Central African Republic 2016 National Commune Monography Survey February 2017 Foreword At the international Donor Conference held in Brussels on November 17, 2016, the Government presented its strategy for Recovery and Peacebuilding as reflected in the Recovery and Peacebuilding Assessment (RPBA). The RPBA: (i) identified recovery and peace-building needs and priorities and their associated financial costs; (ii) identified specific operational, institutional and financial arrangements to facilitate the implementation of identified priorities; and (iii) presented an approach to monitor implementation progress. The RPBA was widely supported and donors expressed their confidence in the Government by pledging over $ 2 billion in resources for projects to be implemented over the next 3 years. With resources available, it is now time to implement. Doing so efficiently and effectively requires timely and reliable information. The results of the first National Commune Monography Survey offers just this. Not only are its results timely, they are collected using an innovative survey design that allowed to collect information across the territory, for every commune and for households living in every commune, rapidly and at the fraction of the cost of a typical household survey. As such, by repeating the survey regularly, this survey not only presents the first complete photo of the state of the nation, it becomes the basis of a monitoring framework going forward. The Local Development Index derived from the information collected by the survey and presented in this report serves as an important tool for decision about the allocation of resources. It allows to assess localities that are most deprived, and, by deriving the indicator regularly, to assess progress over time. This work is not only very relevant for decision makers in Government and other agencies, it represents the frontier in statistical development. Up to date, representative information and the associated monitoring tools will improve the way we work to better address the needs of the citizens of CAR. M. Félix MOLOUA Ministre de l’Economie du Plan et de la Coopération Internationale 1 Acknowledgements This report is the result of joint efforts between the ICASEES and the World Bank. Data collection efforts were led by Alfred SABENDO, Director of Economic, Demographic, and Social Statistics at ICASEES, under the supervision of Magloire Constantine DOPAMAS, Director General of ICASEES. Arsène ZIA KOYANGBO of ICASEES was responsible for carrying out the data entry. Alain Maxime NDOMO and Gervais DANBOY of ICASEES supported training and coordination of enumerators. Gervais Chamberlain YAMA, Roy KATAYAMA, and Johannes HOOGEVEEN of the World Bank provided technical assistance in the conceptualization and implementation of the survey. The report was prepared in collaboration with ICASEES by Felipe DIZON of the World Bank. He conducted the analysis of the household survey, and Bienvenue TIEN conducted the analysis of the census of the communes. The team would like to thank the interviewers as well as the participants in the meetings reviewing questionnaires and the draft report, in particular Silvio Paul GIROUD and the Directors of ICASEES for their comments. The team would like to thank the United Nations - World Bank Fragility & Conflict Partnership Trust Fund. All photo credits in the report are attributed to Stephan GLADIEU. 2 Executive summary To collect information about the state of the nation and development priorities, l’Institut Centrafricain des Statistiques, des Etudes Economiques et Sociales (ICASEES) conducted in August 2016 a national survey, the Enquête Nationale sur les Monographies Communales. The survey consisted of two components: a commune census which elicited information from all commune authorities in the country, and a household survey which elicited information from households in almost all communes across the country. This survey is the first source of objective information on the state of the nation since the crisis. As such, it covers some important data gaps. The commune census provides information on conditions in all communes across the territory, including on local infrastructure, health and education facilities, local governance, economic activities, conflict, security and violence, and on local perspectives of security and policy priorities. The household survey complements the commune census. It provides information on socio-economic well-being, perceptions of security and economic conditions, and opinions on policy priorities. The household survey was conducted in all but two communes, with 10 households interviewed in each. By conducting the commune and household surveys simultaneously, it is possible to explore differences in perspectives between local authorities and citizens. State of the Nation These survey results allow us to provide a snapshot of the state of the nation and to quantify major needs, the extent of regional differences, and changes relative to 2012 (before the crisis). The report presents a Local Development Index (LDI) that combines a range of policy relevant indicators into a single measure. The LDI comprises variables across three different pillars that are important for development: local administration, local infrastructure, and access to basic services. By shedding light on conditions in a straightforward way, the LDI allows decision- makers to quickly monitor progress across communes. This is the first time an LDI has been prepared for CAR and in future rounds of the survey the definition of the composite index can be refined as needed. Based on the current LDI, the commune census reveals that commune administration offices are under-staffed and lack funding, that access to basic infrastructure such as electricity, mobile phone coverage, banking services and road networks is low, and that access to basic social services, such as public primary schools, health centers, and clean water is limited, with large disparities between Bangui and other regions of the country. Results from the household survey echo the challenging context that the commune census presents. Education levels of individuals have been stagnant, and are lower in rural areas and for women. Households generally have very few assets, and significantly less now than before the crisis in 2012. Six out of every ten households are categorized as food insecure. Households on average eat less than two meals per day and tend to consume low quality food, staples mostly. Households are vulnerable to a variety of shocks and have inadequate coping strategies. The household survey allows for explorations into crucial sub-populations such as internally displaced persons and those working in mining. It is estimated that roughly 15 percent of individuals are internally displaced. Half of the households had at least one member which was ever displaced since the onset of the crisis in 2012. Households which have currently received 3 displaced individuals tend to be worse off in terms of asset ownership and food consumption, and displaced children tend to miss more days of school. On another note, rough calculations from this survey indicate that the artisanal mining sector is sizeable. A back of the envelope calculation suggests that (during the harvest season in August) 100,000 individuals are involved in mining. It appears that households in mining communes have better food consumption, suggesting that mining has positive spillovers to households in the community. Development Priorities There is broad consensus among local authorities and citizens alike that the highest priority for the country is achieving peace, reconciliation, and security. About 9 out of 10 communes, and 8 out of 10 households rank this as most important. The next priority is good governance and the provision of basic services, followed by economic development. With respect to specific policy interventions, the top five interventions identified by survey respondents are in order of importance: (i) improve security through a reformed FACA (army) operating across the territory; (ii) re-integrate ex-combatants; (iii) improve access to basic services (health, education, water); (iv) improve road infrastructure; and (v) strengthen agricultural and livestock production. Household perceptions of living conditions and opportunities indicate high expectations for economic recovery and welfare improvements. Most households felt that their standard of living was about the same or worse relative to 6 months prior to the survey. Looking ahead, the majority of households were optimistic that their standard of living and employment opportunities would improve in the coming year. Both urban and rural residents expressed this optimism. Across prefectures, residents of Vakaga and Haut-Mbomou were the most doubtful, while those in Mbomou, Ombella-M’Poko, and Nana-Mambere were the most upbeat about the future. The majority of households perceived that ethnic and religious tensions had improved. Building a Monitoring System The Monographies Communales survey, with its nationwide coverage and demonstrated ability to collect information rapidly and in a cost-effective manner, presents an objective feedback loop for the state and citizens. It will form a key component of a new monitoring system that assesses progress towards RPBA objectives. In future rounds of the survey, the commune census and household survey can be fine-tuned and tailored to provide answers to new policy questions that emerge. The survey can also be adapted to enable better inference about important subgroups such as those in displacement camps, those in Bangui or in mining settlements. A dashboard can be created to present information on tangible development outcomes (e.g. number of functional schools; reach of the mobile phone network; availability of transport) as well as to report on citizen perceptions (e.g. security; trust; whether life is improving). The dashboard can integrate complementary information from other sources (e.g. prices; displaced people; quality of roads) to offer a more complete picture of the state of the nation. When these indicators are measured regularly over the course of the RPBA, they can demonstrate trends over time. 4 To facilitate transparency and confidence building, the data from the commune census and household survey are publicly available on the Ministry of Economy, Planning, and Cooperation website and will be posted on the new ICASEES website when it is launched. They can also be downloaded from: http://bit.ly/2k7wFlq. 5 Table of Contents 1. Introduction ............................................................................................................................. 8 2. Commune census ................................................................................................................... 10 2.1. Introduction .................................................................................................................... 10 2.2. Local administration ....................................................................................................... 11 2.3. Infrastructure .................................................................................................................. 12 2.4. Access to basic social and economic services ................................................................ 14 2.5. Local Development Index (LDI): a composite measure of commune capacity ............. 17 3. Household Survey.................................................................................................................. 20 3.1. Profile of households ...................................................................................................... 20 Agro-ecological zones ........................................................................................................... 22 Mining ................................................................................................................................... 23 Internal displacement ............................................................................................................. 24 3.2. Education ........................................................................................................................ 26 3.3. Wealth and food consumption ....................................................................................... 28 3.4. Shocks and Coping Strategies ........................................................................................ 34 4. Perceptions and Priorities ...................................................................................................... 40 4.1. Perceptions of security and socio-economic welfare ..................................................... 40 4.2. Priorities for Development ............................................................................................. 43 5. Conclusions ........................................................................................................................... 45 5.1. Summary of Findings: Commune census ...................................................................... 45 5.2. Summary of Findings: Household Survey ..................................................................... 46 5.3. Building a Monitoring System ....................................................................................... 47 6. Annex..................................................................................................................................... 49 6.1. Survey Design ................................................................................................................ 49 6.2. Methodology .................................................................................................................. 50 Local Development Index ..................................................................................................... 50 Wealth Index.......................................................................................................................... 55 Food Consumption Score (FCS)............................................................................................ 55 Database Access .........................................................................Error! Bookmark not defined. Commune Questionnaire ............................................................Error! Bookmark not defined. Household Questionnaire ...........................................................Error! Bookmark not defined. 