TRANSPORT GLOBAL PRACTICE Comoros Road Sector Public Expenditure Review Toward Improving Sustainability, Resilience, and Safety 1 JUNE 2021 CONTENTS Acknowledgements .......................................................................................................................................................4 Abbreviations and Acronyms ........................................................................................................................................5 Executive Summary .......................................................................................................................................................6 I. Introduction ......................................................................................................................................................... 13 II. Overview of the Road Sector: Current Network and Spending ..................................................................... 17 Existing Road Network............................................................................................................................................ 17 Rural Accessibility .................................................................................................................................................. 17 Current Road Network Condition ......................................................................................................................... 18 Current Road Sector Spending .............................................................................................................................. 20 Road Traffic Growth and Road Safety................................................................................................................... 22 Chapter Summary .................................................................................................................................................. 24 III. Available Resources and Financial Needs..................................................................................................... 25 Institutional Setup.................................................................................................................................................. 25 How Much Is Currently Available?........................................................................................................................ 26 How Much Is Needed—Ideally? ............................................................................................................................ 28 Climate Resilience Requires Even More Resources............................................................................................. 30 How Much Is Needed—A Practical Solution ........................................................................................................ 31 Chapter Summary .................................................................................................................................................. 33 IV. Increasing Fiscal Efficiency and Improving Road Asset Management ......................................................... 34 Enhance Road Asset Management .......................................................................................................................... 34 Increase Sustainability of the Road Fund ................................................................................................................ 37 Improve Efficiency in Budget Execution ................................................................................................................ 39 Increase Efficiency in Road Procurement and Contract Management .................................................................... 40 What About the Remaining Gap? ............................................................................................................................ 42 Chapter Summary .................................................................................................................................................... 43 V. Strategic Prioritization and Potential Interventions ....................................................................................... 44 Wider Socioeconomic Benefits from Transport Connectivity ................................................................................. 44 Rural Access and Poverty Reduction ................................................................................................................... 44 Affordability and Public Transportation .............................................................................................................. 45 Market Access and Inter- and Intra-Island Connectivity ..................................................................................... 47 Agricultural Production ....................................................................................................................................... 49 Climate Resilience ............................................................................................................................................... 50 Road Safety ......................................................................................................................................................... 50 Strategic Prioritization—Multicriteria Analysis ...................................................................................................... 53 Chapter Summary .................................................................................................................................................... 56 VI. Conclusion..................................................................................................................................................... 57 References ................................................................................................................................................................... 59 2 Appendix A. Road Prioritization Results .................................................................................................................... 63 Appendix B. Road Condition Data .............................................................................................................................. 70 Appendix C. RONET Analysis.................................................................................................................................... 73 Road Network Data ................................................................................................................................................. 73 Unit Road Preservation Project Costs ...................................................................................................................... 76 Vehicle Fleet Characteristics ................................................................................................................................... 77 Appendix D. Wider Socioeconomic Benefits from Transport Connectivity ............................................................... 78 Rural Accessibility and Poverty Reduction ............................................................................................................. 78 Affordability and Public Transportation .................................................................................................................. 82 Market Accessibility and Inter- and Intra-Island Connectivity................................................................................ 85 Agricultural Production ........................................................................................................................................... 88 Access to Healthcare Services ................................................................................................................................. 90 Appendix E. Road Safety Measures ............................................................................................................................ 93 Appendix F. Road Safety Risk Assessment ................................................................................................................. 96 Conceptual Framework............................................................................................................................................ 96 Detailed Definition of Parameters ........................................................................................................................... 97 Results ................................................................................................................................................................... 102 3 Acknowledgements The World Bank team led by Atsushi Iimi (Senior Economist, Transport Global Knowledge & Expertise Unit) and composed of Rodrigo Archondo-Callao (Senior Highway Engineer); Andry Rakotoarisoa (Infrastructure Specialist); Soames Job (Lead Transport Specialist); Alain Labeau (Transport Consultant); Antonino Tripodi (Road Safety Consultant); Massoundi Miradji (Transport and Data Consultant); and Desta Wolde Woldeargey, Sitti Fatouma, and Matoiri Ramlat (Program Assistants) prepared the study. Idah Pswarayi-Riddihough (Country Director), Maria Marcela Silva (Practice Manager), Raymond Bourdeaux (Operations Manager), and Boubacar Sidiki Walbani (Residential Representative of Comoros) provided overall guidance. The team acknowledges insightful comments and suggestions provided by the following World Bank colleagues: Fiona Collins (Lead Transport Specialist), Satoshi Ogita (Senior Transport Specialist), Jose Luis Diaz Sanchez (Economist), Kavita Sethi (Senior Transport Economist), and Antoissi Said Ali Said (Country Operation Officer). The team would like to express special thanks to the government of the Union of Comoros, including the Road Fund of the Directorate of Roads, Bridges, and Flood Control, for its close collaboration in data collection and analysis. 4 Abbreviations and Acronyms CMF Crash modification factor DRP Directorate of Regional Planning DRTR Directorate of Roads and Road Transport FER Fonds d’Entretien Routier GDP Gross domestic product IRI International Roughness Index KMF Comorian franc NMCRS National Multisectoral Committee on Road Safety RN National Road RUC Road user charge WHO World Health Organization 5 Executive Summary Background. Comoros has potential for economic significant. growth in various sectors, such as agriculture and There is substantial need for road improvements, but tourism, but growth has been relatively modest. fiscal space is limited. The government has ambitious Approximately 40 percent of the total population was development goals, with the Plan Comores Emergent estimated to live below the national poverty line in calling for massive rehabilitation projects on more 2014. Regional inequality is persistently high and has than 800 km of roads (nearly the entire road network) widened in recent years (Figure ES1). Many rural and construction of 100 km of new roads over the next Comorians do not have good access to market 10 years at an estimate total cost of KMF100 billion opportunities. (UDD250 million), but resources are limited, and The Comorian national markets are highly fragmented implementation capacity must be enhanced to because of poor transport connectivity between and manage the road assets more efficiently. within the islands. For example, banana prices more Comoros is highly vulnerable to climate events such as than triple on their journey from Nioumachoua tropical cyclones, flooding, and landslides, which (Comorian franc (KMF)2,500/50 kg) to Volo Volo affects the transport sector. Resilient infrastructure market in Moroni (KMF8,500/50 kg) partly because of can minimize life cycle costs but is more costly initially. inter- and intra-island transportation costs (Figure Road safety is becoming a matter of increasing ES2). High profit margins for retailers can also be concern in Comoros. As the economy recovers, traffic attributed to the limited supply of agricultural will increase, and without proper measures, road commodities. Price differences within Comoros are crashes will as well. Figure ES1. Regional poverty head count (%) Figure ES2. Cost of bananas, including cost to transport from Nioumachoua to Moroni Source: Haazen et al. 2016. Objectives. Against the above background, the sector; examine available resources and potential objectives of the current report are to examine the financial needs; explore ways to meet the existing fiscal sustainability of road sector spending and gap; and propose a prioritization scheme based on explore a potential intervention framework in the a multicriteria analysis, suggesting a list of priority road sector that will allow Comoros to address its roads that the government could consider in its connectivity challenges while considering the next road investment programs. country’s macro fiscal situation. It will review Current road network. Comoros has a well- current performance and practices in the road 6 established road network of approximately 815 conditions are poor and have deteriorated in km (Figure ES3) with an asset value of some recent years. With the assistance of external USD400 million; 436 km are national roads, 282 development partners, some major national roads km are regional roads, and 58 km are unclassified. were rehabilitated in 2017/18, but Comoros is According to national standards, all classified highly vulnerable to climate events. In 2019, roads are supposed to be paved. The national Cyclone Kenneth devastated approximately 90 km roads carry approximately 80 percent of traffic, of roads (more than 10 percent of the total and traffic volumes on 60 percent of lower-level network). As a result, overall road conditions were roads are very low (fewer than 300 vehicles per deteriorated in recent years (Figure ES4), with half day). Although network coverage is of the road network in poor or very poor comprehensive, connecting most major towns and condition. villages (rural access index 65 percent), road Figure ES3. Road network conditions, 2021 7 Figure ES4. Comoros: Changes in road condition, maintenance projects tend to be sacrificed in favor 2014-2021 of