. Likewise, access to higher education in CAR is astonishingly limited and stands only at 2 percent which could be because higher education is mainly limited to Bangui. Given that post-primary education is key for the development of skills needed in the labor market, this relative low access rate to post-primary education would suggest a negative implication on youth employability in productive sector (Annex Figure B.3-B.5). 9 The formal education system in Central African Republic uses the 3-6-4-3-4 system: the first three years are pre-primary, the next six years are primary, the next four years are lower-secondary, the next three years are upper-secondary, and the final four years are tertiary education. Official school age for preschool is 3 –5, primary 6–11, lower secondary 12–15, upper secondary 16–18, and tertiary 19–25 (See Annex D Table 5). The education sector falls under four ministries: (a) Ministry of Primary and Secondary (MEPS), (b) Ministry of Higher Education (MHE), (c) Ministry of Scientific Research and Technological Innovation (MSRTI), and (d) Ministry of Technical Education and Literacy (META). 8 Inequalities 5. There are huge disparities in access to education across different socio-economic and girls and children living in rural areas and remote regions are the most disadvantaged. The gender gap is significant at all levels of education except at the preschool level. The GPI is currently standing at 82 percent, 70 percent, and 74 percent at the primary, lower and upper secondary levels respectively, albite the target was set to be achieved by 2005 as part of the MDG 3. Although there is no gender gap in access at the preschool and higher education levels, most preschool and higher education services are in urban areas mainly the capital city (Bangui), serving a very small group of the population. 6. The gap in access between children from the poorest and richest households is large and it tends to increase with each level of education. For example, at the primary level the GER for the poorest quintile is 81 percent compared to 112 percent for the richest quintile; at the lower secondary level, the GER is 7 percent for the poorest and 26 percent for the richest; and at the upper secondary level, the GER is 2 percent for the poorest and 37 percent for the richest (Annex Figures B.2-B.5). 7. Concerning the rural-urban inequality in access, it is generally small because the crisis affected mostly urban areas; however, it tends to be relatively higher at the post-primary education level since most post-primary schools are in urban areas. For example, in 2010 (prior the 2013 crisis) the GER in rural areas was 98 percent compared to 128 percent in urban areas, however, in 2016 (after the crisis), the GER in rural areas was 92 percent compared to 91 percent in urban areas (Annex Figures B.2-B.5). There are also regional disparities across all levels of education although these disparities tend to be lower given the crisis contributed to reducing the gap through its impact on regions who have a high number of cities (Annex Figures A.6-A.6) Overall, the impact of the crisis has reduced the GER in urban areas by about 30 percent against 5 percent for rural areas and the reduction in region 7 (Bangui) at primary level is 40 percent. Out-of-school 8. About 27 percent of children aged 6-15 are out-of-school; while the majority has never been to school, the conflict has increased the dropout rates. Before the crisis, the trends of out-of-school show tremendous improvement, but there has been minor improvement between 2010 and 2018. For example, the out-of-school rate for primary school age children (ages 6-11) dropped from 56.4 percent in 2000 to 32.8 percent in 2010 but only decreased to 30 percent in 2018. Similarly, at the lower secondary level, out-of-school rate dropped from 48.6 percent in 2000 to 26.9 percent in 2010 and decreased only to 24 percent in 2018. At the upper secondary level, the out-of-school rate dropped from 77.2 percent in 2000 to 54.1 percent in 2010 and decreased to 49.3 percent in 2018. Although the out-of-school trends at the secondary level are like the primary education level, the share of children who have dropped out at the lower secondary is high and has increased rapidly. For example, at the national level, the share of dropouts increased from 8 percent to 35 percent at the primary level, and from 44 percent to 65 percent at the lower secondary level. The share of dropouts among boys and young people living in urban areas is the highest, as urban areas are mainly affected by the conflict and young boys tend to engage in conflict- related activities (Annex Figures A.9-A.11). For example, at the lower secondary level, the share of boys who dropped out of school accounts for 84 percent of total out-of-school rate for boys compared to only 57 percent for girls. 9 Learning outcomes 9. CAR does not have reliable learning outcomes assessment data which is one of the reasons it was not part of the Human Capital Index (HCI). However, results from the latest 2006 PASEC light10 show that the quality of education in CAR is very low, with an average score of 23.8 out of 100 in French and 28.8 out of 100 in mathematics. Moreover, national exam success rates indicate that that only 32 percent pass the grade 6 exam, and this success rate ranges from 20 percent to 59 percent by academic inspection. The success rate at grade 10 national exams is relatively better and it stands at 59 percent at the national level, ranging from 47 percent to 85 percent across academic inspections. This relatively higher success rate could be due to the very low transition rate to lower secondary and the extremely low access rate at the lower secondary. The success rate at the grade 13 national exams is only 13 percent. While many factors could contribute to the low learning outcomes, key supply side factors—high class size, high STR, high share of unqualified teachers—are among the key contributing factors. 10PASEC Light is national learning evaluation survey for students at the primary level based on the international PASEC instruments. 10 IV. Out-of-School Youth Characteristics Magnitude of the out-of-school youth population (15-24) 10. Approximately 660,000 (i.e. 60 %) youths in the Central African Republic are not in school and, a large share remains inactive and registers low levels of education. The Central African Republic is characterized by a large youth cohort as about 60 percent of the population is below the age of 25. The majority of out-of-school youth leave the system with low levels of education and without acquiring the basic foundational skills. For example, about 70 percent (460,000) of the out-of-school youths are dropouts (Figure 1). The proportion of dropouts among the 15-19 years cohort is 73 percent (191,000 youths), while it stands at 68 percent for the 20-24 years cohort (268,000 youths). The children cohort (6- 14 years) also exhibits a great number of out-of-school which eventually contribute to the large out-of- school youth who will have limited skills for employability. The remainder of this section presents a diagnosis of some factors for school dropouts. Figure 1: Number of Out-of-School Children and Youths by Age Group 400,000 393,704 350,000 Never inschool Droput 263,135 224,130 300,000 250,000 268,347.86 200,000 84,853.09 191,115.84 150,000 66,207 100,000 139,277 125,356 50,000 43,594.52 72,019 - 22,612 Age 6-11 Age 12-14 Age 15-19 Age 20-24 Source: ENMC 2018 11. The 15-24 years old cohort of youth population is characterized by a high level of inactivity, i.e. many youths are not enrolled in school, nor working. The level of inactivity is generally higher among young girls, youths from poor family, and those living in urban areas. At the national level, about 26 percent of the youth population is inactive11 and this increases as the youth grows older. Youths from wealthy households tend to remain in school longer and acquire higher levels of education which prepare them for better labor market outcomes. However, young girls and youths from poor households tend to engage in low productive sectors, register low schooling enrollments or simply remain inactive. For example, more than half of the youth from the richest quintile is still in school compared to only 18 11 only 17 percent of youth engaged in agricultural work which is consistent with the survey (2018) and almost all youth said they were looking for work (89 percent) which show those who are employed are somewhat underemployed i.e. not satisfied with their current job; 11 percent of the youth from the poorest quintile (Figure 2). Additionally, the low proportion of out-of-school youths from wealthier households can usually afford to be economically inactive and tend to be less engaged into work activities; in comparison with the relatively large proportion of youths from poorer households who might be constrained to take any available job12. Moreover, there are large disparities between regions in terms of youth inactivity. For instance, the level of youth inactivity is below the national average in region 1 (23%), region 3 (21%), region 4 (20%) and region 7 (23%), while more than a third of the youth in region 2 (36%) and region 6 (33%) is neither enrolled in school nor working. Lastly, the 20-24 years old register a higher-level inactivity in comparison to the 15-19 years old. This could be explained by the fact that the 15-19 years are still enrolled in school. Figure 2: Youth School Work Status 13 School only Both work and school School only Both work and school Work only Inactive Work only Inactive 21% 24% 24% 25% 25% 26% 26% 26% 28% 30% 30% 30% 32% 33% 34% 23% 21% 20% 23% 5% 28% 33% 8%2%5% 4% 36% 23% 2% 12% 33% 30% 40% 2% 38% 21% 49% 49% 38% 48% 50% 50% 60% 3% 31% 65% 34% 47% 53% 6% 36% 36% 37% 36% 5% 62% 26% 27% 27% 4% 9% 3% 19% 15% 18% 48% 9% 7% 33% 32% 27% 22% 27% Age 15-19 Age 20-24 Age 15-24 Age 15-19 Age 20-24 Age 15-24 Age 15-19 Age 20-24 Age 15-24 Age 15-19 Age 20-24 Age 15-24 Age 15-19 Age 20-24 Age 15-24 Region1 Region2 Region3 Region4 Region5 Region6 Region7 National Rural Female Poorest Richest Source: ENMC 2018 12. Many central African youth lack the basic foundational skills such as numeracy and literacy needed to engage into productive labor market activities. About 22 percent of the youth aged 15-24 has no formal education, only 44 percent has some primary education and only about 1 percent has been beyond secondary education (Figure 3). Young girls, youth living in rural areas and those from poor families have the lowest level of educational attainment. However, the distribution of the youth population by age group shows that the educational attainment of the youth force is improving as the younger cohort register lower levels of no education. For example, about 18 percent of the 15-19 years old has no formal education compared to 28 percent for the 20-24 years old. Yet, given that many youths do not complete secondary education, coupled with the low-quality of education in CAR which puts youth 12 Deon Filmer (2005) 13 To complement the analysis, a qualitative survey was conducted to understand the characteristics of the youth in CAR. About 186 youth (137 men and 49 women) were interviewed and the detailed results and methodology used can be found in Annex C. 12 employability at risk; the youth population will most likely remain functionally illiterate and lack foundational skills in literacy and numeracy if this issue is left unaddressed. Figure 3: Youth Educational Attainment by Socioeconomic Group No formal education Some primary Some secondary Some higher 100% 0% 1% 2% 0% 0% 1% 3% 2% 0% 0% 0% 1% 4% 12% 16% 14% 8% 90% 18% 18% 18% 33% 34% 33% 27% 25% 26% 80% 70% 59% 36% 46% 61% 60% 44% 54% 65% 59% 53% 37% 43% 50% 48% 36% 44% 40% 49% 30% 47% 32% 20% 38% 35% 29% 33% 39% 18% 26% 28% 22% 22% 28% 25% 10% 18% 7% 9% 8% 0% Age 20-24 Age 15-24 Age 15-19 Age 15-24 Age 15-19 Age 20-24 Age 15-24 Age 15-19 Age 20-24 Age 15-24 Age 15-19 Age 20-24 Age 15-19 Age 20-24 Age 15-24 National Rural Female Poorest Richest Source: ENMC 2018 13. Despite the economy being dominated by the informal sector14, education is not only associated with higher wage earnings, but it also increases the chances of finding employment in sectors with higher returns and contract employment which offers greater stability. As shown on Figure 4, higher educational attainment is associated with higher probability of finding wage employment and employment in industry and service sectors. A series of multinomial logit regressions when controlling for the main observable characteristics of the labor force, confirm the importance of education in determining the sector and status of employment. With family farming as the base category, the likelihood of finding wage employment for individuals with some primary, some lower secondary, some upper secondary and some higher education increase by 9 percent, 138 percent, 261 percent and 396 percent, respectively; compared to individuals with no formal education. Similarly, with agriculture as the base category, the likelihood of working in the services and industry sectors increases with higher levels of education. Figure 4: Educational Attainment and Opportunities of Employment in Wage and Productive sector 14Youth Survey (2019)-- about 48 percent of youth are engaged in economic activities of which more than half involved in informal self-employment jobs—54 percent in self-employment, 14 percent in public sector and 31 percent in private sector; 13 Some primary Some lower secondary Some upper Secondary Post-secondary 428% 396% 300% 261% 268% 273% 235% 196% 161% 138% 101% 95% 55% 43% 47% 9% Wage Non-wage household Industry Services enterprise Employment Status Sector of employment Source: ENMC 2018 TYPOLOGY OF OUT-OF-SCHOOL CHILDREN AND YOUTH POPULATION (6-24) 14. The typology of out-of-school children and youth reveals that out-of-school youth tend be a systemic issue as most children do not start on time and dropout early. As shown on Figure 5, half of the children start school late and only 44 percent remain in school at age 17. While the incidence of dropouts tends to be higher among older children, delayed entry could be one of the factors contributing to early dropout; as delayed entry leads to overage enrollment which later lead to dropping out without acquiring adequate skills for labor market activities15. Figure 5: Typology of Out-of-School Children and Youth by Age 15Overage enrollment is a very important predictor of dropout patterns. Across Sub-Saharan Africa, delaying enrollment by one year increases the probability of dropping out by 9.2 percent, but the estimates vary greatly across countries. For example, in Uganda, each year of over-age enrollment increases the probability of dropping out by 2.2 percent, whereas this rate is more than 20 percent in Burundi and Swaziland. Late school entry is one sources of difficulties for students, school teachers and administrators from a number of perspectives. For students, older students and repeaters are ridiculed by their younger classmates and this could lead to dropout. Older students are also likely to perform poorly in class and often have difficulty in coping and may dropout (Taylor, Thabo Shindler & Akoobhai, 2010) 14 Inschool Late entry Dropout Never attended 16 8 1 7 5 5 4 17 19 18 34 39 39 36 56 51 63 66 71 82 78 84 79 79 75 76 67 65 65 50 44 45 48 27 33 20 18 13 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Age (in years) Source: ENMC 2018 15. The incidence of out-of-school disproportionately affects young girls, youth living in rural areas, and those from economically disadvantaged groups. Figure 6 shows large gender, urban-rural and wealth disparities among the out-of-school youth. The out-of-school rate among young girls is higher, standing at 70 percent; as opposed to young boys, where the rate is 49 percent. Besides, the out-of-school rate varies largely across regions. For instance, the out-of-school rate is lower in region 7 (35 percent) while it reaches 70 percent in region 4 (Annex Figure A.13). Similarly, the out-of-school incidence affects the youth living in rural areas more than those living in urban areas. The out-of-school youth is 69 percent in rural areas whereas it is 48 percent in urban areas. Likewise, the out-of-school incidence is relatively higher among youths from poorer households than those from wealthier households; which like the urban-rural gap, further reinforces the existing inequalities and hinders the poverty reduction agenda of the country. Figure 6: School Status of the Youth Population by Socioeconomic Group In school Dropout Never attended 10 11 4 18 25 23 25 39 37 42 42 45 46 44 51 52 55 40 30 31 31 Male Female Rural Urban Poorest Richest National Gender Area Wealth Quintile Source: ENMC 2018 16. Many youths do not start school on time, and the majority either dropout at the primary level or are enrolled at a level below their age cohorts—primary school. Figure 7 shows the school attendance tree of youth (15-24). At the national level, only 6 percent of youth are enrolled in upper secondary and 15 2 percent have completed upper secondary education16. However, close to a third of the youth is still enrolled either in primary or lower secondary, below their age cohorts. As shown, youth from wealthier households tends to get benefits from the education system, as they enrolled on time and stay in school longer. Figure 7: Youth School Attendance and Dropout Pattern by Socioeconomic Group No formal schooling Left at primary Still in primary Left at lower secondary Still in lower secondary Left upper secondary Still in upper secondary Competed upper secondary 100% 80% 60% 40% 20% 35% 18% 26% 24% 0% 6% National Female Rural Poorest Richest Source: ENMC 2018 Close to three-quarter of the out-of-school youth (15-24) who attended school in the past, drop out early before completing primary education. Figure 8 shows the educational attainment of the out-of- school youth who attended school in the past. As shown, at the national level, 69 percent of the youth aged 15-24 who had dropped out, did so before completing primary school. This incidence is even higher standing at 75 percent for the out-of-school youth living in rural areas. Figure 8: Educational Attainment of Out-of-School Youth Who Attended School in the Past Some primary Some lower secondary Some upper Secondary Post-secondary 3% 3% 2% 1% 19% 18% 10% 27% 24% 44% 72% 79% 81% 69% 42% National Female Rural Poorest Richest Source: ENMC 2018 16Youth survey(2019)similarly shows about 40 percent of youth dropped out at the primary education level and 40 percent dropped out at the lower secondary level. 