6 List of Tables Table 1.1 Prefectures in each region ............................................................................................... 9 Table 2.1: Water, energy, and telecommunications infrastructure by region ............................... 13 Table 2.2: Constraints in health and education (percent of communes) ....................................... 15 Table 2.3: Local Development Index- components and weights .................................................. 17 Table 3.1: Basic demographic characteristics ............................................................................... 21 Table 3.2: Reasons for not enrolling in school or missing school days ........................................ 27 Table 4.1: Priorities for development ........................................................................................... 44 Table 6.1 Indicators used to construct the sub-indices and overall composite index ................... 50 Table 6.2 Ranking of communes on the Local Development Index (LDI) .................................. 52 List of Figures Figure 2.1: Main and secondary economic activities.................................................................... 10 Figure 2.2: Local administration- funding and staffing in communes.......................................... 12 Figure 2.3: Local infrastructure- mobile phone coverage, banking services, and roads .............. 14 Figure 2.4: Access to basic services- primary schools and health centers.................................... 15 Figure 2.5 NGO Presence in the CAR .......................................................................................... 16 Figure 2.6: Local Development Index across communes ............................................................. 18 Figure 3.1: Economic activity of the household head................................................................... 22 Figure 3.2: Map of agro-ecological zones, collected from commune level data .......................... 23 Figure 3.3: Age pyramid of individuals in sample ....................................................................... 23 Figure 3.4: Map of communes with mining in the top three economic activities ........................ 24 Figure 3.5: Internal displacement experience ............................................................................... 25 Figure 3.6: Education levels of individuals in the sample ............................................................ 26 Figure 3.7: Ownership of household assets .................................................................................. 28 Figure 3.8: Food consumption ...................................................................................................... 30 Figure 3.9: Food consumption of those who produce food vs those who purchase food ............. 30 Figure 3.10: Food consumption by wealth and region ................................................................. 31 Figure 3.11: Food consumption and wealth in mining communes ............................................... 32 Figure 3.12: Food consumption, wealth and internal displacement ............................................. 33 Figure 3.13: Change in food prices ............................................................................................... 34 Figure 3.14: Experience of negative shocks ................................................................................. 35 Figure 3.15: Household coping strategies to deal with negative shock ........................................ 37 Figure 3.16: Wealth, negative shocks, and food consumption ..................................................... 38 Figure 3.17: Membership in community groups........................................................................... 39 Figure 4.1: Household perception of the security situation .......................................................... 40 Figure 4.2: Trust in different armed forces among local authorities and households................... 41 Figure 4.3: Perception on the economic situation among local authorities and households ........ 41 Figure 4.4: Perceptions on future improvements in the quality of life ......................................... 42 Figure 4.5: Priorities for development .......................................................................................... 43 7 1. Introduction Few doubt the enormous challenges the Central African Republic (CAR) faces. Few also doubt the need for reliable and up to date information to inform decision makers about the conditions in the country. In its absence decision makers are restricted in their capacity to efficiently allocate resources, to monitor progress of development programs, and to readapt programs as needs evolve. In a first effort to fill the existing data gap following the return to constitutional order in February 2016, the National Statistics Institute (ICASEES) conducted in August 2016, a national survey, with financial and technical support from the World Bank, to collect information about the state of the nation and the development priorities of local authorities and citizens. This survey, labelled Enquête Nationale sur les Monographies Communales, comprised of two elements: a commune census and a household survey. The commune census collected information on current conditions in all 179 communes of the CAR and on the perspectives on security and policy priorities of local officials. In spite of ongoing insecurity in parts of the country, this census was conducted successfully. Through interviews with local authorities, information was gathered on local infrastructure, health and education facilities, local governance, economic activities, conflict, security and violence. In addition, the census asked about perceptions of socio-economic and security conditions in the principal town of the commune, and about policy priorities for the commune with respect to three themes: peace, reconciliation, and security; good governance and the provision of basic services; and economic recovery. The household survey collected information on household socio-economic well-being, perceptions of security and economic conditions, and opinions on policy priorities, complementing the data from the commune census. The household survey was conducted in all but two communes, with 10 households interviewed in each. The two communes that could not be covered due to armed groups and impassable rainy season roads are Ouandja-Kotto in Haute- Kotto prefecture and Ouada-Djalle in Vakaga prefecture, both in in the northeastern area of the country. By conducting the commune and household surveys simultaneously, it is possible to explore potential differences in perspectives between local authorities and citizens, and to map both the provision of services at the commune level and socio-economic well-being at the household level. The Monographies Communales survey will form the basis of a feedback system to inform decision makers about progress. By conducting these surveys every six months, it will be possible to obtain timely and reliable evidence of evolving needs and development progress in CAR to inform decision making. With additional donor support, this feedback system can be strengthened to improve monitoring and would complement other initiatives to strengthen data production in CAR. 8 In this report, Section II characterizes recent conditions of the communes based on data elicited from local authorities in the commune census and presents a composite index of indicators on local administration, infrastructure, and basic services, called the Local Development Index (LDI). Section III presents results from the household survey that describe the situation of households and individuals with respect to topics such as education, wealth, consumption, and vulnerability. Section IV presents the perceptions and development priorities of both local authorities and households, and Section V concludes by summarizing key findings and by describing the survey’s potential role in a monitoring system. Additional details on the survey design and analytical methodology can be found in the Annex. As the report makes references to regions in the analysis, the corresponding prefectures located in each region are summarized below in Table 1.1. Table 1.1 Prefectures in each region Region Prefectures Region 1 Ombella-M’Poko, Lobaye Region 2 Nana-Mambere, Mambere-Kadei, Sangha-Mambere Region 3 Ouham, Ouham-Pende Region 4 Ouaka, Kemo, Nana-Gribizi Region 5 Bamingui-Bangoran, Vakaga, Haute-Kotto Region 6 Basse-Kotto, Mboumou, Haut-Mboumou Region 7 Bangui (capital) 9 2. Commune census 2.1. Introduction Smallholder agriculture is the predominant economic activity, with 85 percent of communes reporting agriculture as its primary economic activity. Livestock (35 percent of communes), mining (18 percent of communes), and commerce (16 percent of communes) are listed as the most common secondary economic activities (Figure 2.1). The capital, Bangui, is a notable exception in which commerce and civil service are reported as the main activities. The most common type of agriculture activity is cultivation of staple crops. For instance, 145 local authorities list cassava as the first cultivated crop in their communes, and another 100 authorities indicate peanut as the second most cultivated crop in their localities. This predominance of agriculture highlights low economic diversification, and is accompanied by no change in the structure of the economy compared to 2012. Nine out of 10 communes indicate that their current key economic activities were similar to those before the 2012 crisis. Figure 2.1: Main and secondary economic activities Agriculture dominates the economy of CAR, which highlights low economic diversification. In most communes, there is a predominance of agriculture activity, with …Moreover, livestock, mining activity, and commerce are the most the exception of Bangui common secondary activities While agricultural and livestock activities present the bulk of the economic activity in most communes, services which are crucial to support these activities have declined in the crisis years. For instance, only 4 percent of communes indicate that there is a store selling inputs such as fertilizers, seeds, and phytosanitary products in the capital of the commune, down from 10 percent before the 2012 crisis. And, only 41 percent indicate the existence of veterinary services in the commune, down from nearly 60 percent in 2012 before the crisis. Communes being the smallest administrative divisions in the country, they are at the forefront for the provision of services. Using data from the commune census the remainder of this section assesses the state of communes across the nation. This assessment is organized around three 10 pillars: state of the administration, availability of basic infrastructure, and provision of basic social and economic services. Consequently, in an attempt to produce a straightforward and sensible metric with which to compare the current level of development across communes, indicators across the three pillars are consolidated to produce a composite index, called the Local Development Index (LDI). The LDI, shedding light on current conditions, may aid decision makers in allocating resources and monitoring progress in communes. The LDI serves as a diagnostic tool intended to capture the current conditions across CAR. Before presenting the LDI and discussing the three pillars which comprise the LDI, this report first briefly describes the economic activities across communes. This report proceeds to discuss the three pillars of development in communes— local administration, local infrastructure, and basic social and economic services— and then presents the composite LDI. 2.2. Local administration To be effective, local administrations require personnel, resources, and a secure environment to work in. In most communes, local government offices, if they exist at all, are understaffed and underfunded, suggesting that basic government functions cannot be executed. Furthermore, gendarmerie and police forces are absent in a large share of the communes across the country. Commune administration offices are under-staffed and lack funding; and the difficulty in mobilizing resources has increased compared to 2012. Only 24 communes have 20 or more staff in the municipal office. Furthermore, regular payment of civil servants remains a challenge, where 9 in 10 communes report significant irregularities in paying their staff (Figure 2.2). One third of all communes indicate not having received a budget allocation for 2016. Only 6 communes have an annual per capita budget of 2,000 CFAF or higher, while 41 communes have an annual budget per capita ranging from 500 to 1,999 CFA Francs, and 75 communes have less than 500 CFAF. Nearly all communes (95 percent) report difficulties in mobilizing resources for government activities. This is a sharp increase compared to 2012 at which point only about half of the communes recall having such difficulties. Gendarmerie and police forces are absent in much of the central and eastern regions of the country (Figure 2.2). Only 1 in 3 commune capitals have a gendarmerie present, and only 1 in 5 commune capitals have a police force present. These forces are concentrated in the southwestern part of the country. 11 Figure 2.2: Local administration- funding and staffing in communes There is low capacity for local governance as communes lack staff and funding. This is combined with the absence of gendarmerie and police forces in many communes. Resource mobilization has become increasingly difficult after the crisis Difficulties at mobilizing resources before the cris and today (% of communes in each region that respond yes) 100 100 97 96 94 93 93 88 60 61 55 50 43 35 Before the crisis Today 2.3. Infrastructure Transport and communication infrastructure enables economic activity. However, essential infrastructure— such as electricity, mobile phone coverage, banking systems, and roads— is lacking throughout the country, and there are stark disparities between Bangui and the rest of the nation. 