investments. In particular, in recent years, a maintenance backlog has accumulated partly because of the transition between the former road maintenance fund (Fonds d’Entretien Routier) and the new road fund (Fonds Routier), Cyclone Kenneth, and the COVID crisis (Figure ES6). Comoros has a road fund mechanism, with revenues from the fuel levy assigned to the sector, but its sustainability is not guaranteed. Internal resources from the fuel levy have been Sources: EC 2015; World Bank surveys. diminishing, mainly because it is not adjusted for inflation and is delinked from fuel consumption. In Current spending. The government of Comoros real terms, road fund revenues from the fuel levy has been making particular efforts to rehabilitate have fallen 20 percent over the last 8 years (Figure major national roads in recent years. Including ES7). The transfer mechanism is also highly internal and external resources, Comoros spent on discretionary and unpredictable, preventing the average USD5.4 million per year on road sector Fonds Routier and Directorate of Roads and Road development from 2016 to 2019, of which Transportation (DRTR) from planning and approximately 70 percent was for capital implementing their road programs in a timely, investment (Figure ES5). The availability of consistent manner. resources is highly variable from one year to another. Because of this unpredictability, road Figure ES5. Road sector spending according to activity Figure ES6. share of spending according to activity Source: World Bank estimates based on government data. Financial need—Ideal. Under an optimal, far from affordable given the country’s current unconstrained budget scenario that minimizes the fiscal and external balances. With current internal present value of total road user costs and and external available resources, even the do- government intervention costs, USD17.5 million minimum scenario of performing only per year would be needed to bring the network rehabilitation projects when the roads reach very back to its original design standard, reduce the poor condition and losing approximately 10 average roughness of the network, and slightly percent of the asset value over the next 5 years increase the road asset value, but this scenario is cannot be implemented (Figure ES8). 8 Figure ES7. Actual and theoretical fuel levy Figure ES8. Annual road agency costs—ideal scenario Source: Comoros government. These simulation results mean that, in Comoros, Financial need—Practical. One option is to focus it is extremely challenging to improve or even on investing in and maintaining priority roads maintain the existing road network under current while allowing lower design standards for low- road standards. Based on official road standards, volume roads (approximately 400 km). Under this all classified roads are supposed to be paved by “practical� scenario, the resource requirement default, and most roads were paved at some can be contained at USD12.4 million per year for point, but in reality, half have significantly the first 5 years (Figure ES9) and then lowered to deteriorated and are in need of reconstruction USD9.6 million per year in subsequent years, because of lack of maintenance. Given limited achieving a steady state in road sector resources and because some roads are more development. By focusing limited resources on important than others, a more practical approach priority roads with more traffic, average road could be considered to ensure that basic road roughness could be improved from a current 5.3 connectivity is maintained over time. m/km to 4.2 m/km by Year 5 and 3 m/km by Year 10 (Figure ES10). Figure ES9. Annual road agency costs - practical solution Figure ES10. Average road roughness Remaining gap. There remains a substantial measures to partially make up the shortfall. shortfall of approximately USD7 million given How to fill the gap? First, to use available available resources (USD5.4 million per year). It is resources wisely, it is essential to enhance road an ambitious target, but there are feasible policy 9 asset management, including strategic performance-based maintenance contracts and prioritization. community-based maintenance contracts (which can also generate employment opportunities). Second, the new Fonds Routier must be operationalized. The Fonds Routier’s managerial Using all these policy measures, Comoros could and financial autonomy must be confirmed, with a perhaps mobilize an additional USD2.4 million per full transfer of fuel levy collected ensured. In year. In total, Comoros could potentially mobilize addition, the current fixed-amount fuel levy USD7.7 million per year. Still, an unfilled gap of transfer mechanism must be adjusted. To ensure approximately USD5 million may be left. Subject to sustainability of road fund revenue, the fuel levy the country’s fiscal and external balances, support mechanism must be at least inflation proof and from external development partners may be linked to actual fuel consumption. needed to fill such a gap (Figure ES 11). Third, current budget preparation and execution Tradeoffs must also be considered to address in the road sector are inefficient; only important externalities. With a constrained approximately 20 percent of what the government budget scenario, the government must make plans in its medium-term public investment plan informed decisions and consider tradeoffs in has been budgeted annually, whereas terms not only of network standards, but also of approximately 66 percent of the allocated budget building back better to reduce the climate was actually executed. By establishing a more vulnerability of priority roads and to improve professional, autonomous road agency or entity safety. Comoros is highly vulnerable to extreme equipped with financial and implementation climate events. Building in climate resilience could capacity and autonomy, budget execution can be add to investment needs (potentially 30 percent more efficient. additional costs) but could reduce lifecycle costs and increase the net present value of road assets. Fourth, road unit costs are inherently high in As overall road conditions improve, improving Comoros because of limited market competition safety will require additional investments that are and high import prices of materials and not currently being made. (Under the current equipment, but there are opportunities to scenario, at least approximately USD1 million increase efficiency in public procurement and would need to be dedicated to increasing road contract management, for example, through safety.) Figure ES11. Potential internal and external resources 10 Recommendations. To achieve sustainability in prioritize roads serving agricultural road sector spending while maintaining overall production areas and providing road connectivity, the main recommendations are accessibility to social services (e.g., health) as follows: and those where economic risk is high because of extreme climate events • Gradually move toward a steady-state (floods), road safety black spots (areas condition of the road network. The road where crashes are concentrated), and network requires reconstruction, roads with high traffic (e.g., connecting rehabilitation, and periodic maintenance main cities and ports) (appendix A). Next interventions to upgrade its overall road programs could be prepared on the condition and overcome the backlog of basis of such a prioritization framework. deferred maintenance. This requires a greater financial commitment than under • Provide enhanced road asset a steady-state scenario, under which management. Current road programs are network deterioration reaches a relatively not prepared based on actual data and constant level year after year, and future evidence. To use available resources interventions can be anticipated and wisely, it is essential to establish a reasonably taken care of with existing comprehensive road asset management resources. system, although it does not have to be data intensive. As underlying data are • Critically review road standards. To updated, the national road program must ensure long-term sustainability in road be updated based on the road asset sector development, at least USD12.4 management analysis. million would be needed per year for the next 5 years. Under this scenario, • Conduct regular road condition and available resources would need to be traffic surveys. To implement good road focused on priority roads while allowing asset management, it is critical to update lower design standards for low-volume underlying road conditions and traffic roads (approximately 400 km). Half of the data regularly. The necessary budget for network has significantly deteriorated and these surveys should be secured over the is in need of full reconstruction. For low- medium to long term. priority roads with limited traffic, climate- • Use cost-effective data collection tools. resilient but more cost-effective Various new, cost-effective technologies standards, such as low-volume sealing or are available, such as mobile phone– well-maintained gravel roads, could be based software. It is recommended that considered. such tools be used to the maximum • Strategically prioritize roads. Strategic extent. prioritization is critical to meet the • Spend more on maintenance. Resource financing gap. A prioritization framework allocation is inconsistent with optimal based on a multicriteria analysis can allow distribution. If USD12 million is available roads to be prioritized to accommodate per year, more than half of available various development objectives. It is resources should be dedicated to periodic important to reach consensus across all and routine maintenance. Focusing on economic and social sectors (e.g., network maintenance will allow the agriculture, health, education, private government to consolidate a road asset sector development). management strategy that pays adequate • Develop a list of priority roads. A attention to preserving existing assets in multicriteria analysis was conducted to good condition and break the vicious cycle 11 of rehabilitating roads, not doing anything management. Cost efficiencies in delivery for 10 years, and then having to of road assets could be achieved by rehabilitate or reconstruct them again. providing high-powered incentives following results- and user-oriented • Operationalize and make the Fonds approaches by, for instance, introducing Routier autonomous. To restore stability performance-based maintenance and predictability of road fund operations, contracts- and community-based the road fund’s managerial and financial maintenance contracts. autonomy must be reconfirmed. A transparent decision-making process • Mobilize more external resources. Even if must be established to transfer a full all the above policy measures are taken, it amount of dedicated resources. is likely that a financing gap of USD5 million per year would remain. Subject to • Adjust the fuel levy mechanism. The the country’s fiscal and external balances, current fixed-amount transfer mechanism support from external development is not sustainable. To ensure sustainability partners may be needed to fill such a gap, of road fund revenue, the amount although it is essential to implement the transferred it must be linked to fuel above-mentioned institutional reforms to consumption. ensure long-term sustainability of the • Examine the possibility to increase and sector. diversify road user charges. The current • Consider additional tradeoffs to build fuel levy is minimal in Comoros. Although resilience and safety in the road network. they are limited, there are other possible Within a constrained budget scenario, the sources of revenue, such as vehicle license government must make informed and registration fees, axle load taxes, decisions and consider tradeoffs in terms parking fees, and driver’s license fees. not only of road network standards, but • Establish a more professional road also of building back better to reduce the agency or road management unit. climate vulnerability of priority roads and Planning and implementation capacity improve safety. Building climate resilience must be increased. A more professional, and improving safety would add to autonomous road agency or entity investment needs but bring overall equipped with financial and positive economic returns in the long run. implementation capacity should be Because there are no resilient design established, possibly by clarifying the standards available in the country, and it roles of the current DRTR and Directorate is not possible to determine with precision of Regional Planning and transforming what the specific additional costs would them into a more professional, be, a key next step would be to invest in accountable road management unit, development of design specifications and “Cellule de Gestion Routière.