16 V. Determinants of Out-of-School Youth Reasons for never attending school 17. Family refusal, cost of education, distance to school and being too young are the principal reasons cited for out-of-school youth who have never been in school. Figure 9 shows the main reasons given for out-of-school youth (15-24 years) who have never been in school. As shown, the main reasons vary greatly by region, gender, area of residence and wealth quintile. For example, although family refusal is cited as the key reason at the national level, cost of education and distance to school are the main reasons cited for out-of-school youth from poor and rich households who have never been to school. Although insecurity is not among the principal reasons for being out-of-school, it tends to disproportionally affect young boys and youth from poor households and those living in region 3 and 6. Lastly, cost of education is not the main reason for being out-of-school in in region 1, region 6 and region 7 (Annex Figure A.14). Figure 9: Reasons for Never Been to School, Youth Age 15-24 Family refusal Cost School too far Too young Insecurity Illness Quality Work Other 7% 4% 1% 0% 15% 9% 11% 8% 15% 13% 16% 24% 25% 15% 29% 32% 27% 26% 23% 28% 36% 29% 34% 28% 33% 19% 16% 16% Total Male Female Rural Urban Q1 Q5 Source: ENMC 2018 Reasons for dropping out The cost of education, failed examinations, insecurity and pregnancy are widely cited as the main reasons for dropping out of school. As mentioned above, many youths in CAR do not start school on time and dropout early without completing primary education17. The reasons for dropping out vary across social groups. For example, pregnancy and marriage are among the top reasons accounting for girls’ dropouts (Figure 10). Pregnancy is also cited as one of the top reasons among youth from wealthy households. In addition, it is worth to mention that opportunity cost seems not to be a reason why central African youth dropped out of school given that only 1 percent cited work as the main reason for dropping out. Similarly, lack of public teachers and distance to school are not part of the main reasons cited for dropouts18. 17 Youth Survey (2019) also shows the majority of youth cited poor economic condition as the main reason for being out-of- school 18 Under the Family Code 1998 the minimum age of marriage is 18 years. However individuals can marry under the age of 18 with parental consent or if a state prosecutor dispenses the age requirement based on serious grounds (https://www.girlsnotbrides.org/child-marriage/central-african-republic/). In CAR, 68% of girls married before the age of 18 and 29% married before the age of 15. However, the recent household survey (ENMC 2018) shows that women in CAR 17 Figure 10: Reasons for Dropout from School, Youth Age 15-24 Cost Abandonment/failure Insecurity Pregnancy Married Quality Illness School too far away School closed Family refusal Absence of teacher Work 0% 8% 10% 11% 13% 18% 11% 13% 17% 11% 25% 10% 18% 7% 10% 25% 23% 20% 19% 20% 15% 41% 35% 34% 33% 26% 31% 28% Total Male Female Rural Urban Q1 Q5 Source: ENMC 2018 Household socioeconomic factors 18. Households’ socioeconomic factors could also contribute to the incidence of out -of-school youth in CAR. Figure 11 summarizes key socioeconomic factors affecting out-of-school rate. The following are observed: (i) about 12 percent of CAR’s population has been displaced since the beginning of the conflict and only 7 percent of the population has returned as of 2018; (ii) about 18 percent of households hosted displaced families, and the total share of the refugee’s population during the conflict reached 20 percent; (iii) about 30 percent of CAR’s population has been affected by the conflict and about 24 percent experienced illness or disease due to the conflict; (iv) about 18 percent of households also experienced death due to conflict; (v) about a third of households reported valuable household items were stolen from them; and (vi) more than half of the households feels the security situations has been improving, however, a large share of households believes that economic conditions and public services are worsening. All these socioeconomic factors negatively affect youth’s school attendance19. For example, the out-of-school rate among youth who are subject to displacement is 67 percent compared with those who did not experience displacement (58 percent). Youth from households who returned to their original area of residence also face high incidence of out-of-school rate in comparison to those who have never been displaced, 73 are generally unmarried, 13% of the adult population is married, 4% of the youth population is married, 23% of the youth say they are in union. Among the youth population, 36 percent of them are married and under the age of 18. Similarly, from those in union, 23 percent the union happened before age 18. From the out-of-school youth who drop out, 47% were pregnant at age 17 and common-law, and 40% were pregnant at age 18. 19 Youth survey (2019) shows only 37 percent of youth have received some form of skills training and the majority of those who have received skills training are employed; (vi) about 5 percent of youth reported having served in an armed group during the conflict and about 40 percent of those who served in the armed group joined before they reached their 14th birthday. More than half of the youth have lost either both or one of their parents and the majority of the youth lives with family members; 18 percent vs 58 percent (Figure 12). Figure 11: Household Perception on Key Recent Socioeconomic Indicators 79% 79% 86% 72% 62% 53% 60% 62% 48% 55% 30% 24% 32% 37% 12% 7% 18% 20% 18% Death Returned Diseased Satisfied with Feeling safe at transport Refugee Affected security improved Access to Job Access to market education Security improved Economic Access clean Displaced Stolen property Employment Access to condition Access to Hosted displaced Community water Youth safety home Internally Displaced Conflict Improved Worse Worse People Source: ENMC 2018 Figure 12: Correlation of Youth School Attendance and Household Perception Never been in school Never been in school Dropout Dropout Total out-of-school 70% 58% 62% 58% 62% 60% 58% 57% 73% 67% 66% 65% 58% 58% 58% 58% No Yes No Yes No Yes No Yes No Yes No Yes No Yes No Yes Displaced Returned Hosted Refugee displaced Affected Diseased Stolen Death property Never in school Never been in school 61% Dropout 61% Dropout 57% 60% 66% 55% 54% 62% 61% 63% 63% 56% 56% 54% 55% 52% No Yes No Yes No Yes No Yes No Yes No Yes No Yes No Yes Access to Access to Access Access to Access Economic condition Access to Job Youth market transport clean water education to Employment health Source: ENMC 2018 Supply side factors 19. Shortage of trained teachers and school infrastructures is associated with high rates of out-of- 19 school youth. Supply side factors also contribute to the high incidence of out-of-school youth. Figure 13 shows the supply of teachers measured as the share of public teachers in the total teaching force and the availability of school infrastructure measured as the ratio of total classrooms to total teachers. The results from this analysis show that at the national level, the inadequate supply of public teachers is associated with high rates of out-of-school youth, which is 67 percent at the bottom tercile of public teacher ratio, and 49 percent at the top tercile. The inadequate supply of public teachers mostly affects out-of-school youth who have never attended school. For example, the out-of-school rate for youth who never attended school stands at 28 percent at the bottom tercile as opposed to only at 9 percent at the top tercile; whereas the dropout rate is almost the same for both terciles (40 percent). Similarly, the inadequate supply of school infrastructure tends to affect more out-of-school youth who have never attended. Generally, supply side factors tend to affect mostly young girls and youth living in rural areas. Figure 13: Supply Side Factors: Effect of Teachers and School Infrastructures on Youth School Status Never been in school Dropout 82% 76% 75% 72% 68% 67% 67% 55% 53% 52% 49% 42% 38% 33% 31% 28% 22% 25% 12% 18% 9% 6% 8% 9% NationalFemale Rural NationalFemale Rural NationalFemale Rural NationalFemale Rural <.1 >.56 <.66 >1 Share of public teachers of total teachers Classroom teachers ratio Source: ENMC 2018 20. The report used several econometric models to investigate how demand and supply side factors affect the probability of being out-of-school and decisions to enroll in school, and dropout of school. The modeling aims to determine the following four aspects of out-of-school youth: (i) demand and supply side factors that affect the decision of the youth to be in or out of school using a binary logit model; (ii) factors that affect the youth’s decisions to first enroll in school vs factors that lead to dropping out of school using a multinomial logit model; (iii) factors that affect youth’s schooling decisions at different times and youth’s attendance status using a sequential logit model; and (iv) factors that vary by gender, geographic area and poverty status using the Oaxaca decomposition model. The Oaxaca decomposition model aims to determine whether certain known factors (endowments or explained) justify the difference observed between young male vs female; or whether the differences are attributed to some unknown factors (coefficients or unexplained). The detailed methodology of this analysis is presented in Annex E. 21. The following are the two key supply side factors used in the analysis. First, the public provision 20 of teachers is used as a measure of resources allocated to the education sector since a large share of recurrent education spending goes to teachers. Additionally, there are 72 subprefectures and the teacher’s data is linked to the household survey at the subprefecture level to determine whether there is an even distribution of public resources devoted to education at the subprefectures level. Second, the provision of classroom at the subprefecture level is used as a measure to capture the distribution of capital resources across the country. The following are the five key demand side factors included in the model: (i) household head’s education level, gender, and sex; (ii) household’s wealth status and family size; (iii) prevalence of conflicts; (iv) areas of residence; and (v) youth’s age and gender. Factors that affect youth’s decision to be in or out of school 22. The results from the analysis show that gender, area of residence, wealth status, and supply side factors are strong predictors of youth’s school participation (Figure 14). Holding all other factors constant at the mean (marginal effect), being a young girl increases the incidence of being out-of-school by 17 percent. Similarly, being from the wealthiest quintile, reduces the incidence of being out-of-school by 17 percent. The incidence of out-of-school youth is also reduced by 14 percent for the youth living in urban areas. On the supply side, the shortage of classroom increases the out-of-school incidence by 14 percent and 15 percent for the middle and top terciles distribution, respectively. The inadequate supply of teachers also matters but it is not as strong as the limited supply of classrooms. This would suggest that financial resources are a key determinant of youth educational outcomes. 23. Household head’s education level, gender and age are not major determinants of youth’s school attendance. Although other factors such as household head education, head age, household size, conflicts and children’s age play some roles in the youth’s schooling decisions, when the effect of other factors are controlled for, the effect tend to be lower. For example, conflict only affects youth decisions if it leads to displacement or death of household members after all other factors have been controlled. The detailed regression results are presented in (Annex E Table 1). Figure 14: Marginal Effect of Determinants of Out-of-School Youth, Age 15-24 21 20% 17% 15% 14% 15% 10% 7% 6% 4% 5% 3% 3% 2% 0% 0% 0% -5% -2% -1% -4% -5% -10% -6% -8% -15% -11% -14% -20% -17% Female Age Urban Rich Richest Female Age Poor Pop displace Some primary Some secondary Family Size Death of household members Middle HH affected by conflict Middle tercile Top tercile Middle tercile Top tercile Some post secondary Student Area Wealth Head education Head . Conflict Teachers Classroom Source: ENMC 2018 Factors that affect youth decisions to enroll and dropout 24. Although many demand and supply side factors affect both youth enrollment and dropout decisions, some factors only affect enrollment decisions. Gender, age, area of residence, family size and shortage of classroom affects both enrollment and dropout decisions while household wealth and availability of teachers mostly affected decision to enroll. Conflict also seems to affect more the decision to dropout rather than the decision to enroll. This suggests that a package of remedial solutions should be considered when thinking of policy options for youth who have dropped out (Annex E Table 1). Factors that affect decision of youth school attendance at different level 25. Some factors affect all transition stages while others are specific to certain stages in the transition. For example, the effects of household wealth status are most evident in the first few transitions then disappears. The effect of head education is more relevant at the later transition implying better educated parent has influence on educational attainment of youth. Area of residence and sex of student have less effect on last transition stage. Only age of student affects all transitions levels. Supply side factor—teachers provision affects only the first two transition while shortage of classroom affects the first transition and transition corresponding to secondary school level as infrastructure shortage is huge in CAR in post primary education. As such, policy actions can be sequenced to address the importance of 22 determinants at each level. In particular, availability of schools nearby and the wealth quintile of a household should be considered when making policy choices that support enrollment and reduce dropout at the primary level. To determine the factors that influence decisions for schooling at each transition level, we employed a sequential logit model and the result is presented the (Annex E Table 2). The model includes five transitions: first, decision whether to enroll in primary school (vs. never enrolling); second, decision whether to continue/finish primary school (vs. dropping out of primary school); third, given that the youth continued/finished primary school, whether to get into lower secondary education or not; fourth, given that the youth enrolled in lower secondary education, whether the youth eventually dropped out or not; fifth, given that the youth enrolled in