12 Table 2.1: Water, energy, and telecommunications infrastructure by region Availability in the principal town of the commune (percent of communes) Clean water Mobile International National National Electricity Internet (SODECA or phone radio radio television (ENERCA) access pump) coverage reception reception reception Region 1 36 18 55 95 50 9 14 Region 2 24 0* 50 97 6 0 15 Region 3 28 0* 35 74 2 0 7 Region 4 60 0* 33 83 3 0 3 Region 5 50 0 33 92 25 0 25 Region 6 17 3* 37 97 0 0 17 Region 7 88 100 100 100 100 100 100 CAR 36 12 43 89 15 6 16 * Infrastructure exists in an additional 6-7% of communes within the region but are not currently functioning; overall, only 57 % are functioning. Access to water and electricity infrastructure is low, with only 1 in 10 communes connected through the national electricity company (ENERCA) and 1 in 3 communes connected through the national water company (SODECA). Only 1 percent of commune capitals in rural areas have connectivity through ENERCA. On the other hand, all communes in Bangui report having electricity. Electricity infrastructure is functional in Bangui, in a few communes in the neighboring Ombella M’Poko prefecture, and in one commune each in Lobaye and Basse -Kotto prefectures. Water infrastructure is limited with only 1 out of every 3 communes reporting either distribution by the national water company (SODECA) or the availability of water pumps in the commune capital. Coverage in Bangui is high, but in other regions, 17 to 60 percent of communes have access to clean water infrastructure There is limited mobile phone coverage and minimal access to banking services. Only four out of 10 commune capitals have at least one mobile phone provider in the commune capital (Figure 2.3).1 The national radio and television stations and Internet are available in even fewer communes. In contrast, all communes in Bangui have mobile phone coverage. In other urban areas and rural areas, only 68 and 23 percent of communes, respectively, have at least one mobile phone provider in the commune capital. Furthermore, only 1 out of every 10 commune capitals have some form of banking system, which is either a bank agency or a local credit mutual. 1 In CAR, reported mobile phone providers are TELECEL, MOOV, ORANGE, AZUR, and SOCATEL. 13 The roads connecting communes to Bangui are often impassable through much of the year, and the cost of transportation is high, especially in remote areas. Half of the communes report that roads to Bangui are only accessible for some months out of the year with 50 percent reporting that roads are inaccessible 4 to 6 months of the year (Figure 2.3). About 3 in 4 communes report transport costs to Bangui at more than 25 CFAF per kilometer per person. Traveling from the Northeast to Bangui is the most expensive, while travelling from communes in the South close to Bangui is relatively cheaper. Most of the roads which connect the Northern communes to Bangui are not accessible for some months out of the year, while those which connect the Southwest communes to Bangui are accessible all-year round. Figure 2.3: Local infrastructure- mobile phone coverage, banking services, and roads Essential infrastructure—such as mobile phone coverage, banking services, and roads— is lacking in many communes. 2.4. Access to basic social and economic services Access to basic social services, such as public primary schools, health centers, and clean water is limited, especially outside commune capitals. Among the 10 largest localities (villages/quartiers) 14 of each commune, functional public primary schools and clean water sources (i.e. tap, pump, or protected well) are present in 44 and 43 percent of localities respectively, and only 18 percent have functional health centers (Figure 2.4). Access to basic social services in commune capitals is better. Most commune capitals have public primary schools and health centers, 83 and 62 percent of communes, respectively, implying that 17 respectively 28 percent of commune capitals lack basic education and health services! However, access to clean water and sanitation systems remains a challenge even in the commune capitals. Only 36 percent of the communes report having clean water access points in the capitals, with the national water company SODECA operational in just a handful of communes. Only six percent of communes report having a sanitation system. Figure 2.4: Access to basic services- primary schools and health centers Access to basic social services, such as primary schools and health centers, is limited. According to local authorities, a lack of facilities and shortage of qualified personnel are the largest barriers to provision of basic services. Local officials indicate that the provision of health services is hindered by a lack of health facilities and a shortage of doctors and nurses. For primary schools, the main challenges cited the insufficient number of teachers or the lack of teacher qualification (40 percent of communes), the lack of functioning schools (24 percent of communes), and insecurity (14 percent of communes). For secondary schools, similar challenges are cited with the lack of functioning schools cited as the greatest constraint. The number of public schools has decreased compared to the pre-crisis year, in particular in Region 3, Region 4, and Region 6. Table 2.2: Constraints in health and education (percent of communes) Health care (most important issue) Bangui Rest of country CAR Lack of functioning health facilities 43 30 31 Lack of physicians 14 23 22 Lack of nurses 0 22 21 Other 43 25 26 Total 100 100 100 15 Primary education (most important issue) Bangui Rest of country CAR Lack of functioning schools 43 23 24 Lack of teachers 0 24 23 Teachers are not qualified 14 17 17 Insecurity 14 14 14 Physical conditions (buildings) 0 6 6 Other 29 16 17 Total 100 100 100 Secondary education (most important issue) Bangui Rest of country CAR Lack of functioning schools 29 47 46 Lack of teachers 0 11 11 Insecurity 14 10 11 Physical conditions (buildings) 14 9 9 Teachers are not qualified 14 6 7 Other 29 16 17 Total 100 100 100 Note: These results represent the perspectives of the local authorities in the communes. NGOs are active in most communes across the country, partially compensating for some gaps in public service delivery. More than 83 percent of communes indicate that NGOs are active in their communes (Figure 2.5). Medecins Sans Frontières (MSF) is present in 28 communes in CAR, and most of its activities are in the Prefectures of Ouham, followed by Mambere Kadei. The second and third largest NGO present in the country are respectively, the Agency for Technical Cooperation and Development (ACTED) and the Danish Refugee Council (DRC). Most of their activities are in Ouham Pende. Figure 2.5 NGO Presence in the CAR NGOs are active across CAR, partially compensating for some gaps in public service delivery. Most established NGOs are the one that are more active in CAR Top 5 NGOs (% of communes) 18 17 16 14 14 11 12 10 8 6 6 6 4 2 0 MSF1 ACTED IRC DRC NRC MSF fills a gap for health services, with a focus on malaria, vaccinations and maternal and child health. They also provide assistance to refugees. ACTED has a broad range of programs, including, for example, aiding the return and recovery of internally displaced people. Its 16 emergency actions focus on the South-Eastern part of the country and on Bangui. IRC focuses its activities on medical care, water and sanitation services, and protection for vulnerable women and girls. Their activities are focused in Bangui, Ombella M’Poko, Nana Gribizi, Ouham Pende, and La Kemo. The DRC implements various programs, with emergency education and reintegration of children serving integral parts of its efforts. The NRCs largest programs are education and water and sanitation. In combination with limited health and education services, a formal judicial system is effectively absent. At the time of the survey, a total of 24 courts are reported to exist, but six of these 24 are not functioning, and four prefectures (Ouham, Kemo, Nana-Gribizi, Mbomou) report having no functioning court to settle legal disputes. As a result, citizens must resort to other conflict resolution mechanisms. In about 9 in 10 communes, traditional authorities are sought after to settle disputes. 2.5. Local Development Index (LDI): a composite measure of commune capacity The preceding sections have described the situation of the communes along three pillars: local administration, local infrastructure, and access to basic social and economic services. This section presents a Local Development Index (LDI) which combines 12 indicators from these three pillars into one composite index. The composite index is thus a measure of the state of service provision in each commune. Table 2.3: Local Development Index- components and weights Sub-index Weight Indicators Weight 2016 budget per capita in CFA (census 2003 1/3 Local population data) 1/3 administration Number of staff in the Mairie 1/3 Security, Gendarmerie or Police 1/3 Transport cost to Bangui (CFAF per km) 1/3 Infrastructure 1/3 Mobile phone reception in the commune capital 1/3 Banking in the commune capital 1/3 Share of 10 largest localities in the commune with 1/3 functional primary public schools Commune capital has a maternity 1/18 Commune capital has a hospital or a health center 1/18 Basic services 1/3 Share of 10 largest localities in the commune with 4/18 functional health centers Presence of SODECA or Adduction d'Eau 1/18 Sommaire in the commune capital 17 Share of 10 largest localities in the commune with clean water (bornes fontaines, forages, ou puits 5/18 proteges) The LDI ranges from 0 to 100, with higher values indicating the best possible scenario of service provision. Each pillar is also attributed a score which ranges from 0 to 100. The LDI is a weighted sum of these pillar scores, with equal weight (equal to one-third) for each pillar. However, within a given pillar, each indicator is not necessarily attributed the same weight, so that some indicators are weighted more heavily in the LDI than others. Definitions of the set of indicators for the index are further explained in the Annex. The LDI score is low for most communes; but, compared to other communes in the country, those located in Region 1, Region 2, and Region 7 are more likely to be in the top quintile of the LDI (Figure 2.6). The LDI is low for a large share of communes, indicating the need for substantial improvements across the country. Among the three pillars which form the LDI, local infrastructure has a higher variance across communes, whereas access to basic services is relatively homogeneous across communes. Communes in the southwestern regions Region 1, Region 2, and Region 7 rank higher in the LDI rankings. The regional average LDI for Region 1, Region 2, and Region 7 are above the national average, while the regional average LDI for Region 3, Region 4, Region 5, and Region 6 are below the national average (Figure 2.6). Five communes (Baboua, Berberati, Haute-Kade, Bouar, Yobe-Sangh) out of the 10 communes with the highest LDI are in the region 2. On the other hand, six communes (Lili, Yambele, Ngandou, Seliba, Kotto, Mbelima) out of the 10 communes with the lowest LDI are in Region 6 (See Table 6.2 in Annex for a comprehensive list of communes and their respective LDI). Figure 2.6: Local Development Index across communes The LDI score is low for a large share of communes; but, communes located in the southwest regions have relatively higher LDI. LDI score by commune LDI quintile by commune 18 Availability of local infrastructure varies most across communes, whereas access to basic services is more homogeneous across communes. South-western region perform better than average in the LDI ranking Average LDI scores by region 60% 47.7% 46.8% 50% 42.2% 38.8% 37.8% 37.8% 40% 34.2% 30.9% 30% 20% 10% 0% 19 3. Household Survey This part of the report presents descriptive analysis of data from the household survey. The household survey sample consists of 1,767 respondents. Via a household roster the household survey includes information on 10,551 individuals and information on school attendance for 3,535 school-aged children between 6-17 years old. Based on the household survey estimates, about 15 percent of individuals are currently internally displaced. Weights were used in analyzing the household survey data to adjust for the population shares based on the 2003 census in rural and urban areas in each of the 17 prefectures of the CAR, for a total of 33 strata (Bangui, which is its own prefecture, is strictly urban). Each household within a given stratum is assigned an equal weight. Recognizing the limitations of the sampling design and outdated sample frame, more complex weighting schemes were not utilized. Moreover, the broader findings are generally unaffected whether weights are used or not, including a simpler weighting strategy based on three strata – Bangui, other urban areas, and rural areas. Given the limitations to the sampling procedure and its implications on how representative the sample is of the current population, we present data from the survey, without necessarily making absolute claims. Due to the small sample sizes in smaller geographic areas, we limit the level of disaggregation of the analysis to higher levels such as by rural and urban, by region, by agro- ecological zone, and by mining vs non mining areas. For reference, the corresponding prefectures located in each region are summarized below in Table 1.1. We further caveat all description of the internally displaced individuals in this survey with the fact that displacement camps were not included in the sampling frame. Moreover, settlements for small mining were also not included, despite such sector potentially covering a non-negligible part of the population. Nonetheless, this survey fills an enormous gap as it collects information on a comprehensive range of outcomes from households spread across the entire country. 3.1. Profile of households Households are primarily large, young, male-headed, and rural. Approximately three out of five households are in rural areas, and two out of three households have five or more household members (Table 3.1). Households are large in both urban and rural areas. Only one in five households are female-headed households. While there are relatively few married individuals, most of the individuals are in formal unions. Alongside considerable ethnic diversity, there exists a small Muslim minority, around six percent of individuals. This is smaller than the 15 percent estimate from the 2003 census, and we caveat that this is likely due to limitations to the survey design. The population is fairly young; seven out of every ten men and of women are under the age of 25. 