� estimate unit costs for various road works. • Improve procurement and contract 12 I. Introduction Comoros has potential for economic growth in Regional inequality is persistently high and has various sectors such as agriculture and tourism, but widened in recent years. Comoros comprises three growth has been relatively modest. Comoros is a major islands, Grand Comore, Moheli, and Anjouan, small island state with a total population of with populations of 379,000, 327,000, and 52,000, approximately 850,000 in 20191 situated off the coast respectively. Moroni, the national capital and largest of Mozambique in the Indian Ocean. In the last decade city, is located on the western coast of Grande Comore (before the COVID crisis), the economy grew at a and has approximately 151,000 inhabitants, or relatively modest average rate of approximately 3 approximately 16 percent of the total population. The percent (c.f., 4.5 percent in Sub-Saharan Africa). other two islands have the regional capitals, Although the service sector is growing rapidly, other Mutsamudu and Fomboni. Moheli is the smallest and sectors remain stagnant, leaving many rural poorest island. Moroni has a poverty rate of 30.7 Comorians behind (Figure 1). Approximately 70 percent and Moheli, 53.2 percent (Figure 2). Such percent of the total population lives in remote rural regional inequality seems to have widened in recent areas, and approximately 57 percent of the labor force years. The Gini index, measured according to is engaged in agricultural production, which is mostly household consumption, was estimated at 45 in 2014, subsistence farming. Many farmers are disconnected which is higher than in Comoros’ structural peers from market opportunities. It was estimated that (32.9) and is considered to have been increased for the approximately 40 percent of the total population lived last decade if no redistributional effect is taken into below the national poverty line in 2014. account (World Bank 2020a). Figure 1. Gross domestic product growth (%) Figure 2. Regional poverty (%) Source: WDI. Source: Haazen et al. 2016. Transport connectivity between and within the from those production areas (Figure 3). Comoros has islands is poor, leaving the Comorian national market attractive tourist sites, such as Nioumachoua Beach in highly fragmented. The widening regional inequality is Moheli, but they are unreachable for international partly attributed to this. For example, Moheli and tourists because domestic air and maritime transport Anjouan have significant agricultural potential, but services are unreliable, and the condition of the road Moroni, the major consumption area in Comoros, is far network is poor. 1 According to World Development Indicators.\ 13 Figure 3. Major transport networks and major towns The COVID crisis revealed the importance of April 2019, directly affecting more than 345,000 maintaining transport connectivity between and people. Total damages and losses were estimated at within the islands. There are significant differences in USD185 million plus an estimated USD277 million commodity prices between areas because of required for recovery and reconstruction, including 62 inefficient and unreliable or even unavailable intra- km of primary roads, 16 km of regional roads, and 12 and inter-island transport services. For instance, km of rural roads, as well as Port Fomboni. The poorly bananas were two and three times as expensive in maintained road network increases the country’s Moroni as in Mutsamudu and Fomboni, respectively. vulnerability to extreme climate events. The unreliable Similarly, tomatoes were 60 to 80 percent more transport network hampers smooth evacuation and expensive. In 2020, the government banned rapid recovery from disasters, limiting access to basic interisland travel because of the COVID pandemic, and products and services such as food, water, and medical the price differentials widened, except for bananas supplies. The Comorian people struggle with poor road (Figure 4). Transport connectivity is critical for accessibility every day (Figure 5). Resilient efficient market transactions and to ensure food infrastructure is more costly at the initial investment security across the country. stage but can minimize life cycle costs by avoiding potential reconstruction costs in subsequent years. Comoros is highly vulnerable to climate events such as tropical cyclones, flooding, and landslides, which adds challenges for transport. Cyclone Kenneth hit in 14 Figure 4. Price differences between the islands Figure 5. “Anjouan: Les routes de l'impossible!� Source: World Bank market survey. Source: Habarizacomores.com, news article on April 5, 2021. Fiscal space is limited in Comoros. As shown below, wave of COVID-19 in Comoros in 2020, mobility unmet financial need for universal access is significant. restrictions and suspension of international travel The government often has ambitious development disrupted many economic activities, resulting in a goals. For instance, the goal of the latest national sharp drop in tourism receipts. Demand for development plan, Plan Comores Emergent, is agricultural products and services (leisure and structural transformation of the economy to transport) was hit the hardest. With strong support accelerate agricultural growth and the tourism and from the diaspora and sustained government health fishing industries. In the road sector, the plan calls for and social spending, the limited spread of the massive rehabilitation projects on 800 km of roads and pandemic led to a relatively small contraction of new construction of 100 km of roads over the next 10 economic growth (–0.5 percent). In early 2021, the years, which is estimated will cost approximately country had a new, stronger COVID-19 wave, but a KMF100 billion (USD250 million) in total (Comoros modest recovery is still expected. Depending on the 2019). speed of vaccination rollout, recovery may start during the second half of 2021, resulting in an expected In reality, however, available financial resources are annual growth rate of 0.2 percent. The economy is far from sufficient, and implementation capacity in the expected to continue recovering in 2022 and 2023, public sector is limited. The road sector has long had depending on the pace of vaccination in Comoros and an insufficient, unpredictable budget allocation. A the rest of the world, although there are significant road maintenance backlog has accumulated, downside risks to the outlook, including a prolonged preventing efficient movement of people and goods. pandemic and growing vulnerabilities in the financial The country’s fiscal balance is negative and tight sector that could further decrease debt sustainability because of persistently low fiscal revenue and large and the financial health of state-owned enterprises. public sector activities, including state-owned Increased government social and health spending enterprises. In addition, debt sustainability has would mitigate the impact of the pandemic but deteriorated in recent years.2 It is a challenge to meet continue to increase debt levels. the huge financial need with limited available resources. Inclusive, sustainable growth is called for more than ever. As the economy recovers, it is of particular The COVID crisis may exacerbate Comoros’ macro- importance to support local business development fiscal challenges. Despite a relatively contained first 2 Since 2016, public and publicly guaranteed external debt analysis under preparation largely expected to place has significantly increased, with a World Comoros in the category of high risk of debt distress (from Bank/International Monetary Fund debt sustainability the current moderate risk). 15 and create jobs. Comoros has a predominantly young government statistics, micro surveys administered in population that is rapidly growing (~3 percent per the country, and spatial open-source data. Some data year). Given the global economic slowdown and are not up to date or were not available. To fill the gap, restricted economic activities, it is becoming even some assumptions and approximations were made more important to integrate the domestic market and based on global and regional experience. In addition, stimulate domestic demand. To this end, establishing the World Bank conducted some field surveys, such as a reliable, efficient transport system between and road condition and road safety risk assessments, using within the islands is critical. simplified methods that were allowed even under COVID restrictions. The quality of data may not be Road safety is becoming a matter of increasing perfect but is reliable. The analysis mainly used data concern in Comoros. As the economy recovers, traffic from 2016 to 2020. This report does not capture the will increase. Without proper measures, road crashes potential impacts of COVID. will also increase. In Comoros, the local newspaper reports road traffic crashes almost every day. Road safety is not only a fundamental social good, but also a necessity for sustainable economic growth. Crashes could drive families into poverty with the permanent or temporary loss of a family income earner (death or disability). Crash trauma is also likely to occur for those with fewer financial resources (and thus less access to safe transport and likely to be living in poorer areas with less safe roads). Road crash costs can be a burden on a growing economy. In Comoros, the costs of road crash deaths and injuries are estimated to be 8.8 percent of gross domestic product (GDP) (World Bank 2019a). Objectives. The goal of the current report is to examine fiscal sustainability in road sector spending and explore a potential intervention framework in the road sector including key institutional reforms and priority investment roads that will allow Comoros to address its connectivity challenges while remaining cognizant of the country’s macro fiscal situation. To this end, the report is organized as follows: Chapter II provides an overview of current road sector performance, including road network conditions and government spending. Chapter III examines available resources and potential financial needs for road sector investment and maintenance. Chapter IV discusses challenges and opportunities to fill the gap. Based on a proposed financing framework, Chapter V presents a prioritization scheme based on a multicriteria analysis suggesting a list of priority roads that the government could consider in its next road program. Limitations of the report. In Comoros, it is a challenge to collect complete, consistent data. The report uses a variety of global and country data, including official 16 II. Overview of the Road Sector: Current Network and Spending Existing Road Network regional average in Africa (Figure 7). In general, small countries tend to have high road density because The Comorian road network is well established, many people live in small areas (high population covering the whole nation and connecting most major density). Of small island countries, Comoros does not cities and towns (Figure 6). The network covers 815 km compare favorably with neighboring countries over the three main islands, including national (404 (Seychelles (110 km per 100 km2 of land), Mauritius km), regional (296 km), urban (54 km), and unclassified (102 km per 100 km2 of land)) but has greater road (61 km) roads. This translates into a road density of density than other island states such as Cabo Verde 43.1 km per 100 km2 of land, which is well above the (33 km per 100 km2 of land). Figure 6. Road network and populated areas Figure 7. Road density and population density Rural Accessibility Approximately 65 percent of the rural population has good access to the road network, which is among the highest in the region. The rural access index, used to measure people’s accessibility to the road network, is one of the most important global indicators in the transport sector. The index is determined according to the proportion of the rural population with access to an all-season road within an approximate walking distance of 2 km or walking time of 25 minutes. Using the latest road condition data, from 2021, the rural Source: WDI. access index is estimated to be 65.3 percent in 17 Comoros, which means that approximately 240,000 relatively good accessibility to the road network rural residents do not have access to a good road The Sustainable Development Goals envisage network. The index is among the highest in Sub- universal access to the road network by 2030. Rural Saharan Africa. Because of the high population density connectivity is among the most important challenges and high road density, small countries tend to have in many African countries. Sustainable Development greater rural accessibility. For example, the index is 55 Goal Target 9.1 is to “develop quality, reliable, percent in Rwanda and 42 percent in Liberia (Figure 8). sustainable and resilient infrastructure…to support In Comoros, accessibility is even greater because most economic development and human well-being, with a populated villages are located along main national focus on affordable and equitable access for all� roads, which are relatively well maintained. This is (World Bank 2016). consistent with the above-mentioned well-established network, supporting the fact that Comorians have Figure 8. Rural Access Index in Sub-Saharan Africa Source: World Bank 2016. Roughness Index (IRI) ,was calculated at a 100-meter Current Road Network Condition resolution (Figure 9). Approximately 21 percent of In Comoros, half of the road network is in poor or roads are in good condition, 30 percent in fair very poor condition. In 2021, 594 km (approximately condition, and 49 percent in poor or very poor three-quarters of classified roads) was surveyed using condition. This is worse than in other countries in the a mobile phone-based application, RoadLab (see region. For instance, the share of roads in good technical details in appendix B). 3 For 449 km, a condition is 77 percent in Mauritius, 53 percent in conventional roughness indicator, the International Eswatini, and 26 percent in Lesotho (Figure 10). 3 Because of the nationwide lockdown due to COVID, local consultants conducted a mobile phone-based road condition survey on the individual islands. 18 Figure 9. Road network condition, 2021 such as RN1 and RN2 were rehabilitated, and average Figure 10. Share of roads according to condition, road condition improved, but in April 2019, Cyclone selected countries Kenneth hit Comoros and devastated the road network. Approximately 90 km of roads (more than 10 percent of the total network) was damaged, and the share of roads in poor and very poor condition increased to 49 percent.4 The latest sharp deterioration in road network quality suggests the importance of climate resilience and proper, timely maintenance of the network. It is believed that the poor road condition amplified the damage from the cyclone. As will be seen below, in recent years, maintenance has been postponed. Timely road maintenance is important in protecting Sources: Gwilliam 2011; Iimi and Rao 2018; World the road network from extreme climate events. Bank 2016; 2018a. Although national roads are relatively well The current road condition may compare favorably maintained, rural roads are in very poor condition. with the situation in the early 2010s but is worse than National roads account for half the total network, in 2018. Approximately 57 of classified roads were carrying the vast majority of traffic. Recent road assessed to be in poor or very poor condition in 2014 programs have focused on these primary roads, but 41 (EC 2015) (Figure 11). With the assistance of external percent are in poor to very poor condition. Regional development partners, some major national roads and rural roads are much worse, with more than 60 4 Underlying road condition data were collected based on different methodologies, so the data must be compared with caution. 