upper secondary education, whether the youth eventually dropped, continued with upper secondary education or youth completed upper secondary education by the age of 24. Factors that drive difference between genders, urban vs rural and poor vs non-poor 26. Finally, the decomposition results under the three categories (male vs. female, urban vs. rural, and poor vs. non-poor) show that the endowment (or explained factors), plays an insignificant role in explaining the difference except the urban and rural differences. Annex E Table 3 shows decomposition results. Controlling all factors, except for the difference between rural and urban, know supply and demand side factors accounts very little in explaining the gaps between the other two categories. For example, 43 percent of the difference between rural and urban in terms of youth school participation is explained by the known factors (factors related to students, teachers, schools, parents and conflicts). The know factors play less than 20 percent of the gaps in other two groups. Overall, the factors that drive the gaps between gender and poverty status are not the known factors and additional information is needed to deepen the analysis and develop relevant policy options. These will be carried out in Phase II. 23 VI. Conclusion and Policy Recommendations Conclusions Prolonged conflict has severely affected the education system making out-of-school youth a product of the “broken system� 26. The 2013 crisis which was a civil war and a humanitarian crisis worsened the situation in CAR which in itself was alarming before the crisis. Insecurity and conflict have led to the collapse of state institutions, particularly the state’s capacity to provide basic service delivery and respond to the population’s urgent needs. Not surprisingly, the education sector was also profoundly affected by the crisis-- school facilities were looted or destroyed and thousands of children lost years of schooling. The crisis increased gender inequality in terms of access to education and the vulnerability of girls to exploitation and abuse was even made more acute. The sector is poorly financed, and the government has limited capacity in providing and retaining primary school teachers as 51 percent of public primary schools have no government-provided teachers. The provision of school infrastructure is also extremely weak and existing classrooms are overcrowded, with average class sizes of 95 in primary public schools and even higher in public secondary schools at 124. Key educational indicators trailing the Sustainable Development Goals 27. Key education sector performance indicators show that the education sector had progressed significantly at both primary and secondary levels between 2000 and 2010. However, this progress was disrupted by the 2013 crisis and key education sector performance indicators remained low since. Some of the main challenges impeding the performance of the Central African Republic’s education sector include: (i) low Millennium Development Goals/Sustainable Development Goals (MDG/SDG) indicators; (ii) huge disparities in access to education across different socioeconomic backgrounds; (iii) high out-of- school rates as about a third of primary school age children are out-of-school, which disproportionately affect some regions; and (iv) limited post-schooling opportunities for the youth coupled with the fact that most of the working age population has no formal education. A large proportion of the youth population is inactive and lack numeracy and literacy skills 28. An estimated 660,000 youth (i.e. 60 %) in the CAR are not in school and a large share remain inactive and register low levels of educational attainment. Most out-of-school youth leave the education system with low levels of education and without acquiring basic foundational skills. Many youths start school in CAR, but the majority drop out at an early age and in fact, delayed entry could be one of the factors contributing to early dropout. In fact, 70 percent of the out-of-school youth are dropouts. The level of inactivity is generally higher among young girls, youth from poor family, and those living in urban areas. 29. A large proportion of the youth population has never been beyond primary education and lack basic numeracy and literacy skills needed to engage in productive labor market activities. About 18 24 percent of youth aged 15-24 have no formal education. This low educational attainment makes the transition from school to work difficult and most youth tend to engage in low productivity sectors and informal employment. Although poor educational outcomes are not just a manifestation of poor employment outcomes, the low educational outcome could determine the labor market orientations. Out-of-school youth tend be girls and those from disadvantaged social groups 30. The out of school youth population is made up disproportionately of young girls, youth living in rural areas, and those from economically disadvantaged groups. The out-of-school rate among young girls is higher, standing at 70 percent; as opposed to young boys, who are at 49 percent. Similarly, the out-of- school incidence affects youth living in rural areas (69%) more than those living in urban areas (48%). Likewise, the out-of-school incidence is relatively higher among youth from poorer households than those from wealthier households. This further reinforces the existing inequalities and hinders the poverty reduction agenda of the country. Moreover, the out-of-school rate varies largely across regions as some regions are more affected by the conflicts than others. The out-of-school rate is lower in region 7 (Bangui) standing at 35 percent; while it reaches 70 percent in region 4 (Kagas) Key determinants of out-of-school youth vary by region, gender, area of residence and wealth quintile 31. Family refusal, cost of education, distance to school and being too young are the principal reasons cited for out-of-school youth who have never been in school whereas the cost of education, failed examinations, insecurity and pregnancy are widely cited as the main reasons for dropping out of school. However, the main reasons vary greatly by region, gender, area of residence and wealth quintile. For example, insecurity is not among the principal reasons for being out-of-school, yet it tends to disproportionally affect young boys and youth from poor households and those living in region 3 (Yade) and region 6 (Haut-Oubangui). Likewise, pregnancy and marriage are among the top reasons accounting for girls’ dropout and these are also cited as one of the top reasons among youth from wealthy households. Household’s socioeconomic factors also contribute to the incidence of out-of-school youth in CAR. For example, the out-of-school rate is higher among youths who were subjected to displacement and stands at 67 percent, in comparison with youths who did not experience displacement, which is 58 percent. Youths coming from households who have returned to their original place of living also face high rates of out-of-school (73%) as opposed to youths who were never displaced (58%). Supply side factors such as the inadequate supply of teachers and school infrastructure also contribute to high incidence of out-of-school youth. 32. The results from the analysis show that gender, area of residence, wealth status, and supply side factors are strong predictors of youth’s school participation. The report used several econometric models to investigate how demand and supply side factors affect the probability of being out-of-school and decisions to enroll in school, and dropout of school. First, when holding all other factors constant at the mean, being a young girl increases the incidence of being out-of-school by 17 percent; whereas, being from the wealthiest quintile or living in urban areas reduces the incidence of being out-of-school by 17 percent and 14 percent respectively. Second, supply side factors tend to affect mostly young girls and youth living in rural areas. At the national level, the shortage of classrooms increases the out-of-school 25 incidence by 14 percent and 15 percent for the middle and top terciles distribution, respectively. However, the inadequate supply of school infrastructure tends to affect more out-of-school youth who have never attended school. Although many demand and supply side factors affect both enrollment and dropout decisions, some factors affect enrollment decisions only. Specifically, youth’s gender, age, area of residence family size and the limited supply of school infrastructure affect both enrollment and dropout decisions; however, household wealth’s status and the limited supply of public teachers mostly affect enrollment decisions. Besides, conflict seems to affect dropout decisions more than enrollment decisions; which would suggest that remedial solutions should consider different policy options for youths who have never been in school and youth who have dropped out. There is a demand for skills but there is inadequate supply (provision of training opportunities) 33. Education and skills training are positively correlated with better employment opportunities in CAR. Despite the economy being dominated by the informal sector, education is not only associated with higher wage earnings, it also increases the chances of finding employment in sectors with higher returns and contract employment which offers greater stability. However, youth educational attainment tends to be low despite the benefits. Similarly, the large share of youth who have received skills training are employed but only 37 percent of them can receive some form of skills training due to the low capacity of training centers to provide formal and informal training opportunities. 34. The capacity and quality of TVET institutions is limited. At the national level, there are around 179 training providers in the Central African Republic. Training providers are concentrated in Bangui, the capital city, where more than 82 percent of the training providers are located. The fact that most training programs are in Bangui and urban areas would suggest that most of the youth in rural areas have limited access to these training opportunities. Most training programs are in auto mechanics, masonry and electricity which is offered in 71 percent of the surveyed TVET institutions; followed by carpentry (57 percent); and plumbing and pipping (43 percent). Although most of the employment in CAR is predominantly in the agriculture sector, only one private TVET institution offers agriculture and livestock training programs. Policy Recommendations 35. The Central African Republic is one of the most fragile countries in the world, which has recently begun to come out of the latest in a series of internal conflicts. Out-of-school youths are unlikely to have a productive future without the benefits of adequate training and credentials, and in most cases, work and life skills. There is a critical need to engage these young boys and girls especially those from disadvantaged groups to promote peacebuilding and economic development. The TVET sector (both formal and informal) has limited capacity to address these issues and provide livelihood and economic improvement opportunities in a sustainable manner. Moreover, this sector faces multiple challenges including lack of relevance of the training programs. The findings in this report provided several directions to consider for policy actions and there is no one size fit for all the challenges. 26 36. The issue of out-of-school youth in CAR is a systemic issue but at the same time there is no substitute for strengthening the quality and access to basic education system. Supply and demand side factors affect a youth’s decision of schooling at different stages. As such, remedial policies for out -of- school youth should focus both on preventive strategies through the formal education system and curative strategies for those out of school youth who will not have access to the mainstream formal education system. A more comprehensive set of policy recommendation and priority interventions are expected to be developed during Phase II but the initial following four recommendations can be made from the analysis in Phase I. Improve the school environment to attract and retain children and reduce the out-of-school youth 37. Evidence shows that adequate school infrastructure and teachers are key factors determining on- time enrollment or early dropout. As such, the provision of a conducive learning environment will reduce the incidence of out-of-school by addressing early enrollment and retention in the system. Key school level intervention includes the provision of adequate classrooms facilities, trained teachers and other materials. The incidence of out-of-school among girls, children living in rural areas and children from poor households, requires a combination of interventions including the provision of school level sanitary packages, financial support for poor households, and the promotion of girls’ friendly school environment. Provide alternative education for youth who can no longer enter formal schooling 38. Second chance opportunities should be given to youth who lost the opportunity to attend the formal system. Given many youths missed at least four years of schooling during the prolonged conflicts, the government should consider prioritizing the provision of alternative education for these youth. Several alternative education modules could be designed to fit the characteristics of the youth: (i) condensation of primary education curriculum to give an opportunity to complete primary education in a fast track; (ii) programs that include remedial education plus skills training geared for a specific job or self- employment. These modules can be adjusted to suit different contexts and include extra support for youth with special needs, adolescent girls with children or other family responsibilities. Develop gender-oriented programs to empower young girls and make an impactful change 39. Develop gender-oriented programs to empower young girls and improve the livelihoods for adolescent girls. As evidence shows, adolescent girls face multiple challenges including health and economic ones which are usually associated with early pregnancy, unemployment and other social factors. Key suitable programs for women empowerment including life/soft skills training, entrepreneurship training; advisory services; technical training in information and communications technology; tutoring to girls in transitioning to secondary education; work experience through internships; and job placement support. Develop policy that integrate youth in the labor market by focusing on marginalized groups 40. Promote work-based training and apprenticeships by incentivizing private sector participation and involvement based on key priority sectors. Not all learners can access TVET instruction via formal 27 established institutions in a classroom/workshop setting besides which there may not be enough opportunities at TVET institutions. Work-based learning opportunities provide another solid avenue for learners to access technical, vocational and skills both in public agencies and formal and informal private sector enterprises. In addition, apprenticeships and on-the-job training could also improve the chances of youth to find employment after completing their training programs. These programs should also focus on disadvantaged social groups-girls, youth from rural areas and poor families. 28 References Agresti, A. (2002). Categorical Data Analysis, Second Edition, New York: John Wiley & Sons. Altonji, J. G., Bharadwaj, P., & Lange, F. (2008). Changes in the characteristics of American youth: Implications for adult outcomes. Working paper, Yale University. Ammermüller, A. (2007). PISA: What makes the difference? Explaining the gap in test scores between Finland and Germany. Empirical Economics, 33 (2) (2007), pp. 263-287. Blinder, A. S. (1973). Wage discrimination: reduced form and structural estimates. Journal of Human resources, 436-455. Buis, M. L. (2010). Not all transitions are equal: The relationship between inequality of educational opportunities and inequality of educational outcomes. M. Buis, Inequality of Educational outcome and inequality of educational opportunity in the Netherlands during the 20th century. PhD thesis. Deon Filmer (2005): “Gender and wealth disparities in schooling: Evidence from 44 countries� International Journal of Educational Research 43 (2005) 351–369 Fortin, N., Lemieux, T., & Firpo, S. (2010). Decomposition methods in economics. In Handbook of labor economics (Vol. 4, pp. 1-102). Elsevier. Fox, J. (1997). Applied regression analysis, linear models, and related methods. Thousand Oaks, CA: Sage. Glewwe, P., & Kremer, M. (2005). Schools, teachers, and education outcomes in developing countries. Handbook of the Economics of Education, 2, 945-1017. Greene, H. W. (2005). Econometrie. Paris: Pearson Education 5e ed. Heyneman, P. S. (1979). The career education debate: Where the differences lie. Teachers College Record, 80(4), 659-688. Horrace, W. C., & Oaxaca, R. L. (2001). Inter-industry wage differentials and the gender wage gap: An identification problem. ILR Review, 54(3), 611-618. Inoue, K., Di Gropello, E., Taylor, Y. S., & Gresham, J. (2015). Out-of-school youth in Sub-Saharan Africa: A policy perspective. The World Bank. http://documents.worldbank.org/curated/en/293771468004845225/Out-of-school-youth-in-Sub- Saharan-Africa-a-policy-perspective Jann, B. (2008). The Blinder–Oaxaca decomposition for linear regression models. The Stata Journal, 8(4), 453-479. Jones, F. L. (1983). On decomposing the wage gap: a critical comment on Blinder's method. The Journal of Human Resources, 18(1), 126-130. Krieg, J. M., & Storer, P. (2006). How much do students matter? Applying the Oaxaca decomposition to explain determinants of adequate yearly progress. Contemporary Economic Policy, 24(4), 563-581. 