20 Table 3.1: Basic demographic characteristics Displaced Head of household Head of All individuals Characteristic household individuals (15% of all Urban Rural individuals) Residence 38 38 30 Urban Rural 62 62 70 Household size 36 37 35 1 to 4 members 5 to 8 members 47 48 47 > 8 members 17 15 19 Gender 83 51 52 79 85 Male Female 17 49 48 21 15 Ethnicity Gbaya 32 32 24 29 34 Banda 23 23 19 21 24 Mandja 9 9 8 10 9 Sara/kaba 7 7 30 4 9 Mboum 7 6 3 4 8 Ngbaka-bantou 6 7 2 8 6 Zandé-nzakara 4 4 2 3 4 Yakoma-sango 7 6 3 13 3 Other 5 6 9 9 4 Religion 93 92 87 90 95 Christian Muslim 6 6 12 9 4 Other 1 1 1 1 1 Civil status 51 37 35 40 58 Union libre Single 27 48 53 34 22 Married 10 6 5 10 10 Widow, widower 8 4 4 8 7 Separated/divorced 5 3 3 0 7 3 Agro-ecological zone 37 37 17 58 23 Equatorial forst Sudan-obanguien 34 33 22 29 37 Sudano-guinean 25 25 40 12 34 Soudan-sahelian 4 5 21 1 6 Most households are self-employed and in agriculture. In urban areas, seven in ten household heads are self-employed, and in rural areas nine in ten household heads are self-employed (Figure 3.1). In rural areas, nine in ten household heads have agriculture as the main income- generating activity, while in urban areas four in ten household heads have agriculture as their main income-generating activity. Apart from agriculture, household heads in urban areas are engaged in commerce and public administration. Households in Bangui are, as expected, different than in other urban areas of the country. In Bangui, only four percent report being in agriculture, 37 percent are engaged in commerce, and about 10 percent are students. Moreover, only half of those in Bangui report being self- employed. The analysis on the household survey will not attempt to make more explicit 21 distinctions between Bangui and other urban areas, primarily due to the small sample size in Bangui. Figure 3.1: Economic activity of the household head Households are primarily self-employed and in agriculture. Employment status in main income-generating activity, proportion of Main income generating activity, proportion of household household heads heads 8 4 9 5 2 34 4 20 8 90 22 85 69 37 Urban Rural Urban Rural Agriculture Commerce Public admin Other Own Salaried Welfare Employer Before describing conditions on education, wealth, and consumption, we discuss a few key dimensions that this report will explore: agro-ecological zones, mining vs non-mining areas, and internally displaced persons. Agro-ecological zones The country is divided into four agro-ecological zones (Figure 3.2). The Sudan-Sahelian zone is in the northern-most part of the country, where only 5 percent of the population reside. This overlaps with the hunting and tourism livelihood zone. The Sudan-Guinean zone, with 29 percent of the population, overlaps with the cotton, cereals and livestock zone. The Sudan- Obanguien zone, with 39 percent of the population, overlaps with the zone of cereals and livestock. The Equatorial Forest zone in the southern-most part of the country is where about 27 percent of the population reside. Based on the commune census, across all agro-ecological zones cassava is the main food crop, followed by groundnuts. There is some maize production in all agro-ecological zones, except in the Sudan-Sahelian zone, where millet is relatively more prominent. There is substantial cash crop production in the Equatorial Forest zone where over 70 percent of communes produce coffee, and over 20 percent of communes produce cotton. In Sudan-Guinean zone, 40 percent of communes produce cotton, whereas in Sudan-Obanguien zone about half of the communes produce coffee. Apart from the Sudan-Sahelian zone, there is some mining across agro- ecological zones. About half of the communes in the Equatorial Forest and Sudan-Obanguien zones report some mining activity. The data used in this analysis to classify agro-ecological zones comes from the commune level survey. As the four agro-ecological zones do not overlap perfectly with the administrative regions of the CAR, the classifications for the communes is a rough approximation. 22 Figure 3.2: Map of agro-ecological zones, collected from commune level data Mining There is some indication that some working-age men are missing from the sample, and that these missing men may be working in small scale mines. Males aged 21-40 comprise 23 percent of all males, whereas females aged 21-40 comprise 28 percent of all females. Apart from the obvious candidate explanation which is that the missing men may be involved in conflict-related activities or may have experienced higher mortality rates due to conflict, another potential explanation is that the missing men are working in small scale mining. In communes where mining is one of the top three economic activities 20 percent of males are between the ages of 21-40, whereas in communes where mining is not one of the top three economic activities 24 percent of males are between the ages of 21-40 (Figure 3.4). Moreover, those working in small scale mines may have migrated from non-mining communes to mining communes so that comparing missing males in mining and non-mining communes may be an understatement of the size of small scale mines. Figure 3.3: Age pyramid of individuals in sample >60 56-60 51-55 46-50 41-45 36-40 31-35 26-30 21-25 16-20 11-15 6-10 0-5 25 20 15 10 5 0 5 10 15 20 25 Percent of sample Male Female 23 On one hand, this finding highlights the fact that the survey did not include informal mining settlements, and as such we caution that this informal mining sector will be missing from the core of the analysis. In fact, under one percent of individuals report mining as their primary activity. On the other hand, this finding in and of itself highlights that the mining sector is possibly sizeable and important. A back of the envelope calculation from this household survey would suggest that 4 percent of males aged 21-40 may correspond to 100,000 individuals in a population of 5 million. This is not unreasonable, and in fact, is quite a conservative measure. In 2010, prior to the crisis, a report from the International Crisis Group (ICG) estimates about 80- 100,000 miners in the CAR, while a report from USAID report on Property Rights and Artisanal Diamond Development in the CAR estimate that there were about 400,000 miners including artisanal laborers. In this report, as an initial attempt to capture potential welfare effects of small scale mining, the welfare indicators of households in mining communes are compared to those in non-mining communes. Figure 3.4: Map of communes with mining in the top three economic activities Internal displacement The survey estimates that about 15 percent of household members are currently displaced, and more than half of the households have experienced the displacement of some or all of their household members since 2012 (Figure 3.5). We caveat this result by noting that the UN OCHA reports closer to 10 percent internally displaced persons. This difference again may be owing to the limitations to the survey and sampling design. In urban areas, 68 percent of households have experienced the displacement of some or all of the household members since 2012, and in rural areas 61 percent of households have. Among these households which experienced displacement, roughly half have had their entire household displaced since 2012. Compared to other regions, Region 3 and Region 5 have a larger share of current household members as internally displaced individuals (Figure 3.5). Specifically, about one out of three individuals surveyed in Region 3 and Region 5 is a displaced individual. In contrast, in the southern regions Region 1, Region 2, and Region 6 less than one out of every ten individuals is a 24 displaced individual. We again caveat that the figures for region 5 are more imprecise. Region 5 has fewer observations due to the fewer number of communes in that region. Compared to individuals who are not displaced, the displaced individuals are more likely to be Muslim and of the Sara or Kaba ethnicity (Table 3.1). Apart from this, there are no other striking demographic differences between displaced and non-displaced individuals— the civil status, gender, age, and urban-rural composition is similar between internally displaced and non- displaced individuals. Among the displaced, we see missing males as we previously noted for the overall population, in that the proportion of individuals who are aged 21-40 years old is about seven percent less for males than for females. Many of the currently displaced individuals have only been recently displaced, and those who have been displaced for longer tend to be located in urban areas. Overall, about one of every three internally displaced individuals have only been displaced less than 6 months ago. Those displaced into rural areas seem more transient than those displaced into urban areas. While one of every three urban displaced individuals have been displaced more than 2 years ago, less than one of every ten rural displaced individuals have been displaced more than 2 years ago. Figure 3.5: Internal displacement experience About 15% of individuals surveyed are currently displaced; About a third of these displacements occurred within the last 6 months. Number of months since displaced, proportion of individuals 39 35 32 32 26 14 15 7 < 6 months 6 to 12 13 to 24 > 24 months months months Urban Rural More than half of the households have experienced some displacement since 2012. Proportion of households which experienced displacement of a Of households that experienced displacement, status of the displaced member, since 2012 members 67 26 29 21 19 61 53 52 Urban Rural No return of displaced, past 6 months Urban Rural Some return of displaced, past 6 months All members are already there 25 3.2. Education Figure 3.6: Education levels of individuals in the sample Educational attainment is generally low, in particular in rural areas and for females. Level of education among individuals over 17 years, by area of Level of education among individuals over 17 years, by gender residence 46 41 45 37 Percent of sample Percent of sample 38 32 34 31 28 22 17 15 8 5 1 2 None, alphabet Primary Secondary > Secondary None, alphabet Primary Secondary > Secondary only only Urban Rural Male Female Improvements in education are stagnant for males, but the gender gap in access to education is narrowing. Level of education among males, by age group Level of education among females, by age group 9 7 10 Proportion of females 16 14 Proportion of males 34 27 33 24 65 40 35 72 40 38 42 67 42 48 28 19 19 27 25 6 to 17 18 to 35 36 to 45 over 45 6 to 17 18 to 35 36 to 45 over 45 Age group Age group None, alphabet only Primary None, alphabet only Primary Secondary > Secondary Secondary > Secondary Yet some children today are still missing school, and female children seem to be dropping out of school sooner. Proportion attending school regularly in the past 6 months, by age and Of those attending, proportion that had to miss some days of gender school in past 6 months, by age and gender 100 100 Percent of individuals Percent of individuals 80 80 60 60 40 40 20 20 0 0 6 7 8 9 10 11 12 13 14 15 16 17 6 7 8 9 10 11 12 13 14 15 16 17 Age Age Male Female Male Female Educational attainment is low, especially in rural areas. Among those 18 years and older, two out of every five individuals in rural areas have never gone to primary school, and only one out of 26 every five individuals in rural areas have at least completed primary school and acquired some secondary schooling (Figure 3.6). While education levels are higher in urban areas, they are nonetheless quite low—close to one out of every three individuals over the age of 18 have never gone to primary school. Females have lower levels of education than males, but improvements among younger female age cohorts are narrowing this gender gap. For males, there has been little to almost no improvement in education levels across all age cohorts. Among current school-aged males (6-17 years of age), 19 percent still do not have any primary education. This is only slightly lower than the group of males who were over 45 years old, for which 25 percent do not have any primary education. In contrast, education has been improving for younger female age cohorts. For females over 45 years old, 67 percent do not have any primary level education, whereas for current school-aged females, aged 6 to 17 years, 28 percent do not have any primary education. Although the gender gap is narrowing, females still seem to be stopping school earlier than males and are more likely to miss days of school. For example, the gender gap in current school enrollment is much more prominent for those 15 to 17 years old than those for younger ages. And, there are more females than males who are enrolled but having to miss days of school for those 12 to 14 years old. Among those who are currently enrolled, about 40 percent have had to miss days of school in the past 6 months. Of those who had to miss days of school, insecurity is reported by many to be a reason for having to have missed school (Table 3.2). Table 3.2: Reasons for not enrolling in school or missing school days Missing days Not enrolled Reason of school Percent reporting Lack of means 38 22 Insecurity 17 30 Lack of teacher 8 4 School is closed 8 5 Illness 5 28 Price of work 3 4 Vacation 2 1 Other 17 6 The data from the household survey cannot be used to distinguish between completing grades of primary school versus completing primary school education. But, a few estimates can be provided. The lower bound estimate on primary school completion would be the proportion of those who had completed at least some levels of secondary education more. For males, the lower bound estimate of primary school completion is 39 percent for those 18-35, 33 percent for those 36-45, and 37 percent for those 45 and older (Figure 3.6). For females, the lower bound estimate of primary school completion is 18 percent for those 18-35, 17 percent for those 36-45, and 9 percent for those 45 and older. This echoes the same trends of overall low and stagnant education 27 levels, higher education for men than for women, and the existence of some improvement in women’s education. 