19 percent in poor or very poor condition, although there capitals (Moroni, Mutsamudu, Fomboni) are generally is less traffic (Figure 12). relatively good, but in remote areas, the road network has deteriorated, accumulating a backlog of road Some rural areas are in worse shape than others. On maintenance and rehabilitation projects. each island, road conditions around the regional Figure 11. Changes in road condition, 2014-2021 Figure 12. Road condition according to road class Sources: EC 2015; World Bank surveys. 2016 to 2020 (Table 1).6 On average, the government Current Road Sector Spending has been spending approximately 70 percent of total resources on capital investment. In 2016 and 2017, the Since the early 2010s, Comoros has been working to government spent USD3 million to USD5 million on rehabilitate the major national roads on each island. road maintenance (Figure 14), which accounted for With the assistance of the international community approximately half of total road sector spending, but (e.g., European Union, Arab Bank for Economic such maintenance spending stopped in 2018 and 2019 Development in Africa), several major road because the road maintenance fund mechanism was rehabilitation and resurfacing projects have been suspended while the former road maintenance fund completed or are in process, for instance, RN1 from (Fonds d’Entretien Routier (FER)) was transformed Moroni to Moroni Airport, RN2 from Itsinkoudi to into a new road fund (Fonds Routier). From 2016 to Mtsamoadjou, RN21 from Mutsamudou to Sima, and 2019, on average, only 27 percent of road sector RN31 from Fomboni to Miringoni (Figure 13).5 spending was on maintenance (Figure 15), which Maintenance expenditures tend to be postponed. raises significant concern about deferred maintenance Comoros spent on average KMF2,400 million (USD5.4 and likely degradation of the road network. million) per year on road sector development from 5 6 The World Bank is also supporting rehabilitation of 37 km As in other developing countries in the region, Comoros of feeder roads under the Integrated Development and does not have complete, reliable fiscal data on the road Competitiveness Project (P164584) and 26 km of national sector. These figures are the Bank team’s estimates based roads under the Post Kenneth Recovery and Resilience on available government data. Project (P171361). 20 Figure 13. Recent and ongoing road projects Table 1. Government road sector spending 2016 2017 2018 2019 2020 Avg. 2016 2017 2018 2019 2020 Avg. KMF million USD million Road investment 1,031 1,826 423 2,337 1,404 2.32 4.19 1.02 5.32 3.21 Maintenance 1,229 2,354 0 72 325 796 2.76 5.40 0.00 0.16 0.67 1.80 Road fund operating costs 97 57 0 50 79 57 0.22 0.13 0.00 0.11 0.16 0.13 Total spending 2,357 4,236 423 2,459 2,369 5.30 9.73 1.02 5.60 5.41 % of gross domestic product 0.52 0.90 0.09 0.48 0.50 0.52 0.90 0.09 0.48 0.50 Source: World Bank estimates based on government data. Figure 14. Road sector spending according to activity Figure 15. Share of spending according to activity Source: World Bank estimates based on government data. 21 Although country size affects spending levels, Figure 16. Road sector spending in selected African Comoros’ spending on the road sector (USD5.4 countries million; 0.5 percent of GDP) is among the lowest in Sub-Saharan Africa. Given the country’s network size, this level of spending represents approximately USD7,000 per km of road, which would allow routine maintenance to be financed but not much more. Available fiscal data on the road sector are fragmented in the developing world, but current spending in Comoros is low in the region. Sub-Saharan African countries spend on average approximately 1 percent of GDP on roads, whereas potential needs are estimated at 2 percent of GDP (Bricenio-Garmendia et al. 2008). Even small countries such as Rwanda spend Source: Based on Briceno-Garmendia et al. 2008; approximately 1 percent of GDP on road sector World Bank 2018a; 2019b; 2021. development. Larger countries spend more (e.g., Mozambique, 1.8 percent of GDP; Tanzania, 1.2 Road Traffic Growth and Road Safety percent of GDP) (Figure 16). In Comoros, traffic is generally light but has been In Comoros, road sector spending has been highly rapidly increasing. The number of registered vehicles variable over time, mainly because of unexpected in Comoros (44 per 1,000 population) is lower than in delays of donor-funded projects. Annual road sector other small island countries in the region (e.g., 122 in spending varies significantly, ranging from KMF400 Cabo Verde; 168 in Sao Tome and Principe) (Figure 17), million (USD1 million) to KMF2,900 million (USD9.7 but as the economy and household income grow, million), revealing the country’s challenges in public ownership and use of passenger cars are rapidly finance management to stabilize road expenditure increasing. In Grande Comore, where vehicle and implement road programs in a timely manner. registration data are available, the number vehicles increased by 70 percent during the last decade, complicating challenges for road asset management (Figure 18). Figure 17. Registered vehicles and per capita GDP Figure 18. Number of registered vehicles in Grande Comore Sources: WHO; WDI. Source: Direction Général des Infrastructures, Ngazidja The majority of traffic is concentrated on several fewer than 300 vehicles per day. Based on 2013 traffic major national roads, with many other roads carrying count data, current and future traffic were estimated 22 based on conservative assumptions of annual traffic shows the network distribution according to annual growth rate of 4 percent per year, considering that average vehicles per day. Ninety percent of the road gasoline and diesel consumption increased on average network carries fewer than 1,000 vehicles per day, and by 6.2 percent per year from 2016 to 2020. Table 2 60 percent carries fewer than 300. Table 2. Network length according to road type and traffic Road type Annual average vehicles per day Total <300 300–1,000 1,000–3,000 >3,000 Road network length (km) National 132.8 199.7 61.9 9.6 404.0 Regional 263.6 24.8 7.6 0.0 296.0 Unclassified 61.0 0.0 0.0 0.0 61.0 Total 457.4 224.5 69.6 9.6 761.0 Share of roads (%) National 32.9 49.4 15.3 2.4 100.0 Regional 89.1 8.4 2.6 0.0 100.0 Unclassified 100.0 0.0 0.0 0.0 100.0 Total 60.1 29.5 9.1 1.3 100.0 Road investment and maintenance programs can and Source: EC 2015. should be selective, depending on current and As traffic increases, road safety has been emerging as expected future traffic. On each island, road traffic is a matter of particular concern. The number of road only at the substantial level on a few key national crashes rapidly increased, from 972 in 2015 to 1,670 in roads connecting the regional capitals (Moroni, 2016, particularly in urban areas such as Moroni and Mutsamudu, Fomboni) with major towns. The latest Fomboni, where road traffic is increasing rapidly available traffic count data in Comoros are for 2013. (World Bank 2019c). Road crash data that the The distribution of traffic has not changed much. Most Comoros police provided to the World Health traffic is on RN1, RN21, and RN31 and in major urban Organization (WHO) in 2020 indicated that areas (Figure 19). On other regional and unclassified approximately 23 people were killed on Comoros roads, traffic is generally very low—fewer than 100 roads in 2019. Given the probable high level of vehicles per day—so it may be that different roads can underreporting, WHO statistical modeling suggests be treated differently based on different road that the true number of road deaths is seven times as standards. high. Over the last 10 years, the number of road Figure 19. Annual average road traffic deaths has increased in pace with population growth (Figure 20). These numbers are slightly higher than for Mozambique, a low-income country, but much lower than for Cape Verde, a comparable lower-middle- income country (Figure 21). In addition to the high loss of life in road crashes, many people are disabled and seriously injured each year in motor vehicle crashes. In 2016, there were an estimated 41 road deaths per 10,000 vehicles—much higher than the two to five in developed countries. Comoros does not spend any meaningful resources on road safety. Although the National Multi-sectoral Committee on Road Safety exists, it is not funded in the national budget. The African Road Safety Action Plan 2011-2020 that the heads of state adopted in 23 2012 requests that all African countries allocate 10 Safe System Approach (e.g., OECD 2008; World Bank percent of road investment and 5 percent of road 2020b), holistic road safety measures must be taken to maintenance resources to road safety (Small and Runji address all aspects of road safety, including road 2014). If this formula had been applied, Comoros design, vehicle standards, traffic speed, and road user should have allocated at least KMF200 million behavior. (USD500,000) per year for road safety. Following the Figure 20. Road crashes, 2009-2019 Figure 21. Road traffic deaths in selected countries Source: WHO. Source: WHO. network from extreme climate events. Chapter Summary • The government has been making particular efforts to rehabilitate major national roads Key takeaways from this chapter are as follows. (e.g., RN1, RN21, RN31), but maintenance • The Comorian road network is well expenditures tend to be left out. On average, established, connecting most major cities and only 30 percent of total spending in the sector towns in the country. is used for maintenance. • Rural accessibility is also high (65 percent) • Current road spending in Comoros is because of high population density along the approximately USD5.4 million, or 0.5 percent main national roads, which are relatively well of GDP, among the lowest in the region, and is maintained. highly variable over time, making it difficult to • Half of the road network is in poor or very implement a timely, consistent road program. poor condition, particularly in remote areas, • Road traffic is low but rapidly growing, primarily because of lack of road especially along the major national roads, maintenance. Although national roads are which should be prioritized for investment relatively well maintained, regional and rural and maintenance. roads are in poor condition. • As traffic increases, road crashes also increase, • Average road condition improved between calling for holistic road safety measures. 2014 and 2018 thanks to some rehabilitation projects on national roads supported by donors, although it has deteriorated again because of the devastating damage from Cyclone Kenneth. • Climate resilience and proper, timely maintenance are essential to protect the road 24 III. Available Resources and Financial Needs Before discussing potential policy measures to managed the planning and supervision of road improve efficiency and sustainability of road sector projects in Moheli and Anjouan. spending, this chapter describes how the government DRTR is a key player in the management of large currently manages the road network, how much is investment projects, including extensive currently mobilized in the sector, and how much rehabilitation, upgrades, and new construction would be needed “ideally� and “practically� to support funded by the government and external road sector development. development partners. It is responsible for planning, Institutional Setup programming, procuring civil projects and studies, and managing contracts. DRTR plays a key role in the daily Comoro’ road assets are administered by the operations of road asset management. The national Ministry of Land Planning and Urbanization, in charge roads account for approximately 55 percent of the of Land Affairs and Land Transport (Ministère de network and carry more than 80 percent of traffic in l’Aménagement du territoire, de l’Urbanisme, chargé the country. des affaires foncières et des Transports terrestres), Road maintenance and limited road investment administers Comoros’ road assets, under which the projects (e.g., minor rehabilitation and resurfacing) Directorate of Roads and Road Transportation (DRTR) are financed from a road fund, which is in theory and the Directorate of Regional Planning (DRP) jointly responsible for funding road maintenance, and DRTR manage the road network (Figure 22). Before 2016, is responsible for implementing them. the regional road departments of the governorates Figure 22. Road sector organizational chart Grande Comore Anjouan Moheli ▪ National roads ▪ National roads ▪ National roads ▪ Regional roads ▪ Regional roads ▪ Regional roads ▪ Urban roads ▪ Urban roads ▪ Urban roads Regional Directorate of Planning Ministry of Land Planning, Urbanization and Land Transportation Transferred in 2016 Regional Directorate of Roads Directorate ▪ Planning ▪ Implementation ▪ Funding of Planning PIU Donors ▪ Coordination Investment ▪ Planning Government Maintenance Road Fund Comoros has had a road financing mechanism road maintenance fund, le Fonds d’Entretien Routier through a road fund since 2012. A second-generation (FER), was created under the Ministry of Regional 25 Planning in 2012. The second-generation road funds Figure 23. Comoros road sector resources common in many Sub-Saharan African countries inspired the FER statutes and gave the FER managerial and financial autonomy. The main responsibility of the FER was to fund the road maintenance program from a RUC embedded in the road fuel excise tax. In 2019, the FER was restructured into the Fonds Routier, a third-generation road fund, to finance road investments and maintenance, but it has not been fully operationalized. The new statutes open the door to new sources of road maintenance funding. Source: Comoros