29 Krueger, A. B. (2003). Economic considerations and class size. The Economic Journal, 113(485), F34-F63. Machado, J. A., & Mata, J. (2005). Counterfactual decomposition of changes in wage distributions using quantile regression. Journal of applied Econometrics, 20(4), 445-465. Maddala, G.S. (1983). Limited Dependent and Qualitative Variables in Econometrics Cambridge: Cambridge University Press. Mare, D. R. (1981). Change and Stability in Educational Stratification. American Sociological Review 46 (1), 72-87. Nielsen, H. S. (2000). Wage discrimination in Zambia: an extension of the Oaxaca-Blinder decomposition. Applied Economics Letters, 7(6), 405-408. Oaxaca, R. (1973). Male-female wage differentials in urban labor markets. International economic review, 693-709. Oaxaca, R. L., & Ransom, M. R. (1994). On discrimination and the decomposition of wage differentials. Journal of econometrics, 61(1), 5-21. Orazem, P. F and E. M. King (2008). Schooling in Developing Countries: The Role of Supply, Demand and Government Policy. In Schultz, P and J. Strauss (eds.), Volume 4 Handbook of Development Economics: 3475-3559 Amsterdam: North-Holland. Project Baobab. (2009), Teaching Kenyan Youth Skills for Independence. Nairobi: Project Baobab. Rajasekaran, S., & Reyes, J. (2019). Evidence on School Engagement and Re-Engagement: What Works, How & Why, for Whom, and in What Contexts? Taylor, N., Thabo, M, Shindler, J & Akoobhai, B. 2010. Seeds of their struggle: the features of under- and overage enrolment among grade 4 learners in 21 South Africa, CREATE Pathways to Access, Research Monograph No 47. University of Witwatersrand Education Policy Unit. UNESCO, & Johnson, D., Phillips, D., Maclean, R. International Centre for Technical and Vocational Education and Training (UNESCO-UNEVOC); University of Oxford. Conflict and Education Research Group (CERG). (2007). Education for livelihoods and civic participation in post-conflict countries: conceptualizing a holistic approach to TVET planning and programming in Sub-Saharan Africa: a discussion paper. Yun, M. S. (2005). Hypothesis tests when decomposing differences in the first moment. Journal of Economic and Social Measurement, 30(4), 295-304. 30 Annexes Annex A. Preliminary Analysis of Phase II Youth and Labor Market 27. The Central African working-age population (15-64 years) profile is characterized by low levels of educational attainment. On average, about a third of the working age population has no formal education –a concern for future economic growth and global competitiveness. Women, poor households and individuals living in rural areas are the most disadvantaged groups as they tend to register lower levels of educational attainment. The distribution of the working age population by gender, wealth quintile and geographic location, reveals the following: (i) there are relatively more women (39 percent) who have not received a formal education in comparison to the than national average (32 percent); (ii) the working-age population living in rural areas have a high proportion of individuals with no formal education (41percent); and 48 percent of the working-age population from poor households has no formal education; and (iii) there are large regional variations as close to half of the adult cohort (25-64) has no formal education in region 2 and region 3 (Figure B.1 and Figure B.2). Overall, the low educational attainment of the central African working-age population suggests that the labor market is likely composed of workers with low skills who have not completed primary education, and this increases the likelihood of working in the informal and weak productivity sectors. 28. The current educational composition of the central African working-age population (15-64 years) indicates that the profile of workers in CAR is relatively better in comparison to countries with similar income levels. For example, 32 percent of the central African working-age population has no formal education as opposed to Mali, where 65 percent of the working-age population has no formal education. Figure A 1:Adult Educational Attainment by Socioeconomic Group and Regional Comparison No education Some primary Some secondary Higher 2% 1% 0% 3% 5% 1% 5% 3% 13% 10% 20% 17% 17% 28% 23% 42% 15% 37% 38% 44% 53% 37% 29% 38% 26% 32% 66% 48% 25% 45% 41% 39% 32% 33% 21% 12% Rural Female Poorest Richest National DRC Mali SSA LIC average Average CAR Comparison Source: ENMC 2018 31 Figure A 2: Youth and Adult Educational Attainment by Region No formal education Some primary Some secondary Some higher 100% 90% 80% 70% 60% 50% 40% 30% 56% 48% 20% 36% 34% 35% 36% 39% 32% 10% 24% 18% 17% 14% 0% 5% 7% Region1 Region2 Region3 Region4 Region5 Region6 Region7 Region1 Region2 Region3 Region4 Region5 Region6 Region7 Youth (15-24) Adult (25-64) Source: ENMC 2018 29. Transitioning from school into work is usually difficult in CAR and most youths tend to engage in low productive sectors and informal types of employment. Although poor educational outcomes are not the only reasons for poor employment outcomes, the low educational outcome of the youth could determine labor market orientations. Many youths tend to work in the agriculture sector as it requires less experience and low levels of skills, but the types of jobs in this sector is usually associated with non- wage household enterprises. In addition, many youths in CAR tend to engage in informal sector20, relative to the SSA average. For example, 78 percent of the youth aged 15-19 years works in the agriculture sector and 53 percent in non-wage household enterprises (Figure B.3). With an increasing share of the youth population outside of formal employment markets, CAR should expect continued economic stagnation, sluggish public revenue growth, and high expenditure needs. Figure A 3: Youth and Adults Sector of Employment and Employment Status 20 “More than half of youth in Sub-Saharan Africa are out of school, and youth participation in formal employment markets is very low� Facing Forward et al... 32 Agriculture Industry Services Wage Non-wage HH enterprise Farming 14% 19% 22% 21% 8% 6% 7% 7% 45% 60% 63% 68% 78% 75% 71% 73% 53% 35% 23% 30% 2% 5% 9% 7% Age 15-19 Age 20-14 Age 25-64 Age 15-64 Age 20-14 Age 15-19 Age 25-64 Age 15-64 Source: ENMC 2018 30. The youth and adult cohorts working in the agriculture sector are characterized by low levels of education. On average, 77 percent of the youth population (15-24 years) and 71 percent adult population (25-64 years) work in the agriculture sector (Error! Reference source not found.). Individuals with better with higher levels of education tend to engage in formal and more productive sectors. For example, 88 percent of youth with no formal education tends to work in the agriculture sector as compared to only 18 percent for those with some tertiary education level. The corresponding proportions for the adult cohorts are 84 percent and 8 percent respectively. 31. Formal wage employment tends to be very limited for both youth and adult cohorts and education tends to play a major role in the distribution of employment by sector and employment status. The proportion of the youth population engaged in wage employment is very low; on average only, 3 percent of the youth is working in wage employment compared with 9 percent for the adult cohort (Figure 6). Although wage employment is lower for those with no or low levels of education, about 50 percent of the adult cohort with some higher education are engaged in wage employment compared with only 16 percent for the youth cohort. In fact, many youths who complete their formal education also face underemployment. Thus, programs that target better education and employment outcomes for out-of- school youth must tackle multiple problems. To be most effective, broad-based programs should focus on increasing opportunities for employment for everyone—not just a certain set of youth—which can only be achieved through better economic policies that promote growth (Bashir et al 2018). Figure A 4: Youth and Adults Sector of Employment and Employment Status by Level of Education 33 Agriculture Industry Services Wage Non-wage HH enterprise Farming 88% 84% 79% 78% 77% 71% 58% 48% 50% 18% 16% 20% 8% 9% 2% 3% 4% 3% 3% 3% Ave. Ave. Some primary Some higher Some primary Some higher Some secondary Some secondary No formal education No formal education Ave. Ave. Some higher Some higher Some primary Some secondary Some primary Some secondary No formal education No formal education Youth (15-24) Adult (25-64) Youth (15-24) Adult (25-64) Source: ENMC 2018 32. Although agriculture remains the dominant sector of employment across SSA countries regardless of income status, low income countries tend to engage more in agricultural activities and the Central African Republic does so even more. On average, 73 percent of the central African labor force work in the agriculture sector, whereas the proportion of the labor force in this sector is 72 percent for low income countries and 62 percent for SSA countries (Figure B.5Error! Reference source not found.). The opportunity for employment outside agriculture is limited in CAR even for those with better educational attainment. 33. The Central African Republic’s economy tends to be more informal than other low income and SSA countries. The share of the central African’s labor force in wage employment stands at only 7 percent, which is below the average for low income countries. The chance of finding wage employment is even low for individuals with better educational attainment. For example, only 14 percent of the central African’s workforce with some secondary education are engaged in wage employment as opposed to 24 percent for the low-income countries and 37 percent for SSA countries. Lastly, wage employment is mainly concentrated in Region 7, which is the capital city of the nation, Bangui (Annex Figure A.12). Figure A 5: Comparison of Employment Status and Sector of Employment by Level of Education 34 Agriculture Industry Services 15 23 23 20 21 20 21 25 28 7 43 40 39 35 6 7 48 46 7 8 7 9 12 10 7 10 10 15 9 12 78 72 73 73 72 72 65 65 62 59 42 50 43 42 50 SSA CAR DRC Mali LIC SSA CAR DRC Mali LIC SSA CAR DRC Mali LIC Ave. ave. Ave. ave. Ave. ave. Some primary Some secondary All level of education Wage Non-wage HH enterprise Farming 41 37 43 42 50 65 65 61 63 59 69 72 72 72 71 22 26 50 34 27 17 17 26 25 30 18 28 24 18 37 32 19 18 23 24 22 10 10 14 15 9 11 3 5 7 SSA CAR DRC Mali LIC SSA CAR DRC Mali LIC SSA CAR DRC Mali LIC Ave. ave. Ave. ave. Ave. ave. Some primary Some secondary All level of education Source: ENMC 2018 Programs in Place for Remedial of Out-of-School Youth 34. This section provides some examples of skills training interventions in SSA. Specifically, one of the main objectives of this study is to propose, based on what has already been done in similar contexts, adequate programs or skills training relevant to the labor market for the large number of out-of-school youth in CAR. Some of the key programs or policies in fragile and conflict affected contexts are summarized below. The review relies mainly on the literature on the determinants of out-of-school youths and programs that were recently conducted such as the Out-of-School Youth in Sub-Saharan Africa, A Policy Perspective (Keiko Inoue, Emanuela di Gropello, Yesim Sayin Taylor, and James Gresham, 2015) and the Back to School, Pathways for reengagements of out-of-school in education (Subhashini Rajasekaran and Joel Reyes, 2019). The following section provides some examples of programs that worked well and summarizes key programs focusing on TVET and skills development programs in different fragile countries in Sub-Saharan Africa (Box 1). 35 35. The out-of-school youth study in SSA identified three entry points to address the out-of-school youth problem. The first entry point is retention of at-risk youth in school. Given that most youth drop out before they start secondary school, retention efforts must begin before youth enter secondary education. Retention could be improved through greater early intervention to get children enrolled at the right age and a renewed focus on improving the quality of primary education. To retain students who are interested in learning but cannot afford it, cash incentives or subsidies can be an important tool in keeping children in school. The second entry point is remediation through alternative education programs. For the youth already out of school, the most likely path to complete their education is through alternative education systems. Successful alternative programs also require a greater recognition that youth must work to survive and therefore factor in the necessary flexibility to accommodate this. The third entry point is integration with the labor market. Youth who are not likely to go back to school require practical training and experience to increase their employability. Workforce development programs must consider, in terms of skills and services offered, that most youth will be self-employed or work for a small, informal enterprise. They require coordinated action between government actors, regional entities, NGOs and the productive sector. More and better impact evaluations and longevity of effective workforce development programs will help improve their outcomes. 36. In connection to factors that can be associated with young people leaving education early the study by Subhashini Rajasekaran and Joel Reyes, 2019 identify two key risk factors: (i) individual risk factors which include, demographic, geographic, cognitive, emotional and social and behavioral; and (ii) contextual risk factors; this set of risk factors relates to family and household, school, community, and macro-level structural factors. The authors highlighted that monitoring and psychosocial support and learning are key programs that wieldy works in reengaging out-of-school youth. 37. Several programs have been implemented in SSA countries to address similar issues. For example, Rwanda implemented Akazi Kanoze Youth Livelihoods program between 2009-2016 to address the issues of lack of required skills to enter the labor market for out-of-school youth, especially those from rural areas. This program targeted more than 21,000 in-school and out-of-school youth ages 14 to 24. The main objectives of this project were to provide out-of-school youth with market-relevant life and work readiness training and support, hands-on training opportunities, and links into the employment and self- employment job market. For youth who have successfully completed the work readiness curriculum, the project helped them to launch their own income-generating activities. Through this program, 222 micro and small businesses have received business development services and more than 2,400 new businesses have been launched. About 91 percent of employers stated they were satisfied with the Akazi Kanoze participants they hired as employees or interns. More than 50 percent of enrolled youth were employed six months after graduation, either running their own business or working at a Rwandan company. The Akazi Kanoze work readiness curriculum has been integrated into all of Rwanda’s TVET curricula. 38. Another example is Somali Youth Livelihood Program (SYLP) which was implemented to strengthen the professional skills of young people aged 15 to 24 years. Technology was an important feature of this program. In four years, the program served a total of 10,573 youths. About 78 percent of trainees were attracted by external employers and 52 percent of those with entrepreneurship training 36 were placed in companies or found jobs. More than 60 percent of participants said that training improved their future or independent job prospects. This program has been successful for women. Indeed, about 41 percent of participants in entrepreneurship training were women and 37 percent represented the vocational training component. 