3.3. Wealth and food consumption Households own very few domestic and productive assets. Reflecting the prevalence of subsistence agriculture, small agricultural tools such as a hoe or daba are the most common asset, owned by six out of ten households (Figure 3.7). Ownership of more valuable productive farm assets are rare. For example, fewer than five percent of households own carts, ploughs, or oxen. Having lost many of their assets during the crisis, households are much less wealthy today compared to 2012. Households now have fewer productive assets, means of transport, and household items. For example, the share of households that own the following items decreased noticeably from 2012 to 2016: a bed or mattress (61 to 46 percent), a radio (42 to 23 percent), a bike (30 to 15 percent), a cart, plough, or oxen (from 10 percent to less than 5 percent). A wealth index is used to rank households and categorize them into wealth quintiles for distributional analysis. (See the annex for the construction of the wealth index.) Households in the bottom quintile of the index, or the most asset poor households, own almost nothing. Some of the most asset poor households, about one third, own hoes or dabas and mobile phones. In contrast, almost all the households in the top wealth quintile own household furniture, such as beds and mattresses, and dining and living room furniture. Figure 3.7: Ownership of household assets Households have fewer assets today than before 2012. Urban households are wealthier than rural households. Proportion of households which own an asset, pre 2012 vs 2016 Proportion of households which own an asset, urban vs rural PRODUCTIVE ASSETS PRODUCTIVE ASSETS Hoes, dabas Hoes, dabas Bike Bike Motorbike Motorbike Mill, crusher Mill, crusher Carts, rickshaw Carts, rickshaw Ploughs Ploughs Oxen Oxen HOUSEHOLD ITEMS HOUSEHOLD ITEMS Beds and mattresses Beds and mattresses Armchairs and coffee table Armchairs and coffee table Radio Radio Dining table and chairs Dining table and chairs Mobile phone Mobile phone Coal iron Coal iron Television Television Sewing machine Sewing machine 0 20 40 60 80 0 20 40 60 80 Rural Urban Pre 2012 2016 The bottom quintile own almost no assets. These households are disproportionately in the 28 Yada, Kagas and Fertit regions, which correspond to the two northern agro-ecological zones. Proportion of households which own an asset, by bottom and top Asset poor by region and agro-ecological zone, proportion of wealth quintile households in bottom 40 percent of wealth index PRODUCTIVE ASSETS Hoes, dabas Bike Motorbike Mill, crusher Carts, rickshaw Ploughs Oxen HOUSEHOLD ITEMS Beds and mattresses Armchairs and coffee table Radio Dining table and chairs Mobile phone Coal iron Television Sewing machine 0 20 40 60 80 Top Bottom The poor tend to be in the northern regions Region 3, Region 4, and Region 5, and similarly in the two northern agro-ecological zones (Figure 3.7). In Region 4, 51 percent of households are in the bottom 40 percent of the wealth index, and in Region 3 and Region 5, close to 60 percent of households are in the bottom 40 percent of the wealth index. Similarly, in the northern Sudan- Sahelian and Sudan-Guinean zones, over 50 percent of households belong in the bottom 40 percent of the wealth index. In contrast, in Region 6 and Region 7, only about 15 to 20 percent of households are in the bottom 40 percent of the wealth index. In addition to low wealth, households consume little and low quality food. In both urban and rural areas, households eat less than two meals per day on average. And during a week, households have had to reduce the quantity of food about two days in a week, reduce the number of meals 1.4 days in a week, and reduce the adult share of food about one day in a week (Figure 3.8). Households primarily consume staples (about five days in a seven-day period), and eat less meat and fish products. Over a seven-day period, a household on average consumes staple food product about five days, and in contrast consumes meat and fish products about only one to two days. The Food Consumption Score (FCS) is a tool designed by the World Food Programme to measure the frequency and composition of household food consumption based on the number of days in a seven-day period a household consumed each of nine different food categories. The FCS is used as a basis to categorize households into the following food consumption categories: poor (severely food insecure), borderline (moderately food insecure) or acceptable (food secure). (See the annex for construction of the FCS.) 29 Figure 3.8: Food consumption Households consume under two meals per day on average, and primarily consume staples and very little of meat and fish. Number of meals per day, and strategies to cope with lack of food Number of days household consumed various food products over 7- over 7-day period day period 2.3 7 1.8 6 1.7 1.6 1.5 1.5 Number of days 5 1.2 1.0 4 3 2 1 Meals per day Days reduced Days reduced Days reduced 0 quantity of food number of adult share of meals food Urban Rural Urban Rural The survey estimates that about six in ten households are categorized as having poor or borderline food consumption. Despite urban households being wealthier, there is little difference between the food consumption of urban and rural households. While 42 percent of urban households have acceptable food consumption, 36 percent of rural households do. While urban households consume slightly more of oil, sugar, and meat and fish products, they consume less of staples relative to rural households. Moreover, urban households have slightly fewer meals per day on average, and have had to more often reduce the quantity of food during a week. There is thus a significant strain in food consumption for urban and rural households alike. Households that produce their own food tend to have better food consumption. This is in line with the observation that urban households do not necessarily have better food consumption than rural households. Among other ways to source food, households tend to either purchase their food or produce their own food. For example, of households that consumed pulses over the past seven days, 51 percent primarily produced their own pulses and 41 percent primarily purchased their pulses. The less wealthy have poorer food consumption, although this relationship between wealth and food consumption is weak (Figure 3.10). Close to half of the households in the bottom wealth quintile are categorized as having poor food consumption, whereas in the top wealth quintile only 22 percent of households are categorized as having poor food consumption. Figure 3.9: Food consumption of those who produce food vs those who purchase food Households that produce their own food tend to have higher FCS than those who purchase 30 their food. Among households that consumed pulses in the past 7 days, food Among households that consumed vegetables in the past 7 days, consumption score by whether pulses were self-produced food consumption score by whether vegetables were self-produced Percent of Households Percent of Households 41 43 52 57 34 29 25 30 28 23 25 14 Purchased, or other Own production Purchased, or other Own production Poor Borderline Acceptable Poor Borderline Acceptable The geography of those with poor food consumption is similar, but not exactly equivalent, to the geography of the less wealthy. Food consumption tends to be worse in the two northern agro- ecological zones, where asset ownership is also low. In Region 1, Region 3, and Region 5, about 70-90 percent of households are considered to have poor or borderline food consumption. While Region 3 and Region 5 are less wealthy compared to other regions, Region 1 is not. Thus, although there exists some positive relationship between assets and food consumption, the relationship is weak. Figure 3.10: Food consumption by wealth and region Those with poor food consumption tend to be less wealthy and located in the two northern agro-ecological zone, which overlap with the Fertit, Yada, and Plateaux regions. Percent of households in food consumption score categories, by Percent of households with poor or borderline food consumption, by wealth quintile region and agro-ecological zone Percent of Households 26 26 37 53 50 29 30 31 28 29 45 44 32 18 22 Bottom Second Third Fourth Top Asset Wealth Quintile Poor Borderline Acceptable Households in mining communes have better food consumption, but do not appear wealthier than those in non-mining communes. In communes with mining as one of their top three economic 31 activities, a smaller proportion of households have poor food consumption and a higher proportion of households have borderline food consumption relative to the proportion in communes which does not have mining as one of their top three economic activities (Figure 3.11). But, households in such mining communes do not appear wealthier than those in non- mining communes. Thus, small scale mining may have some positive spillovers to households in the community in the sense of improved food security, but not in the generation of household wealth. This is not to say, however, that this is the same spillover on households who directly receive benefits from small scale miners, as the households which receive benefits from small scale mining may not be in the mining communities. Figure 3.11: Food consumption and wealth in mining communes Households in mining communes have slightly better food consumption, but are not wealthier than those not in mining communes. Percent of households in asset wealth quintiles, by whether household Percent of households in FCS categories, by whether household is is in a commune with mining activity in a commune with mining activity Percent of Households Percent of Households 44 37 38 40 22 27 20 35 36 41 35 25 Non mining commune Mining commune Non mining commune Mining commune Bottom 40% Middle 20% Top 40% Poor Borderline Acceptable Households who have received displaced individuals are currently less wealthy and have poorer food consumption, and displaced children are more likely to miss days of school than non- displaced children. Among households which have at least received one member who is currently displaced, 59 percent are in the bottom two quintiles of the asset wealth index and 69 percent have poor or borderline food consumption (Figure 3.12). In contrast, among households who do not currently have received displaced members, 36 percent are in the bottom two quintiles of the asset wealth index and 56 percent have poor or borderline food consumption. School-aged displaced children have the same rates of current school enrollment as the non- displaced, but the school-aged displaced children are more likely to have missed some days of school in the past six months. While those who received the displaced are poorer, there does not exist a strong relationship between whether a household had at least one member becoming displaced since 2012 with household wealth and food consumption. On one hand, households which have experienced (sent) displacement since 2012 are only slighlty less wealthy than those who did not experience displacement at all since 2012 (Figure 3.12). 32 Figure 3.12: Food consumption, wealth and internal displacement Households currently hosting at least one displaced member have fewer assets and poorer food consumption. Percent of households in asset wealth quintiles, by whether household Percent of households in FCS categories, by whether household currently hosts a displaced member currently hosts a displaced member Percent of Households Percent of Households 28 31 43 44 13 25 21 29 59 36 44 27 No member displaced At least 1 member displaced No member displaced At least 1 member displaced Bottom 40% Middle 20% Top 40% Poor Borderline Acceptable Households which have experienced (sent) displacement since 2012 are slightly less wealthy than those who did not experience displacement. Percent of households in asset wealth quintiles, by whether a Percent of households in FCS categories, by whether a household household experienced displacement (sent) since 2012 experienced displacement (sent) since 2012 Percent of Households Percent of Households 35 41 40 51 21 35 27 18 32 39 30 33 Not experience displacement Experienced some displacement Not experience displacement Experienced some displacement Poor Borderline Acceptable Bottom 40% Middle 20% Top 40% Currently displaced school-aged children are as likely to attend school than non-displaced children, but more likely to miss days of school. Proportion of school-aged individuals (age 6-17) attending school Of those attending, proportion of school-aged individuals (age 6- regularly in past 6 months 17) who had to miss some days of school in past 6 months 75 74 52 33 Not displaced Displaced Not displaced Displaced On the other hand, households which have experienced (sent) displacement since 2012 have slightly better food consumption than those who did not experience displacement. Altogether, these results suggest less of a clear pattern between displacement since 2012 and welfare. One 33 possible explanation is that the effect of displacement dissipates through time. An earlier caveat is worth mentioning: the survey does not capture huseholds in displacement camps. 