government. However, the Fonds Routier is not financially autonomous, despite what its statutes claim, and it must rely on late, irregular payments from the For internal resources, Comoros depends largely on a Treasury. It is authorized to collect investment funds RUC levied on fuel sales. When Comoros introduced a and finance investment projects through a dedicated fuel levy mechanism, the original rates were KMF37.5 window, so under the new framework, the budget and per liter for gasoline and KMF17.5 per liter for diesel. development partner lending programs jointly fund In 2013, a fixed-amount weekly transfer scheme road investment, and DRTR and DRP manage it, replaced the conventional RUC levied on fuel sales and potentially by creating project implementation units embedded in the excise tax (15 percent). Under this (PIU). scheme, KMF25 million would be transferred to the Fonds Routier every week, (KMF1,304 million, or How Much Is Currently Available? USD3.4 million at the exchange rate at that time). This accounted for approximately 0.34 percent of GDP. Comoros relies heavily on external resources for road Since then, the underlying fuel tax rate or fixed sector funding. From 2016 to 2019, Comoros spent on transfer amount has not changed. average KMF2,400 million (USD5.4 million) per year on the road sector, of which external development The current fixed-amount allowance scheme has partners financed approximately 60 percent. External several flaws, decreasing resources available to the donors normally finance large road improvements Fonds Routier over time. The underlying fuel tax rates (e.g., rehabilitation and reconstruction road projects), per liter have not been adjusted for inflation for a long and the government funds routine and periodic time. At present value, the fixed amount of KMF25 maintenance through the road fund. The external million is only approximately 0.25 percent of GDP, dependency rate varies substantially from year to year more than 20 percent lower than the initial level in (Figure 23). On average, the international donor 2013 (Figure 24). The tax formula should be adjusted community has financed 62 percent. External at least for inflation. Otherwise, there is no financial dependency has increased markedly in recent years sustainability. In addition, the current fuel levy because of the malfunctioning internal road fund transfer mechanism is independent of fuel mechanism. consumption, which has increased by approximately 30 percent over the last 10 years (Figure 25). Based on current gasoline and diesel consumption, in theory, the fuel levy transfer should be approximately KMF1,744 million (USD3.6 million). 26 Figure 24. Actual and theoretical fuel levy Figure 25. Fuel consumption Source: Comoros government. Source: Comoros government. In addition, the Treasury often exercises countries in the region but is far below the regional discretionary power to release only a fraction of the average in Sub-Saharan Africa. In general, the ability amount designated to the Fonds Routier. According to collect a fuel levy may be limited in small countries. to the decree on the creation, organization, and Although Guinea-Bissau and the Gambia (countries for operations of the Fonds Routier (No. 19-025/PR, which recent road fund data are available), are two to January 31, 2019), a weekly allocation of KMF25 three times as large as Comoros in terms of million to the Fonds Routier is the sole source of funds population, they obtain 0.15 percent and 0.13 percent for road maintenance. This allocation is supposed to of GDP from fuel levies, respectively (Figure 26). In the be drawn from the excise tax levied on fuel sales for region, road funds receive average revenue of vehicles, but only approximately 70 percent of this approximately 0.5 percent of GDP (Briceno- allocation on average was released in recent years. Garmendia et al. 2008). In Mozambique, the road fund Therefore, actual resources allocated to the Fonds receives 1.4 percent of GDP from internal resources. Routier have been approximately KMF900 million The size of countries matters, but even with its limited (USD1.9 million) per year, which is approximately 0.18 road network (~800 km total) taken into account, percent of current GDP. available resources in Comoros are too low. The Fonds Routier receives only USD2,400 per km of road (Figure The current level of road fund resources in Comoros 27). This does not allow for even routine maintenance may compare favorably with that of other small for half of the existing road network. Figure 26. Road fund resources per gross domestic product Figure 27. Road fund resources per kilometer of road Source: World Bank estimates for Comoros; World Bank (2021) for Mozambique; World Bank (2019) for the rest of countries. 27 Table 3. Network length according to type of road How Much Is Needed—Ideally? Type of road Length km (%) To assess how much is needed to maintain the National 404.0 (53.1) current road network in Comoros, an Excel-based Regional 296.0 (38.9) model, Road Network Evaluation Tools (RONET), is Unclassified 61.0 (8.0) used. RONET allows road maintenance and Total 761.0 100.0) rehabilitation policies to be assessed and determines proper allocation of expenditures among various Road investment costs in Comoros are high because activities, such as recurrent maintenance, periodic many materials and equipment must be imported. In maintenance, and rehabilitation projects (see addition, market competition tends to be limited technical details in appendix C). The evaluation is because of the size of the construction market in the based on Comoros’ road network traffic data from country. Unit costs of road projects are calculated 2013 and road condition data (roughness) collected in based on recent road projects in the country. 7 2020/21 using a mobile phone-based application, Although capital unit costs vary from USD150,000 to RoadLab, developed by the World Bank. Of 815 km of USD600,000 per km according to road network type classified roads, 761 km have been analyzed and and applied standards, routine maintenance costs are divided into national, regional, and unclassified roads assumed to be constant at USD6,000 per km per year (Table 3). Because of the lack of data, approximately (Table 4). Capital projects on a particular road are a 54 km of urban roads were excluded from this analysis. function of road condition. According to the RONET Under current road standards, all roads are supposed logic, roads in good or fair condition (sustainable to be paved in Comoros. network) require periodic maintenance, and roads in poor or very poor condition require rehabilitation. Table 4. Unit costs of road projects: Paved roads Unit cost Road condition Road work class Road work type (USD/km) Good Periodic maintenance Preventive 15,000 Fair Resurfacing (overlay) 150,000 Poor Rehabilitation Strengthening (overlay) 400,000 Very poor Reconstruction 600,000 Routine maintenance (per year) 6,000 Three scenarios are examined: The optimal scenario, with unconstrained budget, minimizes the present Table 5. Evaluated scenarios value of total transport cost (sum of road agency costs Scenario Description plus road user costs) of the network at a 6 percent Optimal Optimal solution without budget discount rate; the keep-condition scenario maintains constraints (maximize net present the average roughness of the road network as the value) current average network roughness, weighted per Keep condition Maintenance of average road roughness at current level vehicle-km, for the next 5 years; and the do-minimum Do minimum Routine maintenance until roads scenario funds rehabilitation projects only when the reach very poor condition roads reach very poor condition (Table 5). 7 Based on government data, “Tableau récapitulatif des Routier – Projet de Réhabilitation de la RN2 (Moroni- travaux routiers réalisés de 2016 à 2020� and the Foumbouni) : Trancon Moroni Ouroveni.� document “Projet : Projet de Réhabilitation du Réseau 28 Based on current official road standards, it is resources are in theory USD3.6 million, of which only estimated that USD17.5 million per year would be USD2.7 million has been allocated to the road sector. needed to maintain the existing road network under To maintain road conditions at current levels (IRI 5.3 the optimal scenario (Table 6). This is well above the m/km), USD15.6 million per year would be needed. resources currently available and spent on the road Even under the do-minimum scenario, USD11.7 sector. In recent years, as discussed above, the million would be required per year, which is still twice country has spent on average USD5.4 million per year, as much as currently available in Comoros. including external resources. Available internal Table 6. Annual road agency costs: Without budget constraint (USD million per year) Periodic Routine Scenario Rehabilitation maintenance maintenance Total Years 1 to 5 Optimal 8.9 4.0 4.6 17.5 Keep condition 8.9 2.2 4.6 15.6 Do minimum 7.2 0.0 4.6 11.7 Years 6 to 20 Optimal 12.6 2.5 4.6 19.6 Keep condition 14.8 1.4 4.6 20.7 Do minimum 16.6 0.0 4.6 21.2 Regardless of scenario, maintenance plays a key role Under the optimal scenario, the roughness would in preserving road asset value. Under the optimal decrease slightly, to 4.8 m/km, and under the do- scenario, approximately half of total spending should minimum scenario, the roughness would increase be for periodic and routine maintenance. The sharply, to 6.4 m/km, possibly undermining transport government’s current road program is biased toward connectivity in the country. This do-minimum scenario capital investment. Even under the do-minimum may represent past experience in Comoros: Because scenario, 40 percent should be spent on routine of lack of resources and maintenance, road conditions maintenance. Under the optimal scenario, the net have deteriorated. Given severe damage to roads, present value of the asset could be increased from the extra reconstruction spending is required, putting current USD401 million to USD414 million by year 5 fiscal pressure on the budget, which is already tight. (Figure 28), but under the do-minimum scenario, the Because of this, it is extremely challenging to improve asset value would decline to USD361 million; or even maintain the existing road network under approximately 10 percent of the asset value would be current road standards. Based on official road lost in the next 5 years. standards, all classified roads are supposed to be Under the optimal scenario, road conditions could paved, and most roads were paved at some point, but improve gradually, but under the do-minimum half of roads have significantly deteriorated and are in scenario, they would deteriorate rapidly. RONET need of reconstruction because of lack of estimates the condition of the network in year 5 as a maintenance. On some roads, surface layers and function of estimated road deterioration and road structures have been completely damaged and lost. As projects executed over that period. The average discussed below, some fiscal resources could be saved roughness weighted per vehicle-km takes into by improving public finance management and road consideration the length of the roads and the number contract management, but given such a huge resource of road users; thus, roads with higher traffic have more gap, it is almost impossible to maintain the road weight than roads with lower traffic. The current network unless external partners provide substantial average network roughness measured using the IRI, resources and Comoros generates more internal weighted per vehicle-km, is 5.3 m/km (Figure 29). resources. 29 Figure 28. Road asset value according to scenario Figure 29. Average road roughness according to scenario Climate Resilience Requires Even More for a detailed discussion.) In April 2019, Cyclone Kenneth devastated the country, damaging 62 km of Resources primary roads, 16 km of regional roads, and 12 km of Comoros, situated in the western Indian Ocean, rural roads, as well as Port Fomboni. Landslide is where tropical cyclones occur every year, is highly another climate-related risk in Comoros. Only 20 vulnerable to extreme climate events. There are percent of the road network is in good condition, and many flood-prone areas in the road network 30 percent is in fair condition. This poorly maintained (Figure 30). Some roads are exposed to higher transport network exacerbates the country’s economic risk than others. (See the following chapter vulnerability to extreme climate events. Figure 30. Flood-prone areas and economic risk of extreme climate events Making the road network climate resilient would of roads by avoiding unnecessary repair and require additional investment during the initial stage reconstruction projects. A global study estimated that of road construction but could reduce lifecycle costs incorporating resilience into construction of a road 30 network would increase costs by 3 percent to 23 theory estimated at USD17 million. Under the do- percent (World Bank 2017a). In Comoros, building minimum scenario, expenditures could be kept under resilient roads, including, for example, bridges and USD12 million, but the roughness of the network, and culverts, could increase costs by 30 percent. therefore overall transport connectivity, would be likely to deteriorate significantly. Based on RONET analysis with higher estimated unit costs, incorporating climate resilience would require One practical solution is to focus on investing in and USD21.3 million under the optimal scenario, even maintaining priority roads while allowing lower more than under the baseline scenario. More design standards for other roads. This may be a good rehabilitation projects would be required to build compromise because traffic on feeder roads is limited. resilient infrastructure (Figure 31). Doing so could In addition, even though all roads are classified as increase the net present value of the road asset to paved, half have been significantly damaged and are USD538 million at year 5 under the optimal scenario. in need of full reconstruction. Thus, practically Predicted roughness would be the same as the speaking, this reclassification does not have any real baseline case under the optimal scenario because the impact but can help maintain critical connectivity and benefits of the climate resilience projects would not improve long-term fiscal sustainability in the road materialize immediately after the rehabilitation sector under a budget-constrained scenario. projects but over the medium to long term. There are Under this selective approach, Comoros would need no resilient design standards being applied in the to mobilize USD12.4 million from internal and country, and it is not possible to determine with external resources, much less than the above precision what the specific additional costs would be. scenarios. In this selective solution, RONET considers Investing in development of design specifications the costs needed to reconstruct or rehabilitate 347 km should be considered, and unit costs for different work to paved standards and maintain them over the next items should be estimated. 