41. Another successful program is Uganda’s Empowerment and livelihoods for adolescents (ELA) program which reached around 31, 000 members organized into 500 'adolescent clubs' for 13 to 22 years old. The objective of this program was to empower adolescent girls both in terms of their entrepreneurial and business skills, and in terms of their social capacity. This program had a stronger emphasis on financial literacy, livelihoods and microfinance. One of the goals of this program was to strengthen girls' ability to lead independent and dignified lives and become active agents of social change in their own families and communities. This program provided hundreds of thousands of girls the opportunity to live a better life through mentoring, life skills training and microfinance, and targeted only vulnerable adolescent girls. ELA program has been tested in several African countries (Uganda, Tanzania, Sierra Leone, South Sudan and Liberia) and its positive effects have been demonstrated. For example, in Uganda, those who were not enrolled are 11% more likely to intend to go back to school. The program also decreased the teenage pregnancy rate by 20 to 25 percent in the Ugandans villages with an ELA program. In addition, HIV knowledge improved by 11 percent compared to the baseline for girls in the targeted villages. Beneficiaries are 7 percent more likely to know the risk of pregnancy and 21 percent more likely to use a condom during sex. Another important result of this program is the increase in savings levels among girls. Indeed, girls in targeted villages are 12 percent more likely to make savings and their savings are 17 percent higher than the baseline. Prior to the ELA program, only 1 percent of girls had a bank account, which rose to 8 percent in the intervention sample follow-up survey. Girls in targeted villages are also 4 percent more likely to be engaged in an income generating activity, 5 percent more likely to be self- employed, and spend 16 percent more time working outside the house. 42. Another program that has been implemented in Kenya is Project Baobab. The objective of this project was to educate and empower 1600 Kenyan youth to become successful entrepreneurs. Project Baobab provided entrepreneurial education and grants to youth, especially women, through a specialized life skills and business curriculum in Kenya. The results of this project show that since 2001, project Baobab has educated over 1000 students and funded approximately 300 new business ventures. In addition, over 1,300 students have completed the Project Baobab program. Of those, over 350 have been awarded micro-grants of $100 each. Each year, between 70 and 90 students graduate, and about 30 percent receive a US $100 grant to start new businesses. About 50 percent of the beneficiaries operate businesses whose success is evolving gradually. 43. The Sierra Leone program on Youth Reintegration Training and Education for Peace (YRTEP) is another example of out-of-school youth program in fragile contexts. The target population of this program was ex-combatants and other war-affected youth who were to be provided with non-formal education activities in reintegration, livelihood skills development, remedial education and basic literacy and numeracy skills as appropriate. The program targeted about 45,000 ex- combatants and war-affected youth ages 15-34. About 98 percent reported better reading and writing skills; 85 percent had planted 37 crops; 40 percent started a business; 33 percent enrolled in an apprenticeship; 43 percent re-enrolled in school or another vocational training program; 12 percent are employed for a business or company. 44. Another program that has been put in place is the Liberia Skills Training Program for the Economic Empowerment of Adolescents and Young Women (EPAG). The program was launched by the Liberian Ministry of Gender and Development in 2009 and it aimed to increase the employment and income of 2,500 young Liberian women by providing livelihood and life skills training and facilitating their transition to productive work. The intervention of EPAG consisted of a six-month phase of classroom-based training, followed by a six-month placement and support phase in which the trainees were supported in their transition to self or wage employment. This program targeted young women with 16 to 27 years old. The program increased employment by 47 percent and earnings by 80 percent and positive effects on access to money, self-confidence, and anxiety about circumstances and the future for beneficiaries. It also had a positive effect on a variety of empowerment measures, including access to money, self-confidence, and anxiety about circumstances and the future. 38 Box 1 Box 1: Evaluation of selected TVET and skills development programs in different fragile countries Coun try Program Targeted Issues/Problems Program/interventions Key outcomes and name group Year Business development services were received by 222 Many young micro and small businesses. 91% of employers stated Akazi Market-relevant life skills; 21,039 in- Rwan people— they were satisfied with the Akazi Kanoze participants Kanoze: Apprenticeship or on-the- school and da, especially in they hired as employees or interns. More than 2,400 Youth job training; Job match or out-of- 2009- rural areas of the new businesses have been launched. Livelihood mediation; Institutional school- youth 2016 country—lack More than 50% of enrolled youth were employed six s Project capacity building (ages 14-24) the needed skills months after graduation, either running their own business or working at a Rwandan company. 3,000 South High adolescent Participants show high levels of engagement and Adolescen Suda unemployment Skills training; Life skills girls and confidence, reflected in increased awareness of such t Girls n, rate for young training; Microfinance young issues as protection against rape, early pregnancy, Initiative 2013 women women (age contraception and family planning in general, and 15 to 24) increased knowledge and awareness of HIV/AIDs. Basic education (accelerated learning); Life skills General Youth population Somali training on Soma lacks education 78% of youth participants who received vocational Youth entrepreneurship; Access to 10,573 Youth lia, and skills needed training were placed with outside employers and 52% Livelihood youth friendly loans or aged 15-24 2008- to engage into of those in entrepreneurship training were placed in Program stock; Financial literacy; years 2011 productive labor businesses/employment. (SYLP) Apprenticeships or on-the- market activities job training; Job match and mediation; Mentoring ICT Reintegrating Landmine Apprenticeship or on-the- Participants were at least a quarter more likely to be Liberi 1330 ex- and Employing Action job training; Classroom engaged in agriculture, and almost a third more likely a, combatants High Risk Youth agricultur vocational training; General to have sold crops. 39 2009- al training training on or high-risk 2011 program entrepreneurship; Life skills; youth for ex- Vouchers; Access to youth combatan tailored loans or stock; Basic ts education; Psychosocial training or counseling Six-month training in either job skills targeted to sectors with high demand, or Economic business development skills; Empower six-month support for job Young women ment of Increased employment by 47 percent and earnings by Liberi placement or links to micro- 2,500 young face barriers to Adolescen 80 percent and positive effects on access to money, a, credit, and other training women aged entering the t Girls and self-confidence, and anxiety about circumstances and 2009 and support, such as life 16-27 labor market Young the future. skills training, small group Women learning, a business plan (EPAG) competition, mentorship, savings accounts, child care, and transportation Due to limited Participants reported an increase in the awareness of resources, methods to prevent diseases, such as HIV/AIDS and tensions remain Youth 13,391 war- Liberi malaria.; and saw an increase in knowledge and high in many Education Life skills; Civic engagement; affected a, change in attitudes regarding HIV/AIDS post- communities for Life Recreational activities youth 2005 program. The civic engagement activities seemed to between Skills (YES) ages18-30 serve as conduits of cooperation between the youth combatants and and the elders in the community. returnees Adolescents girls Life/ soft skills training Empower HIV knowledge improved by 11% relative to its face multiple Entrepreneurship training; Ugan ment and 31,000 baseline level among girls in targeted villages, girls in challenges Advisory services (e.g. da, livelihood adolescent targeted villages are 7% more likely to know about including health mentoring, business 2008- s for girls aged 13- pregnancy risk and 21% more likely to use a condom and economic development services, or 2010 adolescen 21 during sex. Self-assessed entrepreneurial ability is ones associated business formalization); ts (ELA) 10% higher among girls in targeted villages and they with early Access to microfinance 40 pregnancy, are 12% more likely to have savings, and their savings unemployment are 17% higher. Life/ soft skills training (in classroom and at the workplace); High Keny Entrepreneurship training unemployment Project About 50% of the grantees are running businesses a, Advisory services (e.g. 1000 youth rate among Baobab with good-to-marginal success. 2008 mentoring, business youth development services, or business formalization); Access to microfinance Life/ soft skills training (in Program classroom and at the The incomes of PCY participants were about 26% Disadvantage High for the workplace); higher than those of other community members. Ugan d youth living un/underemploy promotio Entrepreneurship training; Furthermore, for youth promoters/youth group da, in rural areas ed among youth n of Advisory services (e.g. members their main sources of income are from 2003- and currently living in rural children mentoring, business salaries (23%) and from youth group activities 2006 un/underem areas and youth development services, or (38.5%) while other community members are still ployed (PCY) business formalization); Job mainly engaged in subsistence farming (76%) counselling Young women The and girls suffered WINGS Ugan 1800 young economically program Business skills training (BST); da, women aged and in Cash transfers; Regular For the average WINGS beneficiary, monthly cash 2007 15-35 educationally Northern follow-up by trained income increased by 98% and a 33% increase in from the war Uganda community workers householdspending. Keny Disadvantage a, d African About 61% of participants were placed in jobs by the Niger High rate of Youth youth, ages program. Between 52% and 94% of youth surveyed Job match and mediation; ia, youth Empower 16-35: 3,300 found jobs (dependent or self-employment) and/or Life skills; General training Sene unemployment ment in Kenya; participated in internships, community service, or on entrepreneurship gal, in the region Program 2,500 in went back to school. 9% of program graduates Tanza Nigeria; surveyed were operating small businesses. nia, 1,900 youth 41 2007- in Senegal; 2010 and 1,500 Tanzania. Classroom vocational skills Youth High risk for training; General training on 45,000 ex- Sierra Reintegrat About 98% reported better reading and writing skills; youth to become entrepreneurship combatant Leon ion 85% had planted crops; 40% started a business; 33% more susceptible (agriculture); Life skills; and war- e, Training enrolled in an apprenticeship; 43% re-enrolled in to negative and Health education; Basic affected 2000- and school or another vocational training program; 12% violent education Civic education; youth ages 2001 Education are employed for a business or company. influences Conflict mediation, peace- 15-34 for Peace building High Technical training in unemployment 700 out-of- Ninaweza information and The program increased the likelihood of its for young girls school and Keny Youth communications technology beneficiaries obtaining a job by 14 percent and were due to unemployed a, Empower (ICT), training in life skills, more likely to hold full-time employment positions difficulties in young 2011 ment work experience through while those who did not benefit from the program initial workforce women aged Program internships, and job were more likely to work as casual laborer. entry and lack of 18-35 placement support skills Eritre Youth at risk do Girls’ Tutoring to girls 1,440 No evaluation was done, but the girls who a not stay or Education transitioning to secondary academically participated in the project were able to catch up in return to school. Project education and to those in poor girls basic subjects at school and, in general, were secondary school who are promoted from one class to another. academically weak South Multiple psycho- USIKO School-based remedial 1200 At-risk More than 600 at-risk youth have successfully Africa social problems Program support and mentoring that youth (youth completed the mentorship program. Among the , of young people focuses both on academic aged 12 – 18 members of this group, more than 90 percent 2001 in the area which and noncognitive skills. years of age) obtained high school qualifications. placed youth at severe risk to truancy, school dropout, criminal behavior and early death. 42 Mali, Lack of Support Provide literary skills to 105 The literacy program had served 17,637 basic literacy 2003- community for the parents which can help communities learners (of whom 6,524 were female); 6,831 post 2008 involvement to Quality them to become a part of with literacy learners (of whom 2,260 were female). support schools and Equity their children’s learning 2,430,676 Parents and community members had participated in and children's of process inhabitants 756 school management committees. In addition, the education Education project served approximately 700 schools in his last Program year (the 2007–08 school year). Tanza Reductions in Complem Provide education to out-of- Out-of- No evaluation of the program has been conducted, nia, school entary school children, with an school youth an earlier review of the pilot phase found that the 1999 enrollment due Basic emphasis on girls. The aged 11-18 COBET program did not meet its stated goals of to deteriorating Education program provides numeracy years (xx expanding enrollment among girls. school quality in and literacy skills for youth beneficiaries) Tanzania aged 11–18 years and (COBET) vocational training for the program older cohort. Ugan Lack of Non- Basic Offer a condensed 5,884; Youth More than 3,000 students served through 54 centers da Formal basic Education curriculum of three years aged 9–18 in Kampala, 55 percent were girls, more than a Education to for Urban compared with the five-year quarter transferred to formal schools, and only about disadvantaged Poverty basic education cycle 10 percent dropped out children and Areas youth in the (BEUPA) poor urban Program areas. Mada Obtain their Planet of Choose to complete the Out-of- Initial evaluations showed that more than two-thirds gasca basic education Alphas, program and sit for an school youth of the students passed the final test, approximately r credentials and equivalency test for a aged 10–18 one in five enrolled in formal primary schools, and Accelerate diploma or return to school years (xx about one in 20 enrolled in the ASAMA Program d if they can demonstrate beneficiaries) Compress grade-level knowledge in a ed placement test. Learning for Malagasy Adolescen 43 ts (ASAMA) Sierra Lack of Support to A combination of basic 2,500 Out-of- In its pilot phase, program participants generally Leon Community Strengthe literacy and skills, vocational school completed the coursework. e Education n the education, civic education, children and Centres for Capacity and life skills that focus on functionally Literacy and of the financial responsibility, illiterate Vocational Skills Communit gender relations, and youths and for Women and y health. adults, Girls Education particularly Centres those from for deprived Literacy social and backgrounds, Vocational such as war Skills for victims, Women returnees, and Girls orphans, refugees, and internally displaced persons. Ghan Lack of literacy National It mixed functional literacy Adults (15-45 The program served more than 2 million adults a and functional Functional education with occupational years), (including its earlier form, initiated in 1992), and skills Literacy skills, life