3.4. Shocks and Coping Strategies Many households, especially urban households, were negatively affected by food price increases. This increase in food prices is in line with the observed low food consumption for urban households and for households which primarily purchase, instead of produce, their own food. Eight out of every 10 households have experienced an increase in food prices in the past 30 days (Figure 3.14). From a list of 24 negative shocks or events, 20 percent of urban households report that the increase in food prices was the shock that affected the household the most. Food prices are reported to have increased for most food items, especially meat and fish products, items which tend to not exist in the household diet. Food items for which prices seem to not have increased are those which are mostly not consumed by households, such as sorghum and millet (Figure 3.13). Relatively, the most stable prices are those for food items which are traditionally not cash crops such as cassava, maize, and peanuts. Figure 3.13: Change in food prices Food prices are reported to have increased for most food items. Percent of households reporting change in food prices in the past 30 days Game meat Goat, sheep Chicken Fish Pork Beef Sugar Cassava Oil Tea Groundnuts Maize Rice Milk, powder Milk, liquid Wheat flour Beans Pasta Bread Semolina Mil Sorghum 0 10 20 30 40 50 60 70 80 90 100 Increased Not increased Not consumed No response Households are still affected by conflict; one in three households experienced direct negative effects of conflict in the 30 days prior to the survey, such as conflict or violence in one’s locality, displacement, theft, destruction of assets, serious illness and death (Figure 3.14). In urban areas, 15 percent of households report experiencing conflict in their locality in the past 30 days, 34 whereas much fewer households in rural areas report such. Beyond conflict or violence in the village, a similar number of households experienced other direct effects of conflict, namely displacement, theft, destruction of assets, serious illness and death. So that overall, 33 percent of urban households and 36 percent of rural households experienced some direct negative effect of conflict in the past 30 days. More households in Region 4 and Region 6 report experiencing conflict in their locality in the past 30 days. In those regions, about one in ten households report experiencing conflict in their locality in the past 30 days. Households in mining communes have experienced slightly less conflict in the past 30 days, compared to households in non-mining communes. In mining communes, 14 percent of households experienced conflict in their locality in the past 30 days, compared to 21 percent of households experiencing conflict in their locality in the past 30 days. This would indicate that informal small scale mining settlements are experiencing more stable security. Other shocks also affect households in the CAR. In the past 30 days, more than half of the households that did not experience conflict or violence nonetheless experienced increases in transport costs and food prices (Figure 3.14). Moreover, in the past 12 months, households experienced naturally occurring events such as drought, irregular rainfall, and flooding. The food consumption score of those that experienced such shocks in the past 12 months is on average lower than those who did not experience such shocks. Figure 3.14: Experience of negative shocks Households experience a range of negative shocks—such as conflict and violence. In particular, the destruction of farms and increase in food prices affected households the most. Proportion of households that ever experienced a shock, by urban and Proportion of households reporting shock that affected them the rural most, by urban and rural PAST 30 DAYS Displaced or refugee PAST 30 DAYS Displaced or refugee Conflict in village Conflict in village Theft of valuable asset Theft of valuable asset Destruction of valuable asset Destruction of valuable asset Serious illness Serious illness Death of household member Death of household member Increased cost of transport Increased cost of transport Increased food prices Increased food prices Other conflict related Other conflict related PAST 12 MONTHS PAST 12 MONTHS Destruction of farms Destruction of farms Drought, irregular rain, flood Drought, irregular rain, flood High rate of animal disease High rate of animal disease Animal death Animal death Destruction of homes Destruction of homes 0 20 40 60 80 0 5 10 15 20 25 Urban Rural Urban Rural Households experience shocks which are not directly related to conflict, and these shocks 35 correspond to lower food consumption. Proportion of households which did not experience conflict or Percent of households in food consumption score displacement in the past 30 days, but which experienced other categories, by whether they experienced conflict related types of shocks increase in transport costs in past 30 days… 35 47 PAST 30 DAYS 31 Displaced or refugee 27 Conflict in village 34 Theft of valuable asset 26 Destruction of valuable asset Serious illness Not experienced Experienced Death of household member Poor Borderline Acceptable Increased cost of transport Increased food prices Other conflict related ... and by whether they experienced drought, irregular rainfall, or flooding in last 12 months… PAST 12 MONTHS Destruction of farms Drought, irregular rain, flood High rate of animal disease Animal death 34 46 Destruction of homes 0 20 40 60 80 32 Not affected by village conflict 26 Not displaced or refugee 28 34 Not experienced Experienced Poor Borderline Acceptable While households are subject to a variety of shocks, these same households have inadequate resilience to weather these shocks. The lack of resilience contributes to the link between the experience of shocks and low food consumption. Households have low capacity to cope with future shocks either because they depleted an existing resource to deal with a current shock or because they never had such resources to begin with. In the past 30 days, only 40 percent of households did not experience a potentially detrimental strategy to cope with the lack of food. A household is said to experience a given strategy if either they had recently engaged in the strategy or if they did not have that resource, which is a signal for prior depletion of that resource. Potentially detrimental strategies to cope with food shortages can be categorized into the following: stress, crisis, and emergency strategies. Stress strategies, such as borrowing money or spending savings, are those which indicate a reduced ability to deal with future shocks due to a current reduction in resources or increase in debts. Crisis strategies, such as selling productive assets, directly reduce future productivity, including human capital formation. Emergency strategies, such as selling one's land, affect future productivity, but are more difficult to reverse or more dramatic in nature. (Source: WFP CARI Technical Guide, November 2015) 36 Figure 3.15: Household coping strategies to deal with negative shock Many households do not have coping strategies available to them to deal with the lack of food. Percent of households reporting whether they did any of the following strategies to cope with lack of food or lack of money to purchase food in the past 30 days 1. URGENCE STRATEGIES Selling home or land Begging Practice illegal, risky activities 2. CRISIS STRATEGIES Selling productive assets or transport Reduce essential non-food expenses Remove children from school 3. STRESS STRATEGIES Spend savings Borrow from lender or bank Sale of domestic assets Send family member to seek work 0 10 20 30 40 50 60 70 80 90 100 Yes, did this No, did not have this resource No, there was no need Poor coping strategies, by region Poor coping strategies, by agro-ecological zone Regions with low food consumption scores tend to have poor strategies to cope with the lack of food. In Region 6, four in 10 households have exhausted or do not have available an urgency or crisis coping strategy, whereas in the Region 5 which has worse food consumption, eight in 10 households have expended or does not have available an urgency or crisis coping strategy. Thus, the less well-off households are also the less resilient households. While the most commonly used coping strategies by households are the less drastic stress strategies, such as spending savings, borrowing money or selling household assets, many households simply do not have these strategies as an option. Close to one in five households spent their savings or borrowed some money in the past 30 days to cope (Figure 3.15). But, what is more striking is that more households did not have these resources available to them. About one in three households do not have savings to deal with the lack of food, and similarly about one in three households do not have any domestic assets to sell to deal with the lack of food. Echoing the lack of domestic assets is the lack of productive assets or transport to sell in order to 37 deal with the lack of food. While 10 percent of households sold a productive asset (a crisis strategy) to deal with the lack of food, another 42 percent did not have such productive assets to sell. So that overall, at least 52 percent of households are said to have experienced this particular crisis coping strategy. Figure 3.16: Wealth, negative shocks, and food consumption Asset rich households are not more resilient to shocks than asset poor households— food consumption of the wealthy is more negatively correlated with droughts, irregular rainfall or flooding. Percent of households in food consumption score categories, by whether they experienced drought, irregular rainfall, or flooding in last 12 months… …among households in bottom two wealth quintiles …among households in top two wealth quintiles 31 28 44 57 30 29 33 24 39 43 19 23 Not experienced Experienced Not experienced Experienced Poor Borderline Acceptable Poor Borderline Acceptable It is not necessarily true, however, that households with more assets are better able to weather shocks. The food consumption of wealthy households is more affected by droughts, irregular rainfall, or flooding. Relative to households in the bottom two wealth quintiles, the food consumption of households in the top two wealth quintiles is more negatively correlated with experiencing drought, irregular rainfall, or flooding (Figure 3.16). This is in line with the earlier result in which we see a weak correlation between wealth and food consumption. One potential explanation is that previously wealthy households have sold their assets to improve food consumption in response to a shock, so that they now appear poorer. Another explanation is that even the rich do not have the right assets or do not have decent market access to divest their assets in response to a shock, so that the sale of assets is not an effective coping strategy. Regardless, this mismatch between wealth and resilience highlights the fact that coping strategies are likely weak for both the wealthy and the poor. Community groups can be a conduit for mutual support, but membership in community groups is generally low. Thus, these community groups are an unlikely coping strategy for households in the CAR. Apart from saving, borrowing and selling assets, community groups can also be leveraged as a channel for mutual support and exchange for households to cope with shocks, but there is little involvement in such semi-formal community groups. In urban areas, only 24 percent of households in urban areas and 38 percent in rural areas have a household member who is part of a community group (Figure 3.17). Membership in these community groups has been consistently low. Compared to before 2012, there has been no change in the proportion of households which was part of a community group or association. Membership in tontines, small 38 group saving schemes, is also low, roughly one out of every five households were part of a tontine in the past six months. It is likely, however, that other mutual support exists at a less formal level, such as informal networks of friends and neighbors. Figure 3.17: Membership in community groups Community groups can be an important source of mutual support, but membership in community groups is low. Proportion of households which have at least one member in a group Proportion of households reporting a reason for participating in a or association in the past 6 months group or association in past 6 months 83 69 38 50 47 45 24 39 37 34 Mutual Economic Receive aid Moral support Part of a group/association (past 6 months) support benefits from NGOs Urban Rural Urban Rural 39 4. Perceptions and Priorities Objective measures of access to services at the commune level and living conditions at the household level highlight the need for substantial improvements in the provision of a broad range of services. Beyond such objective measures, the survey further elicited from households and commune representatives, their perceptions of security and socio-economic welfare, as well as their views on priorities for development. Information on perceptions of security and welfare provides a glimpse into current progress and optimism for future progress. Moreover, while substantial and simultaneous improvements across all areas are ideal, resources are scarce and understanding the urgency of various interventions can aid decision-makers in timing these interventions. Information on priorities for development contributes to this end. 4.1. Perceptions of security and socio-economic welfare Many households perceive that the security situation in their localities is stabilizing. Most households perceive the security situation to be improving or to have remained the same in the past six months. Only under 15 percent of households perceive the security situation to have worsened (Figure 4.1). Compared to other regions, more households in Region 1, Region 2, and Region 7 perceive the security situation to have improved. This perception is consistent with recent experience of conflict, wherein more households in Region 4 and Region 6 experienced conflict in their locality in the past 30 days. Figure 4.1: Household perception of the security situation Households perceive the security situation to be improving, especially in the southwestern regions around the capital. Proportion of households reporting perceived changes in Proportion of households reporting improved security in security in their locality over the past 6 months their locality in the past 6 months, by region 54 32 14 Become better Remained the Become worse same Among both households and commune representatives, trust in the FACA is high, while trust in rebel armed groups is low. About nine in 10 households and nine in 10 commune representatives 40 trust the FACA, compared to only one in 10 households and one in 10 commune representatives which trust armed groups (Figure 4.2). The same pattern on trust in FACA vs armed groups holds across regions and for both communes with and without armed group presence. Figure 4.2: Trust in different armed forces among local authorities and households Among both households and among commune representatives, trust in FACA is high, while trust in rebel armed groups is low. Proportion of households reporting their trust in various forces Proportion of commune representatives reporting trust in various forces 100 100 90 90 80 80 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 Distrust Neutral Trust Distrust Neutral Trust Perceived improvements in conflict and ethnic and religious tensions are contrasted with the perceived worsening of the general economic condition and access to basic services over the past six months (Figure 4.3). These