20 years and to rehabilitate and maintain 413 km of Figure 31. Annual road agency costs: Baseline and roads that are currently in fair, poor, or very poor climate-resilient cases condition and have fewer than 30 vehicles per day at lower standards (mainly gravel) and lower costs (Table 7). Because neither climate-resilient design nor cost estimates are currently available, the current analysis is based on the existing (nonresilient) standards and costs. Further engineering investigation will be needed to determine what standards would be most appropriate for low-volume roads in the Comorian context and to take into account resilient design considerations for the network. Under this practical approach and optimal scenario, USD12.4 million would be required (Figure 32), which is still above the current level of spending but potentially achievable with enhanced external assistance. Under How Much Is Needed—A Practical Solution this scenario, average road roughness—including paved and gravel roads—could be improved to an IRI As mentioned above, Comoros must meet financial of 4.2 at year 5 (better than the baseline model) needs in the road sector with limited available (Figure 33) because the proposed solution focuses on resources while sustaining transport connectivity. priority roads with high traffic. Thus, even though Approximately USD5.4 million per year is currently design standards are compromised for some roads, available, including external assistance. The amount overall connectivity could be increased. that is required to maintain the existing network is in 31 Table 7. Unit costs of road projects: Gravel roads Unit cost Road condition Road work class Road work type (USD/km) Good Periodic Maintenance Spot regravelling 3,000 Fair Regravelling 17,000 Poor Rehabilitation Partial reconstruction 40,000 Very poor Full reconstruction 60,000 Routine maintenance (per year) 750 – 1,500 Figure 32. Available resources and required expenditures Figure 33. Average road roughness under practical solution Under this practical approach, resource requirements design standards, prioritize investment and to maintain current road conditions are estimated to maintenance, and establish a sustainable financing be USD8.5 million per year. This is even more feasible, mechanism with approximately USD12 million per although under this keep-condition scenario, average year. High-priority roads, which carry the majority of road conditions could deteriorate over the long run. At the traffic in the country, should be kept paved and year 10, it is estimated that average roughness would maintained. Low-priority roads can be treated increase to 6.6 m/km. Therefore, the keep-condition according to lower but more cost-effective standards. scenario could maintain the road network over the Even for unpaved roads, maintenance will be critical to short term but would not be sustainable. ensure the network’s resilience. A steady state of road network development could be achieved and average In sum, to achieve long-term sustainability of road road connectivity ensured. The resource target that sector development, Comoros must reconsider road 32 Comoros should aim at could be USD8 million to Under this approach, the resource USD12 million per year. Over the long run, the norm requirement can be kept to USD12.4 million should be USD12 million, which could ensure per year. improvement of overall road conditions. • By focusing limited resources on priority roads The following chapter examines some potential policy with more traffic, average road roughness— measures to partially fill this financing gap. including paved and unpaved roads—could be improved from a current 5.3 m/km to 4.2 Chapter Summary m/km. Key takeaways from this chapter are as follows: • To achieve long-term sustainability of road sector development, Comoros needs at least • DRTR, under the Ministry of Land Planning and USD8 million (and ideally, USD12 million) to Urbanization, in charge of Land Affairs and follow this practical solution. To address Land Transport is a key player in the resilience and road safety challenges, even management of large investment projects greater resources would be needed. These funded by the government and external additional costs must be examined further development partners. based on more detailed engineering studies, • Comoros has had a road financing mechanism but climate resilience and road safety could through FER, which was restructured into a add 30 percent and 10 percent more, third-generation road fund, Fonds Routier, in respectively, to investment costs. 2019 but has not been fully operationalized. Comoros mobilizes an average of USD5.4 million per year for the road sector, of which 60 percent is financed by external development partners. For the rest, Comoros depends on an RUC levied on fuel sales. • Internal resources have been diminishing because the fuel levy funding mechanism is not tied to inflation and is delinked from fuel consumption. • In real terms, road fund revenues from the fuel levy have decreased 20 percent over the last 8 years. • To maintain the current road network optimally under the current design standards, USD17.5 million per year would be needed. • Even the do-minimum scenario cannot be realized with current available internal and external resources. • To build more resilience, even greater resources would be required (USD21.3 million per year). • One practical solution is to invest in and maintain priority roads while lowering design standards for low-volume roads (~400 km). 33 IV. Increasing Fiscal Efficiency and Improving Road Asset Management As discussed, Comoros must mobilize at least USD8 finance management (Figure 34). million to USD12 million to maintain the current road Even if all these measures are taken, the financial gap network without undermining critical connectivity. In may remain, in which case, support from external recent years, the country has spent approximately development partners may be needed to fill the gap. USD1 million to USD10 million per year (average Comoros’ external stability is at limited risk in the short USD5.4 million). Thus, it is an ambitious target, but the term, but there are medium- and long-term risks (IMF gap can be filled partially using policy measures. This 2020). Nonconcessional borrowing is not chapter explores three important areas to be recommended, particularly in the road sector, where improved in the Comorian context: road asset commercial viability is normally limited. management focused on maintenance and resilience, sustainability of the road fund, and efficiency in public Figure 34. Key policy measures to improve sustainability of road spending • Strategic prioritization Road Asset • Optimization of spending among road works Management • Resilience in road infrastructure • Operationalization of road fund Sustainability of • Fuel levy adjustment for inflation and fuel consumption Road Fund • Full and transparent transfer of fuel levy • Reliablity of investment planning Fiscal Efficiency • Efficiency in budget execution • Efficiency in public procurement and contract management Enhance Road Asset Management developed. See section V for a more detailed discussion of road prioritization. To use available resources wisely, it is essential to Once priority roads are selected, available resources enhance road asset management, including strategic should be focused not only on capital investment, but prioritization. As indicated above, under the also on road maintenance. The RONET analysis constrained budget scenario, Comoros must focus suggests that, under the optimal scenario, in years 1 to investment and maintenance on priority roads while 5, USD5.9 million could be spent on rehabilitation keeping the rest at least passable by adhering to lower projects and that USD4 million would be needed for design standards with minor spot improvement and periodic maintenance and USD2.5 million for routine proper maintenance. maintenance (Table 8; Figure 35). More than half of To select priority roads, strategic prioritization is available resources (which are assumed to be USD12 critical. Based on the RONET analysis, such million in this case) should be dedicated to periodic prioritization must be selective. It is estimated that and routine maintenance (Figure 36). Estimated road approximately 410 km of roads with limited traffic can agency costs for years 6 to 20 would decrease to be treated according to lower standards. A strategic USD9.6 million. By spending resources not only on and systematic prioritization framework needs to be rehabilitation projects to restore the road condition, but also on preservation projects through timely 34 maintenance, the road network will achieve a steady- could be contained over the long run. state condition, and overall government spending Table 8. Annual road agency costs: Practical solution with budget constraint (USD million per year) Scenario Rehabilitation Periodic Routine Total maintenance maintenance Years 1 to 5 Optimal 5.9 4.0 2.5 12.4 Keep condition 5.6 0.3 2.6 8.5 Do minimum 1.0 0.0 2.4 3.4 Years 6 to 20 Optimal 4.1 3.0 2.5 9.6 Keep condition 8.5 0.8 2.6 11.9 Do minimum 8.2 0.9 2.4 11.5 Figure 35. Annual road spending according to type of project Figure 36. Share of road projects Under the keep-condition or do-minimum scenario, climate events such as tropical cyclones, flooding, and government costs could be reduced in the short term, landslides. Road conditions, and thus need for road but this would create more need for capital projects, keep changing over time. Road agencies are investment in the future. To keep the current road supposed to update their road inventory data at least network condition roughly the same, approximately every 5 years. Traffic data should also be updated. The USD8.5 million per year would be needed for years 1 latest available traffic data in Comoros is for 2013, to 5. With this budget, approximately 70 percent which is too old. With updated, accurate date, a road would have to be spent on rehabilitation projects asset assessment and management tool such as while periodic maintenance is compromised, but over RONET should be recalibrated to revise the national the longer term, more resources would be required investment and maintenance program given available because insufficient maintenance would be likely to resources. accelerate road deterioration. Eventually, substantial Various new technologies are available to update full-reconstruction projects would become necessary. underlying data more efficiently at low cost. To Thus, it is of particular importance to keep a good update road condition data, the Bank used a mobile share of available resources allocated to road phone–based application (RoadLab) that can maintenance. automatically record road conditions while driving and To implement good road asset management, it is estimate road roughness (Box 1). High-resolution important to update road inventory and condition satellite imagery can also be a useful source of data periodically. Roads deteriorate with traffic and information to verify local road conditions. Comoros climate events. Comoros is highly vulnerable to could easily take advantage of these advanced 35 technologies because of its relatively small road network. Box 1. New Technologies to Update Road Conditions The World Bank designed RoadLab, a free application available from the Google Play Store, in collaboration with Beldor Center and Softteco and Progress Analytics LLC. The app runs on all android smartphones and tablets. Similar to other road assessment apps, which use accelerometers in smartphones, this app automatically records road conditions such as roughness, average speed, and GPS coordinates of the starting and ending points of each road segment (e.g., at 100-meter intervals). All that is needed is to have RoadLab running on a phone while driving (Figure B1.1). Users can also manually attach other types of data; road safety hazards, pictures of potholes, black spots (areas where crashes are concentrated), or road accidents can be recorded. High-resolution satellite imagery also has the potential to establish road inventory and condition data, although costs may vary depending on the quality of the satellite imagery and the size of the area to be covered. This method is advantageous because it can be applied anywhere on the globe. There have been several attempts to assess road conditions using satellite imagery. In Nigeria, for instance, a pilot study was focused on Kano State, an arid semi-desert area in the northern part of the country. Using a variety of satellite imagery, including SpotMaps at 1.5-meter resolution, 1-meter optical imagery collected to support a polio vaccination program, and Pleiades 0.5-meter resolution (Figure B1.2), approximately 1,000 kilometers of paved and unpaved roads was virtually assessed. The study clearly demonstrates the feasibility of identifying road inventory and structures, such as bridges and culverts, and road conditions to a certain extent. The predicted conditions are approximately 70 percent accurate. Figure B1.1. Display example of RoadLab Figure B1.2. Examples of high-resolution satellite imagery Source: World Bank 2016. 