skills, and health particularly volunteers kept costs low. Program education. women and rural poor (xx beneficiaries) Liberi Youth with Advancing Provide alternative Youth aged It reached out to the most vulnerable groups in the a, limited the education to out-of-school 13–35 (xx country, and achieved consistently high attendance 2005 education and Youth’s in livelihood training. The beneficiaries) rates employable skills Alternativ program supported e community-based training in reading, writing, and math 44 Education as well as work readiness, Project health, and life skills. Ethio Lack of training Ethiopia’s Combine literacy training Adults and The program trains approximately 2,000 people pia, center (mainly Communit with a soft skills and out-of-school annually. Three-quarters of the participants in 2000 community / y Based entrepreneurial skills children who training modules were females. There is no CBEC vocational Nonforma program. The program was have never evaluation of the impact of the program on the training centers, l designed with the goal of completed employment of trainees, but trainees have reported vocational Livelihood linking training with access their formal improvement in their qualifications as well as in their training centers / Skills to credit to improve job education, motivation to engage in income-generating activities. vocational Training outcomes. especially training centers, for Youths those who TVET centers in and Adults live in rural rural areas) (EXPRO) areas of Program extreme poverty. (xx beneficiaries) Keny Youth have Technical Measure the effectiveness 2,160 out-of- Offering young adults’ vouchers that cover program a, dropped out and of training vouchers. school youth costs encouraged them to enroll, and that those who 2008 before eighth Vocational (ranging in can use the voucher for a private training program grade and have Vouchers age from are more likely to sign up and stay in school. not completed Program roughly 18– Participants who had not completed secondary secondary 30) education were less likely to drop out compared to education. those who began the program with a secondary degree, suggesting that the less educated youth attributed a greater value to vocational training Ugan High rate of Non- Practical skills in specific Out-of- 288 vulnerable youth were placed with local Artisans da, vulnerable and Formal trades including school and for training in rural and urban areas. The number of 2005- marginalized Education hairdressing, tailoring, socioeconom local Artisans trained in delivery of HIV/AIDS 2009 youth affected and motor mechanics, ically messages is now 98 in Kampala and Arua. by HIV/AIDS and Livelihood carpentry, electronics, vulnerable other risk Skills for welding, and cookery. Life youth (xx behavior Marginaliz skills training with a focus beneficiaries) ed Street on HIV/AIDS, reproductive and Slum health, nutrition, child 45 Youth rearing, and drug and Training alcohol abuse. Program Soma There is no Horn of Provide vocational training Out-of- The program has trained approximately 3,000 youth. liland return to school Africa for youth intending to school youth , from youth. Voluntary return to school 1998 Youth Committe e Ugan High rate of Youth Offering up to US$10,000 in Unemployed Groups that received community grants: were almost da unemployed or Opportuni grants to youth groups that or four times more likely to participate in vocational underemployed ties submit proposals in which underemploy training and twice as likely to be engaged in skilled youth Program they identify a vocational ed youth work; improved their profits by 50 percent compared skill of interest and a aged 15–35 with the control group; and increased their savings by vocational training institute. years living in 20 percent. They were more likely to engage in civic conditions of activities and less likely to engage in aggressive poverty. (xx activities (especially men) beneficiaries) Rwan Lack of youth Alliance Training and apprenticeship Out-of- The program trained approximately 900 youth in da, employment for Youth school and trades such as hospitality, carpentry, driving, South promotion Employabi at-risk youth tailoring, salon, media, studio arts, and also provided Africa lity aged 14–29 business start-up assistance. In Rwanda and Tanzania, , and years (xx the Orphans and Vulnerable Children (OVCs) who Tanza beneficiaries) became employed because of the project were able nia to support their younger siblings, providing their basic needs and enrolling them in school 46 Mapping of the TVET Curriculum Education sector policy 45. The Central African Republic’s government has begun providing a strategic direction to skills development and a significant amount of national policy debate has been focused around Technical and Vocational Education and Training (TVET) reforms, resulting in a series of policy documents related in whole or in part to TVET. However, only the 2018 report on national strategy on TVET (Stratégie Nationale de l’Enseignement Technique et de la Formation Professionnelle en Centrafrique) planned to develop a national TVET strategy. These reforms include: (i) The National Strategy on TVET in CAR (La Stratégie Nationale de l’Enseignement Technique et de la Formation Professionnelle en Centrafrique); (ii) The Education Sector National Strategy (2008-2020) and the Education Sector Action Plan 2013-2015 (La Stratégie Nationale du Secteur de l’Éducation (SNSE 2008-2020) and Plan d’action du secteur de l’éducation 2013-2015); (iii) The National Policy on Employment and Vocational Training 2016 (La Politique Nationale de l’Emploi et de la Formation Professionnelle (PNEFP 2016)); (iv) The Transition Plan 2018- 2019 of the Ministry of Primary, Secondary, Technical, and Literacy Education (MEPSTA) ( Le Plan de Transition 2018 – 2019 du Ministère de l’Education Nationale et de l’Enseignement Technique); (v) The National Emergency Program to Create Decent, Immediate and Sustainable Jobs for Peacebuilding and Resilience in the CAR ( Le Programme National d’Urgence de Création d’Emplois Décents, Immédiats et Durables pour la Consolidation de la Paix et la Résilience en RCA (OIT 2017-2021)); and (vi) The 2016 National Recovery and Peacebuilding Plan (Le Plan de Relèvement et de Consolidation de la Paix 2017 - 2021 (RCPCA –2016)). 46. The Central African Republic’s government acts as a large provider of skills in the country. TVET is provided at the secondary education level by the Ministry of Technical Education and Literacy (META) and the duration is 3 years at the lower secondary and 3 years at the upper secondary. TVET is also provided in training centers by both META and the Ministry of Labor, Employment, Training and Social Protection (Ministère du Travail, de l’Emploi, de la Formation Professionnelle et de la Protection Sociale). The former focus on formal training while the later mostly concentrates on informal and specific skills training. Based on 2018 EMIS, there are 9 secondary schools that offer TVET of which 3 are managed by the public. The number of public training centers are 10 but only 4 were functioning. There are also 4 private training centers. TVET is also provided as part of the post-secondary education but access is very limited — the demand for post-secondary TVET is very high but admission is very limited. There is also a significant demand for post-secondary TVET graduates, however, there is limited information about the demand for secondary level TVET graduates. Other ministries also offer sector-specific training programs. These include the Ministry of Agriculture and Fishing, the Ministry of Youth, and the Ministry of Trade. 47. The Central African Republic’s TVET programs are divided into two cycles, a lower cycle and an upper cycle. The lower cycle is a 3-year education program which provides two certificates: (i) the Technical Aptitude Certificate (Certificat d’Aptitude Technique) which is accessible to students with at least grade 9 education level and (ii) the Vocational Training Studies (Brevet d’Étude technique) which is accessible to students with at least grade 8 education level. The upper cycle is also a 3-year education program It offers two degrees namely technical 47 baccalaureate and professional baccalaureate. The Technical Baccalaureate is available to students with at least grade 11 education level, while the Professional Baccalaureate is offered to students with grade 10 education or students who have a Technical Aptitude Certificate. There is also post-secondary TVET which offers is offered Higher Technical degree mainly in private institutions. Finally, there are vocational training offered to students who have at least grade 6 education level and a vocational training certificate is issued at the end of 2 years. 48. According to the 2017 Central African Republic’s report on the assessment of technical and vocation training centers (Répertoire des centres de formation professionnelle et technique), there are around 179 training providers in the country. Training providers are concentrated in Bangui, the capital city, where more than 82 percent (147 training providers) are based (Error! Reference source not found.). There are also clusters of training providers in Papoua (23 training providers), Begoua (11 training providers), and Bimbo (8 training providers). The geographic location of most training programs in Bangui and urban areas means that most of the youth in rural locations have limited access to these opportunities. The profile of the subjects/modules classified by sector of activities provided by the 179 TVET institutions in CAR is presented in Annex Table 2. Figure A 6: Number of Technical and Vocational Education and Training institutions by prefecture BOSSANGOA, 1 BERBERATI, 1 BOZOUM, 5 MBA�KI, 2 PAOUA, 23 BIMBO, 8 BEGOUA, 11 BANGUI, 147 Source: Authors’ estimations based data from ‘Repertoire des centres de formation professionnelle et technique 2017’ 49. Curricula were collected from all functional public and private TVET providers located in Bangui to provide a full mapping of TVET programs in CAR. The public institutions include the Technical High School of Bangui (Lycée Technique de Bangui), the Girls' Professional High School (Lycée Professionnel Féminin), the School of Arts and Crafts (Ecole des Métiers d’Arts), and the Apprenticeship Training Center (Centre de Formation et d’Insertion par Apprentissage). Some of the private institutions are the Providence Institute, Institut Moderne des Métiers Spécialisés (IMMS) and Don Bosco. Annex Table 4 summarizes the different fields of study provided by the main TVET institutions in CAR. As shown, the majority of training programs are in auto mechanics, masonry and electricity which represent 71 percent of programs offered in the surveyed TVET institutions. Next are carpentry 48 programs which represent 57 percent of training programs; and plumbing and pipping representing 43 percent. Although most of the employment in CAR is predominantly in the agriculture sector, only one private TVET institution (Don Bosco Damala) offers agriculture and livestock-related training (Error! Reference source not found.). 50. TVET provision is limited and the number of student enrollment is relatively lower, as only 4,000 students were enrolled in TVET institutions. Error! Reference source not found. shows enrollments in the main TVET institutions in CAR. The Technical High School of Bangui (Lycée Technique de Bangui) is the TVET institution with the highest number of student’s enrollment (4030 students), followed by the Women's Professional High School (Lycée Professionnel Féminin) with 897 female students enrolled. Among private TVET providers, Providence is the institution with the highest number of students. In 2017, there was about 1700 enrolled in this private TVET institution. Figure A 7: Enrollment by main TVET institutions in 2017 Providence 1700 Private TVET Institut Moderne des Métiers Spécialisés (IMMS) 500 DON BOSCO DAMALA 407 École des Métiers d’Arts (EMA) 329 Public TVET Lycée Professionnel Féminin 897 Lycée Technique 4030 Source: Authors’ estimations based on data collected from Strategie Nationale de l’Enseignement Technique et de la Formation Professionnelle en Centrafrique, 2018. Figure A 8: Types of training courses provided by TVET institutions (number of institutions) 49 Electricity (Auto and/or building ) 5 Auto Mechanics 5 Masonry 5 Carpentry 4 Plumbing - Piping 3 Computer (office and maintenance) 2 Cutting and sewing 2 Tiling 2 Joinery 2 Painting (decorative art) 2 Drawing 2 Hotel Industry 2 Accounting 2 Metal Construction 2 Refrigeration and air conditioning 1 Agriculture and Livestock 1 Mechanic drivers 1 Furniture 1 Wood carving 1 Stationery 1 Gilding 1 Binding 1 Art ironwork 1 Leather goods 1 Jewelry 1 Cleaners 1 Hotel and Catering 1 Family and Social Economy 1 Administration and Management 1 Administrative and Commercial Services 1 Construction 1 General Mechanics 1 Technical Baccalaureate 1 0 1 2 3 4 5 6 Source: Authors’ estimations based on data collected from Stratégie Nationale de l’Enseignement Technique et de la Formation Professionnelle en Centrafrique, 2018. Focus Group Discussions: TVET Providers and Employers 51. Focus group discussions with TVET providers were carried out in November 2018 and February 2019 to explore the context of the supply of TVET in CAR and to have a better understanding of the main challenges and priorities of TVET providers and employers. Participants included: representatives from the META, principals from public and private technical and vocational schools and training centers representatives from the agency for vocational training and employment (Agence Centrafricaine pour la Formation Professionnelle et de l’Emploi), representatives from key private firms. The main objectives of the focus group discussions were to discuss key challenges and priorities of these TVET institutions and the perception of employers on TVET trainees’ qualification. The key challenges cited by different actors are summarized as follow: (i) limited capacity and resource to accommodate the high demand for technical and vocational education and training; (ii) significant infrastructure and equipment needs; (iii) lack of training curricula; (iv) inadequacy of teaching materials; (v) high 50 degree of mismatch between courses offered and skills to respond to the demand of prospective employers; (vii) lack of specialized trainers and little to none teacher training programs; (viii) lack of analysis and feasibility studies to identify training needs of the labor market; (ix) lack of consultations between TVET stakeholders on the development TVET curricula; (x) limited employment opportunities for trainees; and (xi) lack of formal training curricula for out-of-school youth. Annex B. Figures Figure B. 1: Trends of GDP per capita growth, CAR and comparators 20.0 10.0 11.4 2.7 1.7 1.7 3.7 2.2 1.9 0.2 2.4 4.4 4.0 3.8 3.0 3.1 0.2 4.1 3.4 2.2 2.0 2.7 1.5 2.7 2.9 2.5 0.0 0.4 1.2 0.7 0.8 0.0 0.6 0.5 -1.3 2009 2010 2011 2012 2013 2014 2015 2016 -1.5 -0.2 2017 -7.8 -10.0 -20.0 Sub-Saharan Africa (excluding high income) -30.0 Central African Republic -36.8 Low income -40.0 Source : WDI 2009-2017 Figure B. 2: Trends of Primary GER by Socioeconomic Status 51 138% 128% 125% 133% 125% 127% 120% 118% 114% 112% 109% 108% 111% 111% 102% 103% 101% 98% 98% 98% 89% 92% 92% 91% 80% 79% 82% 83% 81% 72% 70% 64% 55% 51% 40% 2006 2010 2000 2006 2010 2016 2018 2000 2006 2010 2016 2018 2000 2006 2010 2016 2018 2000 2010 2016 2018 2000 2006 2010 2016 2018 2000 2006 2016 2018 2000 2006 2010 2016 2018 National Male Female Rural Urban Q1 Q5 Source: Estimations based on MICS and ENMC. Figure B. 3: Trends of Lower Secondary GER by Socioeconomic Status 68% 58% 56% 54% 46% 41% 36% 36% 36% 30% 30% 32% 28% 26% 24% 23% 22% 20% 19% 21% 17% 16% 18% 17% 11% 13% 13%11% 9% 3% 6% 5% 6% 7% 1% 2000 2006 2010 2016 2018 2000 2006 2010 2016 2018 2000 2006 2010 2016 2018 2000 2006 2010 2016 2018 2000 2006 2010 2016 2018 2000 2006 2010 2016 2018 2000 2006 2010 2016 2018 National Male Female Rural Urban Q1 Q5 Source: Estimations based on MICS and ENMC. Figure A.4: Trends of Upper Secondary GER by Socioeconomic Status 52 87% 2000 2006 2010 2018 41% 41% 42% 37% 28% 23%23% 21%20% 20% 17% 15% 9% 9% 1% 3% 0% 3% 5% 4% 5% 4% 3% 3% 3% 2% 5% Region4 National Femal Urban Region1 Region2 Region3 Region5 Region6 Region7 Rural Male Q1 Q5 Gender Area Wealth Quintile Region Source: Estimations based on MICS and ENMC. Figure B. 4: Higher Education GER by Socioeconomic Status 4% 4% 2% 2% 2% 1% National Male Female Rural Urban Q1 Q5 Gender Area Wealth Quntile Source: Estimations based on 2018 ENMC. Figure B. 5: Trends of Primary GER by Region 53 142% 138% 139% 129% 129% 123% 123% 114% 117% 117% 114% 106% 105% 108% 105% 99% 