perceptions are similar for households and commune representatives. Figure 4.3: Perception on the economic situation among local authorities and households Perceived improvements in conflict and ethnic and religious tensions are contrasted with the perceived worsening of the economic condition. Household perceptions of changes in security and socio-economic conditions in the past 6 months 100 80 60 40 20 0 Ethnic/religious Violence Acces to Access to Access to clean Food security Economic Employment tensions education health care water condition Worsen Stay same Improved 41 Commune representative perceptions of changes in security and socio-economic conditions in the past 6 months 100 80 60 40 20 0 Ethnic/religious Violence Acces to Access to Access to clean Food security Economic Employment tensions education health care water condition Worsen Stay same Improved About eight in 10 households perceive that the employment situation has worsened. About seven in 10 households perceive food security to have worsened, which is unsurprising given the generally low food consumption score across households. Moreover, about half of the households perceive access to basic services, such as education, health care and clean water, to have worsened. This highlights the fact that an 80 percent enrollment rate in primary school is likely missing much of the education picture—such as quality of education and access to post- primary education. However, most households expect that their quality of life will improve over the next 12 months (Figure 4.4). The perceived worsening economic condition matches the household perception that the quality of life has stayed the same in the past six months for 25 percent of households and has worsened for 44 percent of households. But, most household expect that their quality of life will improve over the next 12 months. Similar to the regional patterns on security, most of those who expect improvements in the quality of life are in Region 1 and Region 2 around the capital. Figure 4.4: Perceptions on future improvements in the quality of life But while most did not see improvements in the quality of life over the past 6 months, many are optimistic that the quality of life will improve over the next 12 months. Proportion of households reporting past and future changes in the Proportion of households expecting future changes in the quality quality of life of life, by region 13 44 11 14 25 62 31 In the past 6 months In the next 12 months Improved Stay same Worsen Can't say 42 4.2. Priorities for Development There is broad consensus among both local authorities and citizens that the highest priority for the country is achieving peace, reconciliation, and security. This is followed by good governance and the provision of basic services, and then economic development. With respect to specific policy interventions, the top five interventions prioritized (among all options) by survey respondents are the following, in order of importance: (i) improve security through a reformed FACA (army) operating across the territory; (ii) re-integrate ex-combatants; (iii) improve access to basic services (health, education, water); (iv) improve road infrastructure; and (v) strengthen agricultural and livestock production (Figure 4.5). Figure 4.5: Priorities for development Peace, reconciliation and security are the top priority; and improving security via a reformed FACA is cited as top priority action. Ranking of development priorities, proportion of households Ranking of development priorities, proportion of commune representatives 13 10 12 35 5 29 52 62 86 65 75 53 33 30 11 15 6 8 Peace and Governance and basic Economic Peace and Governance and Economic reconiciliation service development reconiciliation basic service development 1st Priority 2nd Priority 3rd Priority 1st Priority 2nd Priority 3rd Priority Top priority intervention 45 41 13 14 12 8 10 7 7 6 Improve security Reintegrate ex- Improve access to Improve roads Strengthen with reformed combatants basic services agriculture and FACA (army) livestock Household Commune Respondents were also asked about the top priority within each of the three categories. With respect to the first category of peace, reconciliation and security, overwhelming support exists for restoring security across the country through reformed defense forces (FACA), consistent with the results above (Table 4.1). A distant second is the reintegration of ex-combatants. As for good governance and provision of basic services, redeploying the administration throughout the 43 country was cited the most, in particular by residents of Bangui. On the other hand, improving access to basic services is a high priority among residents living in other parts of the country where access to health and education services are extremely limited. When it comes to economic development, job creation is a high priority among residents of Bangui, while strengthening the agriculture and livestock sectors, and building roads are priorities for residents living in other areas, in particular in the more isolated eastern regions, Region 5 and Region 6. Table 4.1: Priorities for development Priorities identified within each category (percent of households) Peace, Reconciliation, and Security Other Bangui Rural CAR (Highest priority within group) Urban Restore security across CAR through reformed defense forces (FACA) 53 57 57 57 Reintegrate ex-combatants 16 11 15 14 Restore security across CAR through reformed security forces (police and gendarmerie) 1 11 9 8 Support reduction of violence 5 6 7 7 Setting a policy for national reconciliation and social cohesion 7 6 5 6 Good Governance and Provision of Basic Services Other Bangui Rural CAR (Highest priority within group) Urban Redeploying the administration across the country 33 23 22 24 Provide basic services to the population, in particular health 12 24 26 23 Provide basic services to the population, in particular education 13 15 20 18 Provide basic services to the population, in particular water 8 14 12 11 Assure food security 4 13 9 9 Economic Development Other Bangui Rural CAR (Highest priority within group) Urban Strengthen and develop agriculture and livestock sectors 22 33 35 33 Build roads 12 27 37 30 Create jobs 36 20 12 18 Strengthen and develop extractive industries 4 5 5 5 Develop professional training programs 5 4 3 3 Note: The questionnaire asked respondents to choose the highest priority intervention within each category. Only top 5 responses shown. 44 5. Conclusions This report presents results from a national survey conducted in August 2016, the Enquête Nationale sur les Monographies Communales, which is aimed at filling a gap on baseline information on the situation in the CAR and forming the basis for a monitoring system on service provision and development outcomes. The survey consists of two components: a commune census which elicited information from local authorities, and a household survey which elicited information from households in nearly all communes across the country. This section summarizes the findings from the commune and household components of the survey, and elaborates on a monitoring system for the CAR. The commune census reveals limited administration, infrastructure, and basic services across communes. Similarly, the household survey reveals low and stagnant education levels, little wealth, poor consumption, and high vulnerability to shocks. Communes and households in the southwest of the country tend to do better, although large improvements have to be made across the country. 5.1. Summary of Findings: Commune census Commune administration offices are under-staffed and lack funding; and the situation has only worsened since 2012. Only 24 communes have 20 or more staff in the municipal office, with regular payment of such staff remaining a problem. Moreover, 57 communes indicate not having received a budget allocation for 2016. Access to basic infrastructure such as electricity, mobile phone coverage, banking services and road networks is low. Only about 1 in 10 communes report having electricity through ENERCA connection or some form of public lighting in the commune capital, and only 1 percent of commune capitals in rural areas have connectivity through ENERCA. Overall, only 4 out of 10 commune capitals have at least one mobile phone provider in the commune capital. Furthermore, only 1 out of every 10 commune capitals have some form of banking system — either a bank agency or a local credit mutual. Half of the communes report that roads to Bangui are not accessible throughout the year. Access to basic social services, such as public primary schools, health centers, and clean water is limited, especially outside commune capitals. Among the 10 largest localities of each commune, functional public primary schools and clean water sources are present in 44 and 43 percent of localities respectively, and only 18 percent have functional health centers. Access to clean water and sanitation systems remains a challenge even in the commune capitals. Only 36 percent of the communes report having clean water access points in the capitals. A Local Development Index is constructed to combine indicators across three pillars: local administration, local infrastructure, and access to basic social and economic services. By 45 shedding light on current conditions in a simple and straightforward way, the LDI may aid decision makers in allocating resources and monitoring progress in communes. 5.2. Summary of Findings: Household Survey The survey estimates that 15 percent of the individuals are currently internally displaced. Furthermore, half of the households had at least one member which was ever displaced since the onset of the crisis in 2012. The displaced are disproportionately in Region 3 and Region 5. Households which have currently received displaced individuals are less wealthy and have worse food consumption, and displaced children tend to miss more days of school. But households which have had at least one member ever become displaced since 2012 are today only slightly less wealthy and have better food consumption, suggesting less of a clear relationship between wealth, welfare and the experience of displacement over a longer time horizon. While households perceive the general economic condition to be worsening, they perceive the security situation to be improving. About one in three households have experienced direct negative effects of recent conflict or violence (in the past 30 days). Nonetheless, many perceive that the security situation has improved, particularly for those in Region 1 and Region 2 around the capital. The majority of households perceive that ethnic and religious tensions have improved and are more optimistic about improvements in the quality of life over the next year. Education levels of individuals have been stagnant. Education levels are lower in rural areas and lower for women. While there has been some improvement in female education levels, female children are still stopping school sooner. While current enrollment rates seem high, at around 80 percent among current school-aged children, half of households perceive that access to education has worsened in the past six months. Thus, current enrollment may not be an adequate indicator of education as it might not reflect the following: completion rates, days of attendance during the school year, or the quality of education. In fact, 40 percent of those currently enrolled had to miss days of school in the past 6 months, with insecurity mentioned by many as a reason for having to have missed school. Among those over 18 years, only 37 percent of males and 17 percent of females have at least some secondary education. Households have very little assets, and they have much fewer assets today than before 2012. Furthermore, households have little and low quality food consumption with very little difference between urban and rural areas. Households only own simple agricultural tools, and almost none of other productive farm assets. Six out of every ten households are categorized as food insecure (particularly having poor or borderline food consumption). On average, households eat less than two meals per day, and mainly consume staples. Households in the two northern agro-ecological zones are poorer and have poor food consumption. Despite urban households being wealthier, their food consumption is not any better than that of rural households. Consistent with this finding, food consumption is worse for household which do not produce their own food and the increase in food prices has affected many households. In combination with low wealth and poor consumption is high vulnerability. They are affected by a variety of shocks, and they do not have the sufficient strategies to cope with the exposure to shocks. Many households do not have (or have already expended) even the more primitive 46 means to cope with the lack of food, such as spending savings, borrowing money, or selling household assets. Furthermore, households which have more assets do not seem more shielded from shocks than households which have fewer assets, suggesting that even those with assets do not seem able to use the sale of their assets as an effective coping strategy. Membership in community groups and tontines is also very low despite the fact that such groups can serve as effective channels to cope with certain types of shocks via mutual exchange in communities. Households and local authorities agree that peace, reconciliation, and security should be the main development priority. About eight in 10 households and nine in 10 commune representatives rank peace, reconciliation, and security as the top priority over good governance and the provision of basic services and economic development. Many trust in the FACA (army) and only a few trust in armed rebel groups, and this pattern holds across regions. Within the peace, reconciliation, and security priority, the majority deem the restoration of security through the FACA as the key policy intervention. 