36 Increase Sustainability of the Road Fund through the Treasury, road fund resources are essentially delinked from the RUC and road fuel Comoros has a road financing mechanism through consumption. Despite the organizational the Fonds Routier, a third-generation road fund, but restructuring, the Fonds Routier inherited the same it has not been operationalized, and it has no mechanism. There is no financial autonomy. There is managerial or financial autonomy. Even the previous also insufficient operational capacity to formulate a road maintenance fund, FER, did not have any proper road maintenance and investment program. dedicated fund for road maintenance, even though it The Fonds Routier is authorized to finance not only was called a second-generation road fund (Table 9). maintenance but also investment projects. Under the current fixed-amount transfer mechanism Table 9. Types of road funds Generation Main characteristics 1st (~1980s) Within the national budget framework Earmarked road-related taxes and charges under special off-budget account for road spending Managed by sector ministries 2nd Autonomous, accountable road fund administration (1990s~2010s) Dedicated road fund resources to fund road maintenance, often depending on fuel levy Road authorities responsible for executing road programs 3rd (2~020s) Meet all requirements of 2nd generation Responsible for funding road investment and maintenance Leverage road fund resources to mobilize private financing (e.g., through performance-based contracts or ring-fenced road restoration window) Road agencies responsible for supervising project companies Sources: Benmaamar 2006; Pinard 2012; World Bank 2019b. One of the most important problems that prevent the delays in road projects and programs. After the Fonds Fonds Routier from planning and implementing a Routier was created, but still in transition, revenue solid road maintenance program in a timely manner prediction and actual spending seemed to become is the highly discretionary and unpredictable road more consistent, although the level of revenue flows fund revenues. Although the Treasury transferred has remained low. approximately KMF800 million in 2016, it transferred only KMF130 million in 2019 and KMF520 million in Figure 37. Road fund revenues 2020 (Figure 37). Depending on the overall fiscal balance, the Treasury releases only a portion of the mandatory allocation. Since 2013, an average of approximately 70 percent has been released, but this discretionary decision creates considerable inefficiency in road work programming and implementation. Because of this unpredictable fund transfer, the Fonds Routier sometimes carries nearly half of its annual resources over to the next year, resulting in a low budget execution rate. Although resource transfers by the Treasury are often too late Source: Government data. to implement, there are also often unanticipated 37 To restore stability and predictability of road fund properly, it is logical to restore a fuel consumption– operations, the Fonds Routier’s managerial and dependent formula. This is also important from a financial autonomy must be established. As required traffic management point of view. A milage-based fuel in the second- and third-generation road fund model, tax can provide an incentive to refrain from excessive the Fonds Routier should act autonomously, with vehicle use. Based on current fuel consumption, this sufficient, regular funding secured and technical could generate additional revenue of USD0.6 million capacity expanded. To stabilize road fund revenue, per year. transparency and governance are key. It is important The fuel levy formula must be adjusted for inflation. to verify how much is collected in RUCs and how much Under the current fixed-amount formula, the fuel levy should be transferred to the road fund. If the full to be collected diminishes over time. If past inflation amount of the fuel levy collected were released to the were reflected in the current fuel levy, an additional Fonds Routier, approximately USD0.8 million would be USD0.3 million could be collected. In real terms, this added to the road fund revenue even under the would not be an additional burden for taxpayers or current transfer formula. road users. To improve sustainability of road fund revenue, the The Fonds Routier could secure at least USD3.6 current fixed-amount transfer mechanism must be million per year by making these minor but important linked to actual fuel consumption. In theory, it should adjustments in the fuel levy mechanism (Figure 38). be linked to fuel consumption on the road. As This should be a starting point toward stabilizing road discussed above, fuel consumption in Comoros is fund revenue and operationalizing the new Fonds rapidly increasing, which means that the road network Routier. is being damaged proportionally. To charge road users Figure 38. Potential fuel levy collection Comoros can explore increasing the fuel levy in real revenues with other type of RUCs. The current fuel terms, but caution must be taken. On average, in Sub- levy in Comoros is 3.7 U.S. cents per liter— Saharan Africa, road funds derive 80 percent of their substantially lower than fuel taxes in other Sub- resources from RUCs. The fuel levy is a funding Saharan African countries. The regional average is 8.9 instrument common to all road funds, unlike the other U.S. cents per liter—still less than the regionally instruments mentioned above. Fuel levy collection recommended norm of 17 cents in the region (Figure accounts for an average 76 percent of RUC revenues, 39)—although small countries such as Guinea-Bissau ranging from nearly 100 percent in Comoros, Senegal, and the Gambia often have difficulty raising taxes and Guinea to approximately 30 percent in Niger, because internal markets are small, and taxable where the road fund has been able to diversify its economic activities are limited. Too much taxation 38 could harm their economic viability. Careful overload fines8, road tolls, and ferry and bridge tolls consideration is needed before the fuel levy is raised. can at least partially complement fuel levies. The second and third largest components of RUC revenues Comoros can explore alternative sources of road are normally license and registration fees and tolls. sector revenue. There may not be many options Comoros may be able to exploit these. Other RUC because of the size of the economy and the relatively components may generate limited revenues. Tolls can modest traffic level, but vehicle license and be financially justified only on roads and bridges registration fees, axle load taxes, driver’s license fees, carrying more than 3,000 vehicles per day; only 10 km load-distance charges for heavy goods vehicles, of roads in Comoros carry that volume. Figure 39. Fuel levy comparison for African countries Source: World Bank estimates for Comoros; World Bank (2021) for Mozambique; World Bank (2019b) for the rest of the countries. program level, approximately KMF2.6 billion was Improve Efficiency in Budget Execution committed in the budget, and only KMF2.3 billion was spent—18 percent of the programmed amount and 90 Current budget preparation and efficiency in budget percent of the budget (Figure 41). execution in the road sector must be improved to use limited available resources based on more realistic To reduce such inefficiency in budget execution, plans. The government normally prepares a medium- planning capacity must be increased, with better term public investment plan in collaboration with contractual management and road asset external development partners. For 2016 to 2019, the management systems introduced. The government anticipated envelope at the program level was on must supervise ongoing or planned road programs average KMF10 billion (USD23 million) (Figure 40), more closely and follow up on progress in a timely although only approximately 20 percent was budgeted manner. Because the Comoros road network is on an annual basis. In addition, approximately 66 relatively small, a simple contractual management percent of the budget was actually executed. In 2019, system may help with this. It is also important to for instance, KMF13 billion was expected at the introduce a comprehensive road asset management system to assess priority needs for investment and 8 These resources are expected to decrease over time thanks to more effective enforcement of axle load regulations and increasing truck compliance. 39 maintenance and adjust priority spending, depending To encourage the road authority to perform better, it on resource availability. Currently, the relevant is important to make it an autonomous information and data onto road projects are managed implementation entity. Inefficiency in budget in a fragmented manner. execution is partly attributable to the unpredictable road fund allocation. As experienced in other Sub- Implementation capacity in the sector must be Saharan countries, a road fund alone cannot bring enhanced, especially for donor-funded road much improvement unless it is paired with a capable programs. The discrepancy between programmed and road program implementation unit that functions budgeted amounts mainly occurs in donor-funded autonomously. This would normally be a road agency, programs, which often support large rehabilitation but given the size of the country, it may not be and reconstruction projects. The government’s road practical to establish a separate agency from the program is often too ambitious and too focused on ministry. In Comoros, the DRTR (and the DRP) could be capital investment. The latest national development a more autonomous implementation entity if plan, Plan Comores Emergent, envisages massive equipped with sufficient financial resources and rehabilitation projects over 800 km of roads and new technical skills. It must have full, autonomous construction of 100 km of roads, which will cost responsibility for implementing road programs and be approximately USD250 million (Comoros 2019). accountable for results. There is a draft decree to Similarly, in annual programs, the average level of road establish such an implementation unit under the investment programmed is twice as much as the ministry, but it has not been signed because of the optimal strategy. This is not a problem unique to the strong objections of DRTR and DRP. Since the road sector, but in general, the government integration of DRP into the ministry in 2016, the underestimates the technical and administrative division of responsibility remains unclear between complexity of donor-funded programs. Technical DRP and DRTR. It is important to clarify their capacity within the government to implement projects responsibilities and increase their capacities to make and programs must be built, including through donor- them more professional and accountable road funded projects. management entities. Figure 40. Planned, budgeted, and executed expenditure Figure 41. Budget execution rates Sources: World Bank estimates. Sources: World Bank estimates. high because of its inherent, geographical Increase Efficiency in Road Procurement characteristics. In general, unit rates for road projects and Contract Management in Sub-Saharan Africa are increasing significantly (AfDB 2014). In Comoros, road program unit rates are Road program unit costs in Comoros are relatively already relatively high, because of the size of the 40 market to a certain extent. Road procurement costs There are always opportunities to increase market normally benefit from economies of scale. Unit costs efficiency: The perception is that corruption in for contracts smaller than 100 lane-km tend to be Comoros is among the worst in the region, even particularly high (Figure 42). In addition, Comoros is an among small countries, although it is gradually island and depends on imports for major construction improving (Figure 43). It is challenging but not materials and equipment in the road sector, which impossible to increase efficiency gains in public increases costs, although from a geographic point of procurement. view, the local construction market may be somewhat protected from global and regional competition. Figure 42. Unit costs of upgrading roads in Africa Figure 43. Corruption Perception Index Source: AfDB 2014. Source: AfDB 2015. Although it may not be easy for Comoros to achieve Community-based maintenance contracts can also be economies of scale or competition effects in used for small maintenance projects, generating local procurement, there are several contractual tools to jobs. Routine maintenance is perhaps the best way to increase efficiency in road contracts. First, preserve existing road assets. Each dollar spent on performance-based road maintenance contracts can routine maintenance can save USD3 to USD5 on future tighten an incentive mechanism for contracting and rehabilitation and reconstruction projects (ADB 2018). reduce spending of public resources, allowing The scope of maintenance activities contracted contractors to integrate different activities, such as depends on the complexity of projects and duration of design, investment, and maintenance, into their final contracts (Figure 44). For routine maintenance on output and deliver agreed-upon performance unpaved roads, 1- to 4-year community-based standards at lower costs. Contracting authorities could maintenance contracts can be used, which have the also minimize administrative costs for implementing advantage of generating local jobs and meeting local and supervising projects. For instance, Lancelot (2010) maintenance needs in a timely manner. Each contract showed that performance-based contracts were 20 can cover between 30 and 50 km (2-3 km per worker). percent to 35 percent cheaper than traditional input- There are some successful experiences of community- based road projects in Brazil. Other developed or labor-based maintenance in Latin America, South countries have reported savings of 10 percent to 30 Asia, and other African countries, including Rwanda, percent (ADB 2018). Performance-based contracting Tanzania (AFCAP 2013), and Morocco (World Bank can also ensure timely implementation of necessary 2019d). maintenance, ensuring road access to communities during the rainy season and supporting livelihoods (Iimi 2020). 