101% 100% 100% 91% 89% 87% 85% 79% 77% 79% 76% 77% 70% 63% 64% 63%65% 61% 59% 2000 2006 2010 2016 2018 2000 2006 2010 2016 2018 2000 2006 2010 2016 2018 2000 2006 2010 2016 2018 2000 2006 2010 2016 2018 2000 2006 2010 2016 2018 2000 2006 2010 2016 2018 Region 1 Region 2 Region 3 Region 4 Region5 Region 6 Region 7 Source: Estimations based on MICS and ENMC. Figure B. 6: Trends of Lower Secondary GER by Region 81% 71% 63% 46% 34% 30% 31% 31% 29% 27% 26% 27% 26% 22% 23% 19% 19% 14% 13%14% 15% 13% 10% 11% 11% 12% 6% 7% 7% 8% 7% 8% 5% 6% 9% 2010 2000 2016 2006 2018 2010 2000 2016 2000 2006 2010 2016 2018 2000 2006 2016 2018 2006 2010 2018 2000 2010 2016 2000 2006 2016 2018 2006 2010 2018 2000 2006 2010 2016 2018 Region 1 Region 2 Region 3 Region 4 Region5 Region 6 Region 7 Source: Estimations based on MICS and ENMC. 54 Figure B. 7: Trends of Upper Secondary GER by Region 87% 42% 41% 20% 16% 12% 8% 7% 9% 5% 6% 4% 5% 4% 4% 4% 4% 4% 3% 2% 1% 1% 2% 1% 1% 1% 2% 3% 2000 2006 2006 2010 2018 2000 2006 2010 2018 2000 2006 2010 2018 2000 2006 2010 2018 2000 2006 2010 2018 2000 2010 2018 2000 2006 2010 2018 Region 1 Region 2 Region 3 Region 4 Region 5 Region 6 Region 7 Source: Estimations based on MICS and ENMC. Figure B. 8: Trends of primary school age out-of-school by categories 55 77 Never in school Dropout 67 64 60 56 53 54 48 49 43 44 41 38 37 39 25 34 33 33 32 33 29 30 29 29 29 27 27 26 22 21 19 19 16 13 2006 2006 2000 2010 2016 2018 2000 2006 2010 2016 2018 2000 2006 2010 2016 2018 2000 2006 2010 2016 2018 2000 2006 2010 2016 2018 2000 2010 2016 2018 2000 2006 2010 2016 2018 National Male Female Rural Urban Q1 Q5 Source: Estimations based on MICS and ENMC. Figure B. 9: Trends of Lower secondary school age out-of-school by categories Never in school Dropout 18 17 21 15 16 21 20 17 18 16 14 14 14 52 16 43 20 11 12 42 17 19 19 15 41 19 33 15 15 16 10 28 30 26 12 24 9 16 12 13 10 22 14 13 22 20 19 16 6 15 15 7 8 8 12 12 10 10 12 7 10 13 7 5 5 5 3 5 3 5 2016 2006 2000 2006 2010 2016 2018 2000 2006 2010 2018 2000 2006 2010 2016 2018 2000 2006 2010 2016 2018 2000 2010 2016 2018 2000 2006 2010 2016 2018 2000 2006 2010 2016 2018 National Male Female Rural Urban Q1 Q5 Source: Estimations based on MICS and ENMC. 56 Figure B. 10: Trends of Youth Age 15-24 out-of-school by categories Never in school Dropout 39 45 41 45 50 44 51 49 47 56 46 60 43 50 59 57 57 62 45 45 46 60 61 61 56 42 51 50 44 45 38 37 43 54 37 46 49 47 42 37 35 34 33 37 33 34 25 23 21 29 25 26 24 23 23 20 18 14 10 12 10 17 16 14 12 11 11 8 8 6 2006 2010 2000 2006 2010 2016 2018 2000 2006 2010 2016 2018 2000 2006 2010 2016 2018 2000 2010 2016 2018 2000 2006 2016 2018 2000 2006 2010 2016 2018 2000 2006 2010 2016 2018 National Male Female Rural Urban Q1 Q5 Source: Estimations based on MICS and ENMC. Figure B. 11: Employment status and sector of employment by level of education and region Agriculture Industry Services 77% 77% 87% 86% 84% 80% 85% 84% 82% 74% 75% 76% 7% 6% Region6 Region3 Region1 Region2 Region3 Region4 Region5 Region7 Region1 Region2 Region4 Region5 Region6 Region7 Youth Adult 57 Farming Non-wage HH enterprise Wage Region5 Region1 Region2 Region3 Region4 Region5 Region6 Region7 Region1 Region2 Region3 Region4 Region6 Region7 Youth Adult Source: Estimations based ENMC 2018. Figure B. 12: Out-of-school rate by region 10 3 12 15 27 29 23 32 37 51 49 43 39 47 65 51 39 37 31 31 30 Region1 Region2 Region3 Region4 Region5 Region6 Region7 In school Dropout Never attended Source: Estimations based on ENMC 2018. Figure B. 13: Reason for never been in schools and dropout by region 58 Cost Family refusal School too far Abandonment/failure Too young Insécuité Illness Quality Work Other 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Region1 Region2 Region3 Region4 Region5 Region6 Region7 Total Cost Abandonment/failure Insecurity Pregnancy Married Quality Illness School too far away School closed Family refusal Absence of teacher Work Studies completed Other 100% 80% 60% 40% 20% 0% Region1 Region2 Region3 Region4 Region5 Region6 Region7 Total Source: Estimations based on ENMC 2018. 59 Annex C. Youth Survey Appocuch and key finding 1. In order to perform an efficient analysis of the magnitude and nature of out-of-school children and youth by socio-economic characteristics, existing quantitative data are not sufficient to do so. One of the primary goals of this Youth Survey is to produce high-quality data to supplement the quantitative data with qualitative information to gain insight on the particularities of out-of-school youth in CAR. Two instruments were used in the collection of data on youth: one for the quantitative data using standard model questionnaires and another for the collection of qualitative data using in-depth individual interviews on a restricted sample of youth. The qualitative instrument was designed to promote an appropriate depth of response on out-of-school youth. 2. The Youth Survey was conducted in February 2019 in Bimbo—the capital of the prefecture of Ombella- M'poko. The data were collected from 186 young people (137 men and 49 women) aged 15 to 24. It collected information on basic demographic and socioeconomic characteristics such as age, gender, marital status, education and training, labor market participation, enrollment status in army groups during the crisis etc. The survey was administered using the paper questionnaires filled by the respondents themselves. Thereafter, the information on the paper questionnaires has been converted to an electronic format. 3. An analysis of key demographic characteristics shows that with an average age of 21, about 74 percent of the sample are male compared with only 26 percent of female. Most are unmarried, but about 14 percent of the sample are married. The data also show that about 48 percent of the sample have at least one child and more than 50 percent of the sample lose one or both parents. In terms of youth participation in armed force (child soldiers), about 5 percent of youth sample reported as been served in any armed force during the conflict. 4. In terms of literacy, around 72 percent of the sample reported having some reading and writing skills. About 4 percent of surveyed youth aged 15-24 have no formal education, 23 percent have not passed primary education and about 14 percent have some post-secondary education. Young women have lower level of educational attainment compared to youth male. The distribution of the youth population in terms of age group shows that about 7 percent of the younger youth cohort (15-19) have no formal education compared to 3 percent for the youth age population aged 20 to 24. A large disparity between young females and male. For instance, about 17 percent of young male have some post-secondary education compared to 6 percent for young women, a 11-percentage point difference. 5. Among the out-of-school youth who did have some schooling, about 15 percent of youth tend to drop out early before completing primary education. Overall, even though they are in youth age group, most of them dropped out at the primary level, on average, about 32 percent of youth aged 15-19 who had dropped out, did so before completing primary school while this incidence is lower for out-of-school youth aged 20-24 with 12 percent. 6. The youth population is characterized by high rate of inactivity. Overall, about 61 percent of the surveyed youth aged 15-24 are inactive – i.e. not working, nor enrolled in school. Inactivity increases with age—older age 60 cohort of youth (age 20-24) are associated with higher rate (67 percent) of in activity than the younger age cohort (40 percent). This is mainly due to younger age group still in school. 7. Only 37% of surveyed youth aged 15-24 received some form of skills training. It also shows that the proportion of older age cohort (20-24) who had received skills training is higher than the proportion of younger age cohort, respectively 49 percent and 20 percent- a 29-percentage point difference. Young males tend more to receive skills training compared to young females. Indeed, about 44 percent of young males have received some form of skills training compared to young female (33 percent). 8. Among the surveyed youth, about 48 percent of youth engaged in economic activities. Employment rate is higher among young males compared to young females. About 51 percent of young males aged 15-24 are engaged in economic activities compared to 40 percent for females. It shows that, overall, about 54 percent of youth engaged in economic activities are involved in informal self-employment jobs, 14 percent in public sector and 31 percent in private sector. This figure also shows that young females are more likely to engage in informal sector compared to males. Indeed, about 84 percent of young females engaged in economic activities are self- employed compared to 45 percent for males. 9. Overall, about 24 percent of youth aged 15-24 engaged in economic activities are involved in commerce related jobs, while only 15 percent of youth are engaged in agriculture, and 12 percent in transport jobs. The majority of young female aged 15-24 engaged in economic activities are involved in commerce related jobs, about half of them. Young male engaged in economic activities are employed in commerce (25 percent), followed by transport (21 percent) and agriculture (17 percent.) Individual Interviews 10. There were five in-depth individual interviews to have a better understanding of the profile of out-of- school youth population in CAR. The individual interviews questions were translated in Sango—the primary language spoken in the country—to accommodate the target population. The main questions explored during the interviews are summarized in Box 1. Box 1: Individual interviews questions • What is your age? • Have you ever attended a formal education program (formal education)? • What were the reasons you never attended or dropped out? • Have you ever tried to re-engage into the education system? • Have you ever participated in a training/apprenticeship program (Formal or informal)? • What is the highest level of formal education that you have completed? • Did you work during the last 7 days (as employee or on your own account)? • Could you describe your employment status during the last week? • What type of employment was this work? 61 • What is your marital status? • Do you have any children? • Are your parents alive? • How many siblings do you have? • Have been affected by the conflicts? • Have you ever been enrolled in armed force? • What circumstances led you to serve in an armed force? • What circumstances led you to leave the armed force? • Do you have optimistic view for the future? 11. Among the youth surveyed, 60 percent are men (3 over 5) and only one over the 5 surveyed was married with 2 children, however 40 of them are single with 2 children. The young people interviewed were on average aged about 20 years. About 40 percent of the sample surveyed dropped out of school at primary level of education and the 40 percent dropped out at lower secondary level. Only one of the surveyed youths dropped out after completed upper secondary education. The majority of the youth surveyed cited poor economic condition as the main reasons for being out-of-school. All the surveyed youths are involved in informal self-employment jobs. Among the youths surveyed, one had been enrolled in an army group during the conflict. The profiles of the five youths interviewed are summarized in Box 2. Box 2: Profiles of interviewed youths • Individual 1: 21 years old unmarried male with 2 children. This young male is the third child of a family of nine children. He has completed upper secondary education but due to poor economic condition dropped out of school. He has received a 6-month professional training on poultry from Jeunesse Pionnière (JPN) training center and attended a workshop on GBV offered by MUNISCA. He is currently working has Moto-Taxi driver. He has not been enrolled to any armed groups during the conflict. • Individual 2: 23 years old unmarried male with 2 children. This young male is from a family of 11 siblings. He dropped out of school without completed lower secondary level due a parental death. He learned on the job the masonry since 2013. He obtained a certification in masonry from construction company, but he is still looking for a decent job in construction. • Individual 3: 16 years old unmarried female. This 16-year-old girl dropped out of school at grade 5. She lost her both parents when she was 3 years old and has 6 siblings. They all live with other family members. She is working as cake-seller. She is part of collective savings group ‘Tontine’ and plans to use that savings to attend sewing training. • Individual 4: 22 years old unmarried male, has been enrolled in an army group. This 22-year-old male has seen his parents killed during the conflict in 2013. He dropped out of high school in 2016 at grade 11. He enrolled in an army group after his parents’ death for about a week and finally dropped of the army group. He is currently working as Taxi-Moto driver and earn on average 5,000 CFA per day. He attended a professional training on mechanic offered by ACFPE and he hopes to work as mechanic in the future. • Individual 5: 20 years old married female with 2 children. 62 This 20-year-old young woman, the youngest of a family of 9 siblings, dropped out of school at grade 6 when she lost her father and got married very young. Her husband has no formal job and has seasonal jobs. She currently sells wood as economic activity and has never attended formal training but would like to become a mechanic. 63 Annex D. Methodology notes of the analysis Logit Model: The determinants out-of-school 7. A set of logit regressions is applied in order to study the probability of being out of school, the probability of dropping out of school, the probability of succeeding studies and the probability of repeating classes. The dependent variables are dummy variables. An ordered probit regression is also applied to study the determinants of delayed entry in primary education. The dependent variable is the age at which children (6 to 11 years old) have started primary school21. Sequential Model-The determinants of transition through the school system 8. In order to investigate the determinants of transition through the education system, a sequential logit model is estimated22. At each level of education, the probability of completion depends on the fact that individuals have completed the previous education level or not. For instance, completing primary education matters for individuals who are engaged in the education system and only people who have already completed primary education are concerned by secondary education completion. Sequential logit model allows modeling the probability of completing each level of education and that of moving to the next level of education taking into account the completion of the previous level of education. The purpose is to model the influence of the explanatory variables on the probability of passing a set of transitions. 9. The model that is estimated for the CAR includes five transitions: first, decision whether to enroll in primary school (vs. never enrolling); second, decision whether to continue/finish primary school (vs. dropping out of primary school); third, given that the youth continued/finished primary school, whether to get into lower secondary education or not; fourth, given that the youth enrolled in lower secondary education, whether the youth eventually dropped out or not; fifth, given that the youth enrolled in upper secondary education, whether the youth eventually dropped, continued with upper secondary education or youth completed upper secondary education by the age of 24. We focus on youth (age 15 to 24) because they seem to be more concerned by transition issues in the education system. 10. A schematic of the model is shown in the figure below. In this chart, one is required to have passed all lower transitions in order to make a decision to continue or to leave the school system. Given the assumption that decisions are independent, one can estimate the model by running a series of logit regressions for each transition on the appropriate sub-sample. Illustration of the transition through the CAR’s education system 21See Maddala (1983) and Greene (2005) for technical details about all these econometric models. 22This regression model is also known under a variety of names, such as Sequential Response Model (Maddala 1983), Continuation Ratio logit (Agresti 2002), Model for Nested Dichotomies (Fox 1997) and Mare model (Mare 1981; Shavit and Blossfeld 1993). For an extended discussion see also Buis (2010). 