5.3. Building a Monitoring System The Monographies Communales survey, thanks to its nationwide coverage and demonstrated ability to collect information rapidly and in a cost-effective manner, can serve as the core for a new monitoring system. By repeating this survey regularly, it can feed a monitoring system that assesses progress towards RPBA objectives and presents an objective feedback loop for the state and citizens. Using a select set of indicators from the survey, a dashboard can be created to present information on tangible development outcomes (e.g. number of functional schools; reach of the mobile phone network; availability of transport) as well as report on citizen perceptions (e.g. security; trust; whether life is improving). The dashboard can integrate complementary information from other sources (e.g. prices; displaced people; quality of roads) to offer a more complete picture of the state of the nation. When these indicators are measured regularly over the course of the RPBA, they can demonstrate trends over time with respect to RPBA implementation. Suggested is to repeat the survey at least once every six months to track changes. As a complement to the dashboard, progress made could be reflected in a national implementation report published on an annual basis. Such a report would highlight achievements and challenges of RPBA implementation and could also include information gathered through focus groups. Monitoring alone serves little purpose if the information collected is not analyzed in a timely manner, put into context, and interpreted in ways that facilitate decision-making. A small team of 2 or 3 dedicated analysts, with an independent statute, that operates under the guidance of RPBA partners and the Authorities could ensure that relevant lessons are drawn from the available information. The team would report on specific topics, produced at the request of its principals. The analytical team would work closely with the National Statistical Institute, ICASEES, responsible for the implementation of the Monographies Communales survey, and help identify areas and issues on which additional information should be collected. 47 Results need to be communicated to different stakeholders to ensure that monitoring processes are meaningful and support better development outcomes. As such, an analysis and communications strategy should be an integral part of the monitoring framework. In addition, to facilitate transparency and confidence building, the dashboard and its underlying data should be made publicly available, including through a web-portal. 48 6. Annex 6.1. Survey Design The survey design for the Monographies Communales survey was tailored to the current context in CAR, that is, a high level of insecurity, the need to collect data in a short period of time, and hard budget constraints. A commune census covering all 179 communes in the country was a central element of the design. As a supplement, a household survey component was added to obtain perspectives of citizens. Given that armed groups continue to be active across the territory, the safety of the survey teams was a critical consideration. The practical and cost- effective solution for the household survey component was to limit the additional travel required of enumerators in the field. Although the ideal sampling methodology could not be used, the household socio-economic characteristics of respondent are in line with other national surveys. The combination of the commune and household surveys provides a good representation of the country and is a valuable source of information in filling data gaps and monitoring progress. The field work for the Monographies Communales survey was conducted by 22 survey teams over a 3 week period, from August 3, 2016 to August 24, 2016. Interviews were conducted primarily in French for the commune census and in Sangho for the household survey, and paper questionnaires were used for both. Enumerators participated in a week long training and piloted both the commune and household survey questionnaires in Bangui and two communes in the Ombella M’Poko prefecture. Commune census As all 179 communes in the country were covered, no selection procedures were required for this component. In the principal town (chef-lieu) of each commune, survey teams interviewed representatives from the local government (mairie) and other community leaders for the commune census. Considering that respondents would have more accurate information on their immediate environment, the questionnaire focused primarily on the situation in the chef-lieu of the commune to improve the reliability of the data collected. Household survey The planned total sample size was 1790 households. In each commune, 10 households were selected to be interviewed: 5 households from a randomly selected neighborhood of the chef-lieu, and 5 households from a randomly selected village located 20 to 40 km from the chef-lieu. Selection of the localities (village or quartier) and households: The list of localities in the commune from the 2003 census was first updated through consultations with local authorities. The names of villages located 20 to 40 km from the chef-lieu were written on pieces of paper, and one was randomly drawn by a local official. This same process was used to select one of the neighborhoods in the chef-lieu. In each of the selected localities, a simple listing of households was completed, up to a maximum of 100 households. Using this list, 5 households were then randomly selected for interviews. Weights were used in analyzing the household survey data to adjust for the population shares based on the 2003 census in rural and urban areas in each of the 17 prefectures of the CAR, for a 49 total of 33 strata (Bangui, which is its own prefecture, does not have a rural area). Each household within a given strata is assigned an equal weight. 6.2. Methodology Local Development Index The indicators used to construct the overall composite index measure the effectiveness of government presence, the state of infrastructure, and the access to basic services in the communes. First, the indicators set out the degree of the state presence. This represents the first pillar and it is captured through a number of indicators, such as the budget per capita (in local currency) allocated to the commune, the number of working staff in the Mairie, the presence of security forces (gendarmerie and police). The second pillar of indicators assesses the availability of basic infrastructures, such the existence of a mobile phone network, a banking system, and the transport cost per km (to conjecture the cost of mobility across the country). The third and final pillar measures the availability basic services in the communes, such as the availability of public primary schools, the availability of a health center, a sanitation system, and access to clean water. The three pillars (local administration, infrastructure, and basic services) build up the overall composite index. The maximum score for each pillar is 100 with each pillar equally weighted. That is the weight for each pillar is one-third. However, some indicators within a pillar were given a higher importance than others, and were therefore attributed different weights (see Table A1, below). Some of the indicators used to construct the sub-indices have missing data points. However, the share of missing values for each indicator was usually below five percent of communes. To address the missing data points and ensure that all 179 communes can be compared on the LDI, values were imputed when missing. Specifically, the median (for continuous variables) or the mode (for categorical variables) across communes for which data was available at the prefecture level was used to replace missing values. Given the small number of communes for which some indicators were missing, the imputed values do not dramatically affect the sub-indices, and therefore do not affect the ranking of each Commune on the LDI. Table 6.1 Indicators used to construct the sub-indices and overall composite index Indicators used to construct the sub-indices and overall composite index Indicator name Measure Weight Local administration 2016 budget per 1 = Highest >4999 1/3 capita in CFA (census 2003 0.75 = High 2000-4999 population data) 0.5 = Middle 500-1999 0.25 = Lowest 1-499 0 = No 50 Indicators used to construct the sub-indices and overall composite index Indicator name Measure Weight Number of staff in the Mairie 0.75 = Highest >19 1/3 0.5 = Medium 10-19 0.25 = Low 1-9 Security, Gendarmerie 1 = Yes 1/6 0 = No Security, Police 1 = Yes 1/6 0 = No Infrastructure Transport cost per km (in 0 = Most expensive >74 1/3 CFA) to Bangui 0.33 = Expensive 50-74 0.66 = Fair 25-49 1 = Cheapest 5-24 Mobile phone reception 1 = Yes, Chef lieu has at 1/3 least one provider 0 = No Banking 1 = Yes, there is some 1/3 form of banking instruments (agences, guichet de banque, caisses d'epargne et credit) in the Chef lieu 0 = No Basic services Share of localities with (share) 1/3 functional primary public schools in the commune Chef-lieu of the commune has 1 = Yes 1/18 a maternity 0 = No Chef-lieu of the commune has 1 = Yes 1/18 a health center 0 = No Share of localities with (share) 4/18 functional health centers in the commune Presence of SODECA or 1 = Yes 1/18 Adduction d'Eau Sommaire in 0 = No the commune capital Share of localities with bornes (Share) 5/18 fontaines, forages, ou puits proteges in the commune 51 Table 6.2 Ranking of communes on the Local Development Index (LDI) Ranking of communes on the Local Development Index (LDI) 5th quintile (top) 4th quintile 3rd quintile 2nd quintile 1st quintile (bottom) (a) (a) (a) (a) Rank Pref. Commune LDI Rank Pref. Commune LDI Rank Pref. Commune LDI Rank Pref. Commune LDI Rank Pref.(a) Commune LDI 1 36 Bah- 72 108 144 Senkpa- 11 Bossembélé 84.3 31 Bessar 50.5 71 8e Arrondi 40.2 12 Baléloko 34.2 21 M'Baéré 26.4 2 37 73 109 145 32 Bossangoa 77.1 43 Grimari 50.3 22 Abba 40.0 43 Kouango 33.9 32 Moyenne Si 26.1 3 38 74 110 Nana- 146 22 Baboua 72.7 61 Kémbé 50.0 41 Galabadja 40.0 32 Marko 33.8 43 Kobadja 26.0 4 39 75 111 147 21 Berberati 71.2 62 Rafaï 49.6 22 Groudrot 39.9 31 Dilouki 33.5 62 Gambo 26.0 5 40 76 112 148 21 Haute-Kade 69.4 63 Zémio 49.6 51 Dar-el-Kou 39.9 21 Ouakanga 33.1 52 Ouadda 25.9 6 41 Birvan- 77 113 149 11 Bimbo 67.2 31 Bol 49.6 31 Lim 39.8 43 Haute-Baïd 32.9 43 Koudoubégo 25.9 7 42 5e 78 114 150 22 Bouar 65.1 71 Arrondi 49.6 31 Bozoum 39.7 43 Pouyamba 32.9 32 Hama 25.9 8 43 79 115 151 Zotoua- 23 Yobé-Sangh 64.7 23 M'baéré 49.5 32 Bédé 39.7 23 Bilolo 32.9 22 Bangarem 25.8 9 44 80 116 152 71 1er Arrond 63.6 61 Alindao 49.0 51 Vassako 38.8 41 Ngoumbélé 32.8 32 Soumbé 25.4 10 45 Bawi- 81 117 153 12 Pissa 63.4 22 Tédou 49.0 31 Yémé 38.3 21 Topia 32.7 61 Bangui-Ket 25.1 11 46 82 118 Ouham- 154 21 Carnot 63.1 12 M'bata 48.7 43 Ngoubia 38.0 32 Bac 32.6 63 Djémah 24.9 12 47 7e 83 119 Sayo- 155 11 Yaloké 63.0 71 Arrondi 48.3 31 Koui 37.1 62 Niaka 32.5 32 Ladi-Gbawi 24.1 13 48 Nana- 84 120 Kotto- 156 21 Basse-Boum 61.7 32 Bakas 48.2 31 Daneyérin 37.0 61 Ouba 32.5 11 Guézéli 23.3 14 49 Bouca- 85 121 157 43 Bambari 61.7 32 Bobo 48.1 42 Botto 37.0 21 Basse-Bato 32.3 62 Vougba-Bal 23.1 15 50 86 Haute- 122 158 11 Begoua 60.2 12 Boda 48.1 21 Boum 36.9 31 Malé 31.4 62 Bakouma 22.7 16 51 87 123 Ouham- 159 12 Mongoumba 60.0 31 Mom 47.6 53 Vokouma 36.2 32 Fafa 31.3 31 Bimbi 22.4 17 52 88 124 Pladama- 160 71 4e Arrondi 59.8 63 Mboki 47.1 61 M'Boui 36.2 43 Ou 31.0 42 Nana 21.7 18 53 89 Nanga- 125 161 41 Sibut 59.7 23 Nola 46.2 32 Bogu 36.1 31 Banh 30.9 32 Ouaki 21.2 19 54 Doaka- 90 126 162 21 Basse-Kade 59.5 22 Kour 46.1 42 Grevaï 36.1 22 Béa-Nana 30.9 61 Yabongo 20.5 20 55 91 Koro- 127 163 31 Paoua 59.0 31 Kouazo 45.8 32 M'pok 35.9 61 Bakou 30.1 21 Mbali 20.3 21 56 Danga- 92 128 Fafa- 164 52 Bria 57.9 43 Gbou 45.3 31 Péndé 35.9 32 Boung 29.7 31 Loura 20.3 22 57 93 129 Baïdou- 165 62 Bangassou 57.0 41 Tilo 44.4 12 Boutélossi 35.9 43 Ngo 29.7 22 Yoro-Samba 19.9 52 Ranking of communes on the Local Development Index (LDI) 5th quintile (top) 4th quintile 3rd quintile 2nd quintile 1st quintile (bottom) Rank Pref.(a) Commune LDI Rank Pref.(a) Commune LDI Rank Pref.(a) Commune LDI Rank Pref.(a) Commune LDI Rank Pref.(a) Commune LDI 23 Basse- 58 94 130 166 21 Mamb 57.0 43 Ippy 44.4 43 Cochio-Tou 35.9 22 Koundé 29.6 43 Yéngou 19.6 24 59 3e 95 131 Bogongo- 167 31 Binon 56.5 71 Arrondi 44.4 41 Mala 35.6 12 Ga 29.4 43 Lissa 19.6 25 60 Ouandja- 96 132 168 11 Boali 55.9 52 Ko 44.1 31 Nana-Barya 35.6 61 Siriki 29.3 62 Zangandou- 19.1 26 61 97 133 169 11 Damara 55.6 61 Mobaye 43.6 12 Lessé 35.4 32 Bakassa 29.2 61 Guiligui 18.6 27 62 98 Ndoro- 134 Mbolo- 170 22 Herman-Bro 54.9 12 Nola 43.3 32 Mbol 35.4 51 Kpat 28.6 61 Mbélima 18.6 28 63 99 135 171 12 M'baïki 54.1 11 La Mbi 42.6 31 Mia-Péndé 35.3 52 Yalinga 28.1 61 Kotto 18.6 29 64 100 136 172 41 Dékoa 53.9 32 Sido 42.4 12 Boganda 35.3 71 2e Arrondi 28.0 61 Séliba 18.1 30 65 101 137 173 42 Kaga-Bando 53.7 31 Bocaranga 41.7 32 Ben-Zambé 35.0 53 Ridina 27.9 22 Niem Yelew 16.6 31 66 102 138 174 42 M'Brès 53.4 12 Moboma 41.6 23 Salo 34.7 53 Ouandja 27.9 22 Bingué 15.5 32 67 103 139 Daba- 175 32 Batangafo 52.7 42 Ndénga 41.3 12 Lobaye 34.6 52 Nydou 27.7 43 Azéngué-Mi 14.4 33 68 104 140 Dan- 176 22 Fô 52.3 41 Galafondo 41.2 21 Haute-Bato 34.5 31 Gbabir 27.4 62 Ngandou 13.4 34 69 105 141 Daho- 177 31 Kodi 52.1 41 Guifa 40.9 61 Ouambé 34.4 52 Mbout 27.3 61 Yambélé 12.9 35 70 6e 106 142 178 63 Obo 51.8 71 Arrondi 40.9 62 Ngbandinga 34.3 62 Ouango 27.2 63 Lili 10.1 71 107 143 179 22 Yénga 40.3 11 Bogangolo 34.2 32 Ouassi 26.8 22 Nadziboro 10.1 Note: a) Prefecture codes and names are below 11 Ombella M'poko 12 Lobaye 21 Mambéré Kadéi 22 Nana Mambéré 23 Sangha Mbaéré 31 Ouham Pende 32 Ouham 41 Kemo 42 Nana-Gribizi 53 43 Ouaka 51 Bamingui-Bangoran 52 Haute-Kotto 53 Vakaga 61 Basse-Kotto 62 Mbomou 63 Haut-Mbomou 71 Bangui 54 Wealth Index The asset wealth index is constructed using principal component analysis (with one rotation) on 11 variables which each indicate how much of a given asset is currently owned. The list of 11 assets consists of assets which are owned by at least 5 percent of the sample. The list includes the following 7 household items (or set of items): beds and mattresses, armchairs and coffee tables, radios, dining table and chairs, mobile phones, coal irons, and televisions, and the following 4 productive assets / means of transport: hoes or dabas, bicycle, motorbike, mill/crusher. Food Consumption Score (FCS) The FCS, designed by the WFP, ranges from 0-112. It combines into a single index the household consumption of 8 different food categories across a 7-day period. The 8 categories are: main staples, pulses, vegetables, fruit, meat/fish, milk, sugar, and oil, and excludes condiments. Generally, the score is categorized into the following bins: 0-21 poor, 21.5-35 borderline, and >35 acceptable. 55