64 Upper P5 secondary dropout P4 P3 Still in or Still in Lower completed P2 Scondary Enrolled in upper secondary P1 1-P5 lower secondary Still in Complete upper 1-P4 Primary Exit secondary by Enrolled in 1-P3 Primary Exit age 24 Never attended 1-P2 Exit 1-P1 Exit After assigning a value to each level of education (pseudo-years) one can study the effect of the explanatory variables on the expected final outcome. The probability that person i passes transition k, pk, is given by: (𝐚�+�𝐛�∗𝐱𝐢) 𝐞𝐱𝐩� p1i = �+𝐞𝐱𝐩�(𝐚�+�𝐛�∗𝐱𝐢) (𝐚�+𝐛�∗𝐱𝐢) 𝐞𝐱𝐩� p2i = �+𝐞𝐱𝐩�(𝐚�+𝐛�∗𝐱𝐢) if pass1 = 1 (𝐚𝟑+𝐛𝟑∗𝐱𝐢) 𝐞𝐱𝐩� p3i = �+𝐞𝐱𝐩�(𝐚𝟑+𝐛𝟑∗𝐱𝐢) if pass2 = 1 (𝐚𝟒+𝐛𝟒∗𝐱𝐢) 𝐞𝐱𝐩� p4i = if pass3 = 1 �+𝐞𝐱𝐩�(𝐚𝟒+𝐛𝟒∗𝐱𝐢) (𝐚𝟓+�𝐛𝟓∗𝐱𝐢) 𝐞𝐱𝐩� p5i = if pass4 = 1 �+𝐞𝐱𝐩�(𝐚𝟓+�𝐛𝟓∗𝐱𝐢) where, the constant for transition k is ak and the effect of the explanatory variable xk is bk. Buis (2010) shows that the effect of the explanatory variables on the highest achieved level of education is a weighted sum of the effects of passing each transition and that the contribution of each transition can be visualized by the area of a rectangle with width equal to the weight and height equal to the effect on the probability of passing the transition (the log-odds ratio or the marginal effect).23 One can thus see how the effect differs by characteristic (such as gender) or cohort. Oaxaca Decomposition Model-Factors that explains differences in the group outcomes 12. The standard conceptual approach for education services is model as the notion of production function (Heynema, 1979; Krueger, 2003 and Orazem and King, 2008) and others also treated as utility function (Glewwe, 23 For this one can use the Stata command Seqlogitdecomp. 65 and Kremer, 2005) to establish causal relationship between educational outcomes and determinants of the outcomes. In the former case, schools are treated as producers and best viewed as organizations that should try to maximize output, subject to their budget constraints. In the latter case, conceptual framework is farmed as household maximizing utility that can be consumed at different points in time (life cycle), and each child’s years of schooling and learning. Both the production and utility maximization approaches could fit to our data and proposed estimation methodologies, one can refer to the cited sources for the details. In this proposal, we aim to develop on the modified production function for education defined as: 𝒀 = 𝒇(𝑿) + 𝜺 Where Y is outcome measures (outcome indicators), X is a vector of variables and ε is unobserved factors or residuals which includes student innate ability, motivations, and other quality improvement efforts. 13. Similar to the literature, this proposal aims to employ the extended Oaxaca -Blinder (1973) decomposition models and estimate the learning achievement (test scores) change over-time both at mean and distributions. Using the above production function, student performance can be modeled as a function of various determinants of educational performance, including both individual/family background characteristics and school characteristics. The school achievement difference changes between the following three categories: (i), gender, (ii) areas of residence, and (iii), poorest and richest wealth quintile. Based on the first category (male and female) this can be represented as: ∆𝒎𝒂�𝒆, 𝒇𝒆𝒎𝒂�𝒆 = 𝑬(𝒀𝒇𝒆𝒎𝒂�𝒆 /𝑿𝒎𝒂�𝒆 ) − 𝑬(𝒀𝒎𝒂�𝒆 /𝑿𝒎𝒂�𝒆 ) 𝑲 𝑲 ̂ 𝒎𝒂�𝒆,𝒌 (𝑿 = ∑ 𝜷 ̅ 𝒇𝒆𝒎𝒂�𝒆,𝒌 ) + ∑ 𝑿 ̅ 𝒎𝒂�𝒆,𝒌 − 𝑿 ̅ 𝒎𝒂�𝒆,𝒌 − 𝑿 ̅ 𝒇𝒆𝒎𝒂�𝒆,𝒌 (𝑿 ̅ 𝒇𝒆𝒎𝒂�𝒆,𝒌 ) 𝒌=� 𝒌=� 𝑲 ̅ 𝒇𝒆𝒎𝒂�𝒆,𝒌 − 𝑿 + ∑(𝑿 ̂ 𝒎𝒂�𝒆,𝒌 ) ̂ 𝒇𝒆𝒎𝒂�𝒆,𝒌 − 𝜷 ̅ 𝒎𝒂�𝒆,𝒌 )(𝜷 𝒌=� 14. Where k is the regressor (k=1 is the intercept) and gap will be there if there is no difference between the two periods. The gap is decomposed in three effects: (i), the endowment (or characteristic) effect (first term) is the difference in scores due to differences in the average for each regressor, weighted by the group male slope. It represents the part of the score gap that can be explained just because of different average characteristics between both groups i.e. it represents the explained component of the performance gap. This term indicates how differences in the average endowments of individual and/or school resource/quality characteristics between the two groups affect the average performance gap, (ii), the returns effect or coefficient (second term) represents the proportion of the score gap that can be explained by differences in the slopes between both groups (given the average group male characteristics). Note that the last two terms on the right of equation collectively represent what Oaxaca (1973) originally in the labor market context referred to as the “discrimination� or “residual� component of the wage gap. However, Blinder (1973) went further in identifying the two separate parts of the discrimination component as one part due to differences in the intercepts and one part due to differences in the coefficients, and the former has since been referred to the “pure discrimination� component and the latter as the 66 unexplained component, and (iii), the interaction effect (third term) is the residual part of the decomposition. In the context of this study, the component is noted as representing the difference in “pure� efficiency; that is, differences in performance that are unrelated to the covariates included in the model such as differences in unobserved school quality or efficiency and individual characteristics (Krieg and Storer, 2006). 15. Studies analyzing differences between two groups have further been interested in the investigation of the individual and collective contributions of characteristics to the explained and unexplained components (Ammermuller, 2007). For example, we might be interested in evaluating how much of the difference in test scores between students in the countries/areas/gender is due to differences in individual and family background characteristics, and how much is due to differences in school characteristics and expressed as: ̂ 𝒎𝒂�𝒆 ̂ 𝒇𝒆𝒎𝒂�𝒆 − 𝒀 𝒀 ̂ 𝒇𝒆𝒎𝒂�𝒆 − 𝜶 = 𝜶 ̂ 𝒎𝒂�𝒆 + 𝜸 ̂ 𝒎𝒂�𝒆 − � ̂ 𝒇𝒆𝒎𝒂�𝒆 (� ̂𝒇𝒆𝒎𝒂�𝒆 (𝑹 ̂ 𝒎𝒂�𝒆 ) + 𝜹 ̂ 𝒎𝒂�𝒆 ) ̂ 𝒇𝒆𝒎𝒂�𝒆 − 𝑹 ̂ ′𝒎𝒂�𝒆 (𝜸 + � ̂𝒇𝒆𝒎𝒂�𝒆 − 𝜸 ̂𝒎𝒂�𝒆 ) + 𝑹 ̂ ′𝒎𝒂�𝒆 (𝜶 ̂𝒎𝒂�𝒆 ) ̂ 𝒇𝒆𝒎𝒂�𝒆 − 𝜹 16. It is fairly simple to identify the contributions of individual characteristics to the explained component given that the total component is merely a sum over the individual contributions (Jann, 2008: 8). For example: ̂ �gap� = β 𝐸𝑥�𝑙𝑎𝑖ned ̂ female (X ̂ female − X̂ male ) ̂ 1,male (X =β ̂1,female − X̂1,male ) + β ̂ 2,female (X ̂ 2,male ) + ⋯ ̂ 2,female − X ̂ n,female (X +β ̂ n,female − X̂ n,male ) where X ̂1,female , X ̂ 2,male are the means of the individual and school characteristics, β ̂ 2,male are the ̂ 1,female , β associated partial regression coefficients? Furthermore, standard errors for the individual contributions are straight forward to estimate (Jann, 2008). However, estimating the contributions of the individual characteristics to the total unexplained component is less straightforward. This is due to the fact that the results offered by the detailed Oaxaca-Blinder decomposition of test score differentials are not invariant to the choice of reference (omitted) group when using dummy variables in the education production functions (Jones, 1983; Oaxaca & Ransom, 1994; Nielsen, 2000; Horrace & Oaxaca, 2001; and Yun, 2005). The choice of reference group does not, however, affect either the total contribution or individual contributions of each categorical variable to the explained component. Conversely, a change of reference category is found to not only change the individual contributions of single categorical variables to the unexplained component, but also to alter the contribution of the category as a whole. Thus, following Jones (1983) detailed decomposition of the unexplained component is given as: ̂ Unexplained ̂ female − α �gap = [(�α ̂ 1,male )] ̂ 1,female − β ̂ B2003 (β ̂male ) + X 17. The first term on the right-hand side of equation is the part of the unexplained gap that is due to “pure� performance differentials; the second term represents that part of the unexplained gap that is due to differences in the education production process (that is, differences in the way that educational input X is transformed into 67 educational outputs or returns to characteristics). Assuming further that the zero point of a continuous variable, Xi, is shifted by adding a constant a, the decomposition is now given by: ̂ Unexplained�gap = [α ̂ 1,female �) − (α ̂female − aβ′ ̂′ ̂male − aβ ̂′ ̂ ̂ 1,male ) + (X male − a)(β1,female − β1,male )] 18. ̂ 1B ) from the “pure� performance effect to the other ̂ 1A − β The scale shift in Xi results in a transfer of a(β part of the unexplained gap that is due to differences in coefficients. Therefore, the detailed decomposition results for the unexplained component changes when no natural zero point exists for one or more of the predictor variables and one can refer to Yun (2005) for the proposed solution to the identification problem, however, more recent studies (Fortin, Lemieux, and Firpo 2010, Machado and Mata (2005), and Altonji, Bharadwaj and Lamge , 2008) considered this solution as less establish and still evolving and caution should be taken when interpreting the detailed decomposition results. Overall, while breakdown of unexplained part into coefficient and instructions seems interesting, for simplicity and due to the narrow coverage of the analysis we used unexplained without further decomposition. 68 Annex E. Regression Results Table 1: Determinates of out-of-school Out of school (in school==1) Mlogit of school status (base in school) Log Marginal Never Dropout effect Gender Female 2.582 0.172*** 1.547 0.759 (7.59)*** (0.0221) (8.49)*** (5.89)*** Age age in years 1.369 0.0550*** 0.360 0.309 squared (12.39)*** (0.00361) (10.19)*** (12.10)*** Area Urban 0.474 -0.135*** -0.813 -0.703 (5.10)*** (0.0268) (4.06)*** (4.55)*** Wealth quintile Poor 0.744 -0.0510 -0.561 -0.128 (Ref. Poorest (1.32) (0.0385) (2.02)** (0.55) Middle 0.524 -0.114*** -1.087 -0.449 (3.18)*** (0.0355) (4.12)*** (2.08)** Rich 0.621 -0.0832** -0.740 -0.351 (2.34)** (0.0353) (2.83)*** (1.63) Richest 0.384 -0.172*** -1.719 -0.603 (4.08)*** (0.0423) (4.91)*** (2.45)** Family Size Family size 0.921 -0.0144*** -0.078 -0.077 (4.17)*** (0.00339) (2.90)*** (3.76)*** Head Education Some primary 1.183 0.0295 0.242 0.083 (Ref. No (1.08) (0.0274) (1.09) (0.51) education) Some lower 0.909 -0.0169 -0.041 -0.216 secondary (0.39) (0.0435) (0.12) (0.84) Some upper 1.220 0.0348 -0.289 0.314 secondary (0.82) (0.0425) (0.80) (1.30) plus Head sex Female 1.128 0.0210 -0.032 0.091 (0.59) (0.0352) (0.11) (0.44) Head age age 1.003 0.000477 0.008 -0.001 (0.52) (0.000914) (1.11) (0.19) Conflict Death in HH 1.277 0.0428 0.332 0.193 indicators (Yes) (1.39) (0.0308) (1.42) (1.06) Affected by 0.992 -0.00147 0.078 -0.091 conflict (Yes) (0.05) (0.0274) (0.35) (0.56) Displaced 1.509 0.0719* 0.189 0.522 (Yes) (1.91)* (0.0373) (0.63) (2.42)** Public teachers Tercile 0.694 -0.0639** -0.873 -0.190 (Ref. Tercile middle 69 bottom) (2.35)** (0.0270) (4.23)*** (1.16) Tercile top 0.810 -0.0365 -0.666 -0.042 (1.14) (0.0324) (2.53)** (0.22) Classroom Tercile 2.111 0.137*** 1.555 0.510 availability (Ref. middle Tercile bottom) (4.56)*** (0.0303) (5.64)*** (3.02)*** Tercile top 2.279 0.151*** 1.352 0.688 (4.55)*** (0.0329) (4.53)*** (3.72)*** Constant -7.730 -5.183 . (9.12)*** (9.10)*** N 1,551 1,551 1,576 1,576 Source: Estimations based on 2018 ENMC. Table 2: School decision tree _1_2_3_4_5v0 _2_3_4_5v1 _3_4_5v2 _4_5v3 _5v4 Gender Female -1.103 -0.868 0.085 -0.709 0.311 (7.04)*** (6.47)*** (0.46) (3.15)*** (1.09) Age age in -0.155 -0.160 0.440 -0.261 0.536 years (5.68)*** (6.36)*** (10.27)*** (5.94)*** (8.15)*** squared Area Urban 0.368 0.880 0.880 0.927 0.500 (2.08)** (5.66)*** (4.31)*** (3.24)*** (1.15) Wealth Poor 0.425 -0.046 -0.112 0.856 0.970 quintile (1.88)* (0.20) (0.33) (1.64) (1.17) (Ref. Middle 0.770 0.643 -0.032 0.075 0.029 Poorest (3.43)*** (2.98)*** (0.10) (0.17) (0.04) Rich 0.413 0.395 0.711 0.762 1.160 (1.92)* (1.84)* (2.26)** (1.82)* (1.55) Richest 1.280 1.155 1.007 0.823 1.627 (4.14)*** (4.42)*** (2.86)*** (1.87)* (2.14)** Family Family 0.033 0.068 0.017 0.046 0.037 Size size (1.39) (3.31)*** (0.62) (1.40) (0.78) Head Some -0.131 0.175 0.249 -0.714 -0.643 Educatio primary n (Ref. (0.67) (1.05) (1.11) (2.31)** (1.78)* No Some -0.143 0.431 0.838 -0.014 -0.471 educatio lower (0.46) (1.54) (2.27)** (0.03) (0.91) n) seconda ry Some 0.403 0.112 0.741 -0.619 0.320 upper (1.29) (0.46) (2.20)** (1.55) (0.67) seconda ry plus Head sex Female 0.029 -0.386 -0.074 0.018 0.496 (0.11) (1.77)* (0.24) (0.05) (1.15) 70 Head age -0.010 -0.002 0.004 0.004 0.014 age (1.50) (0.30) (0.59) (0.40) (1.30) Conflict Death in -0.104 0.126 -0.289 -0.893 0.034 indicator HH (Yes) (0.53) (0.65) (1.12) (2.93)*** (0.07) s Affected -0.174 0.008 0.612 0.402 0.677 by (0.95) (0.04) (2.56)** (1.44) (1.99)** conflict (Yes) Displace 0.077 -0.170 -0.400 -0.769 -0.890 d (Yes) (0.33) (0.77) (1.32) (2.15)** (1.47) Public Tercile 0.730 0.301 0.212 -0.264 -0.170 teachers middle (Ref. (4.14)*** (1.78)* (0.93) (0.82) (0.37) Tercile Tercile 0.699 0.224 0.233 -0.764 0.575 bottom) top (3.06)*** (1.15) (0.93) (2.08)** (1.17) Classroo Tercile -1.264 -0.287 -0.315 -0.758 0.390 m middle availabili (5.44)*** (1.61) (1.34) (2.46)** (0.90) ty (Ref. Tercile -0.945 -0.457 -0.238 -1.390 0.234 Tercile top bottom) (3.86)*** (2.44)** (0.94) (3.90)*** (0.45) Constan 5.114 2.717 -8.828 6.057 -13.428 t (7.45)*** (4.63)*** (9.42)*** (5.40)*** (7.42)*** N 1,539 1,539 1,539 1,539 1,539 Source: Estimations based on 2018 ENMC. Table 3: Decomposition of determinates of out-of-school Decomp Detailed Decom Gender Area Non-poor Differential Prediction_1 0.478 0.694 0.548 (26.16)*** (42.44)*** (38.23)*** Prediction_2 0.703 0.465 0.768 (42.52)*** (24.86)*** (31.42)*** Difference -0.224 0.229 -0.221 (9.10)*** (9.20)*** (7.78)*** Endowments Region -0.004 -0.023 0.025 (0.66) (0.63) (0.78) Student -0.018 0.004 -0.023 71 (2.58)*** (0.35) (2.28)** Parent -0.018 0.078 -0.031 (2.81)*** (3.44)*** (2.90)*** Conflict -0.002 0.002 0.008 (0.84) (0.43) (1.10) School -0.002 0.031 0.010 (0.65) (1.23) (0.46) Services 0.000 0.017 0.017 (0.13) (1.61) (1.19) Coefficients Region 0.056 -0.096 0.042 (0.75) (1.21) (0.54) Student 0.303 -0.150 0.309 (2.12)** (1.07) (1.91)* Parent 0.031 0.038 -0.079 (0.27) (0.30) (0.62) Conflict -0.012 0.032 -0.029 (0.67) (1.58) (1.33) School 0.161 -0.056 0.084 (1.94)* (0.71) (0.77) Services -0.145 0.192 0.116 (1.92)* (2.79)*** (1.15) Constant -0.581 0.230 -0.558 (2.41)** (0.94) (2.04)** Interaction Region 0.001 0.038 -0.072 (0.09) (0.97) (1.86)* Student -0.006 -0.001 -0.007 (1.67)* (0.60) (1.44) Parent 0.007 -0.080 0.049 (0.98) (1.93)* (1.71)* Conflict 0.004 -0.006 -0.008 (1.14) (0.89) (1.01) School -0.005 -0.021 -0.037 (0.80) (0.67) (1.36) Services 0.006 -0.003 -0.034 (1.16) (0.18) (2.07)** N 1,680 1,680 1,680 Source: Estimations based on 2018 ENMC. Table 4: Profile of the subjects/modules provided by the main TVET institutions in CAR TVET institution name Lycée École Centre de Institut Lycée DON Providen Subject offered Techniq des Formation Modern Profession BOSCO ce ue Métie et e des 72 nel rs d’Insertion Métiers DAMA Féminin d’Arts par Spécialis LA (EMA) Apprentiss és age (CFIA) (IMMS) Technical Baccalaureate X Carpentry X X X X Masonry X X X X X Metal Construction X X General Mechanics X Auto Mechanics X X X X X Electricity (Auto and/or X X X X X building) Construction X Administrative and X Commercial Services Accounting X X Administration and X Management Family and Social X Economy Hotel Industry X X Hotel and Catering X Cleaners X Jewelry X Drawing X X Painting (decorative art) X X Leather goods X Art ironwork X Binding X Gilding X Stationery X Wood carving X Furniture X Joinery X X Plumbing - Piping X X X Tiling X X Cutting and sewing X X Mechanic drivers X Agriculture and Livestock X Computer (office and maintenance) X X Refrigeration and air conditioning X 73 Source: Table 5: CAR Education System Age Universite de Bangui et 25 7 7 ème année Doctorat Ecole Normale des 24 6 6 ème année Ecole Normale Instituteurs Superieure Superieur 23 5 5 ème année 5 ème année Master 2 Autres 22 4 4 ème année 4 ème année Master 1 21 3 3ème année 3ème année Licence 20 2 2ème année 2ème année 2ème année 19 1 1ère année 1ère année 1ère année Baccalaureat Secondaire Technique Secondair 18 3 Tle Tle 3 3ème année General Lycee e 17 2 1 ère 1 ère 2 2ème année 16 1 2 nd 2 nd 1 1ère année Brevet d'Enseignement du Premier Cycle Fondamental 2 15 3 3 ème 3 3ème année College College 14 2 4 ème 2 2ème année 13 5 ème 1 1ère année 12 1 6 ème Formation Professionnelle Certificat d'etudes primaires elementaires 11 6 CM2 Fondamental 1 Ecole Primaire 10 5 CM1 Primaire 9 4 CE2 8 3 CE1 Formation non formelle 7 2 CP 6 1 CI Enseignement Superieur Formation professionnelle 5 3 Grande Section 4 2 Prescolaire Moyenne Section Enseignement Seconmdaire 3 1 Petite Section Enseignement fondamental Prescolaire Source: 74