41 Figure 44. Performance-based road maintenance mentioned institutional reforms are implemented to ensure long-term sustainability of the sector. Such a program could cost USD25 million over 5 years. To avoid increasing the country’s debt stress, concessional loans or grants are desirable, but it cannot be overemphasized that fragmented investment and maintenance without a sustainable financial and implementation framework would be a waste of money. As described above, a sustainable institutional setup must first be established. Based on, Source: ADB 2018. or in parallel to, those institutional reforms, additional resources could be used effectively. As more roads are improved and maintained, more What About the Remaining Gap? efforts would also be called for to ensure road safety. Better road conditions mean higher speeds, which It is likely that the existing financing gap in the road would increase road crash risk in the absence of safety sector would remain even if all the above policy measures. A holistic strategy based on the six pillars of measures were taken. As discussed, the adjustments the Safe System approach is needed: road safety and institutional reforms of the fuel levy calculation management, safer roads and mobility, safer speeds (a and transfer mechanisms could raise an additional new pillar), safer vehicles, safer road users, and post- USD1.7 million. It is difficult to estimate precisely how crash response. More detailed assessment is needed much could be saved by improving budget execution to determine how to increase road safety in Comoros. and increasing procurement efficiency. Based on the There is no national road safety agency, no literature (e.g., Iimi 2006; Lancelot 2010), comprehensive road safety system such as post-crash approximately 20 percent could be saved by increasing care, and no comprehensive data system for road efficiency in public procurement in the infrastructure safety. The African Road Safety Action Plan 2011-2020 sector. Together with currently provided external recommends that 10 percent of road investment and resources, Comoros could potentially mobilize USD7.7 5 percent of road maintenance resources be allocated million per year (Figure 45), leaving an unmet gap of to road safety. Under the current scenario, at least approximately USD5 million. USD1 million must be dedicated to improving road Support from development partners may be needed safety. to fill such a gap, under the condition that the above- Figure 45. Potential internal and external resources and remaining gap 42 • Budget preparation and execution in the road sector are inefficient. The government’s road Chapter Summary program is often too ambitious and too focused on capital investment. Key takeaways from this chapter are as follows: • Planning and implementation capacity must • To use available resources wisely, it is be increased. A more professional, essential to enhance road asset management autonomous road agency or entity equipped practices, including strategic prioritization. with financial and implementation capacity • Strategic prioritization is critical. must be established, possibly by transforming Approximately 410 km of roads with limited the current DRTR and DRP into a Road traffic must be “downgraded� and treated Management Unit, “Cellule de Gestion using lower standards. Routière�. • Once priority roads are selected, available • Although road unit costs are inherently high in resources should be focused not only on Comoros, there are opportunities to increase capital investment, but also on road efficiency in public procurement and contract maintenance. management, for example, through performance- and community-based • More than half of available resources (which maintenance contracts. are assumed to be USD12 million) should be dedicated to periodic and routine • Even if all the above policy measures were maintenance. taken, it is likely that the financing gap would remain. • To implement good road asset management, it is important to update road inventory and • Although USD12.4 million is needed, Comoros condition data regularly. Roads deteriorate could potentially mobilize USD7.7 million, over time. leaving a gap of approximately USD5 million. • Various new, cost-effective technologies are • Within a constrained budget scenario, the available such as mobile phone-based government must make informed decisions applications and high-resolution satellite and consider tradeoffs in terms of road imagery. network standards, building back better to reduce climate vulnerability and increase • To restore stability and predictability of road safety. Building climate resilience and fund operations, the Fonds Routier’s increasing road safety would increase managerial and financial autonomy must be investment needs. (Under the current established. Road fund revenue, scenario, at least USD1 million must be transparency, and governance must be dedicated to increase road safety.) stabilized. • Subject to the country’s fiscal and external • To increase the sustainability of road fund balances, support from external development revenue, the fixed-amount transfer partners may be needed to fill such a gap, mechanism must be linked to fuel including to support implementation of the consumption. above-mentioned institutional reforms, which • The Fonds Routier could and should secure at are essential to ensure long-term least USD3.6 million per year by adjusting the sustainability of the sector. fuel levy mechanism, a significant increase from current average revenue of USD1.9 million. 43 V. Strategic Prioritization and Potential Interventions As discussed above, strategic prioritization is critical to Wider Socioeconomic Benefits from ensure sustainable improvement in road connectivity Transport Connectivity with limited resources. It is estimated that half of the road network must be downgraded to lower design There is a wide range of expected economic and social standards in Comoros, so careful prioritization is benefits from greater transport connectivity. Rural required from the viewpoint of the country’s accessibility, affordability and public transportation, development objectives. market access, agricultural production, and road A prioritization framework based on a multicriteria safety are particularly important in prioritizing roads. analysis (MCA) can allow roads to be prioritized See Appendix D for a more detailed discussion. systematically to accommodate various development objectives. To develop a good prioritization Rural Access and Poverty Reduction framework, it is important to understand the wider Rural accessibility is one of the most important economic and social benefits that accrue from greater constraints in developing countries. Limited road road connectivity in the Comorian context. To this end, accessibility tends to affect poor people more. In this chapter reviews key socioeconomic elements Comoros, poverty is particularly persistent in Moheli related to transport connectivity and provides an and remote areas of Anjouan and Grande Comore. example of a prioritization framework resulting in a list Poverty tends to be high where rural accessibility is of priority roads that can be used as a basis for road poor (Figure 46). Although this merely shows sector programs in Comoros. correlation, not causality, poverty is high where people do not have access to the road network. To alleviate poverty, it is essential to ensure that rural communities have access to roads, especially in remote areas. Figure 46. Rural accessibility and poverty incidence Sources: Haazen and Rajoela 2016; World Bank estimates. The rural accessibility improved between 2014 and rose from 51.5 percent in 2014 to 72.0 percent in 2018 2018 but deteriorated more recently. The rural access (Figure 47), which can be attributed to some major index, a useful measurement to monitor road rehabilitation projects along key national roads, but connectivity and plan road programs in rural areas, fell slightly to 65.3 percent because of the devastating 44 impact of Cyclone Kenneth and lack of road a clear difference in transport modes that people maintenance. choose. For inter-island movements, poorer people Figure 47. Rural access index according to region use maritime transportation more, and richer people use air transport more. In some areas, such as Moroni and Moheli Island, the poor tend to spend more on land transportation than the rich. In Moroni, the poorest spend 9.4 percent of total consumption on land transportation, whereas the richest spend 7.5 percent (Figure 48). In Moheli, the poor spend 3.3 percent of total spending on land transportation, whereas rich people spend 2.5 percent. These numbers are on the low end of average household spending on transport in the region (5 percent to 10 percent depending on country (see, e.g., Lozano- Source: World Bank estimates. Gracia and Young 2014). Trends for poor and wealthy households in Comoros are opposite of general Affordability and Public Transportation findings in other countries; in Comoros, poor people tend to spend more on land transportation and thus In Comoros, transportation costs account for a suffer more if road conditions deteriorate and substantial share of household consumption. There is transport service fares increase. Figure 48. Household spending on transport according to region Moroni Other areas in Grande Comore Anjouan Moheli Sources: World Bank estimates based on 2014 EDMC. 45 Public transportation services are particularly There are opportunities to improve the quality of important for the poor. In Comoros, the public public transportation while contributing to road transport sector is active and well organized. There are safety and addressing climate change. Public 29 transport service operators in Comoros, connecting transportation can mitigate traffic congestion, reduce major cities and towns to the regional capitals. The damage to road surfaces under proper operational network covers the entire country (Figure 49). regulations, enhance road safety, and reduce Approximately 500 buses are operating (Figure 50). emissions. Although cooperatives organize current Based on normal operational cycles (before the COVID bus operations well, buses are old, inefficient, and crisis), 9 it is estimated that approximately 36,000 unsafe. Comoros can take advantage of its relatively passengers use public buses every day, or 13 million small market size, for instance, by introducing an trips per year. To ensure safe, efficient mobility, it is electric bus fleet. In addition, Comoros has high important to keep these main bus routes well potential for renewable energy such as geothermal maintained. and solar power. Such projects are emerging even in the developing world (Box 2). Figure 49. Minibus operating routes and travel time to regional capitals Figure 50. Minibus operations in Comoros Source: World Bank 2019c. 9 The average capacity of a bus is 17 seats. 46 Box 2. Potential of Electric Buses New, cleaner bus technologies for public transport, including electric buses, have been rapidly evolving in developed and developing countries. Many Latin American cities, such as Buenos Aires (Argentina), Mexico City, Santiago (Chile), and Sao Paulo (Brazil), have adopted electric buses. In general, electric buses require high upfront costs for vehicle purchase but lower maintenance and operating costs. In addition, electric buses do not emit local air pollutants and thus offer potential greenhouse gas savings of 50 percent to 100 percent, depending on the sources of electricity generation (World Bank 2019e). In Panama City, to establish a clear path toward a sustainable public transport system, an electric bus pilot test was conducted in 2018 to verify efficiency of electric buses based on actual route conditions within the historic city. Electric buses reduce carbon dioxide emissions, especially at low operating speeds. Predicted carbon dioxide emissions could be halved (UNIDO 2019). In Africa, there are several bus rapid transit projects, for example, in Johannesburg, Pretoria, and Cape Town, South Africa; Dar Es Salaam, Tanzania; and Nairobi, Kenya. The majority of these projects are still using conventional diesel buses, but when economic and environmental benefits are taken into account, there is potential to replace them with electric buses. In Africa, the electric bus market is taking off. Morocco has deployed electric buses since 2017. In Egypt, electric buses were launched on several routes in 2019. In Nairobi, Kenya, a project to replace the existing taxi fleet with electric vehicles is underway. In Uganda, Kiira Motors began to build an electric bus factory in Jinja. Air pollution is a major concern in Kampala. It is expected that the electric bus market will continue to grow in the region (Kuhudzai 2020). explains approximately 16 percent of the final price. In Market Access and Inter- and Intra-Island addition, interisland transport costs are KMF1,250. Connectivity Within Anjouan, crop prices also double by the time they are sold at the market. Land transport costs In Comoros, Moroni is the most important primary account for 15 percent of banana prices at market for everyone. To assess general market Mutsamudu. High profit margins for retailers can also accessibility, the Market Access Index is calculated be attributed to the limited supply of agricultural based on the weighted average of purchasing power commodities because of low market participation or market size, with transport impediment taken into rates by farmers. Despite Anjouan’s significant account. Because of its low population density, lack of agricultural potential, market transactions are major cities, and poor road conditions, Moheli is particularly limited (Figure 53), largely because of the considered to be least connected to the domestic lack of reliable, affordable interisland connectivity. market (Figure 51). The primary market, Moroni, is far from many parts of To increase market accessibility, inter- and intra- the country, especially Anjouan and Moheli. island transportation must be improved. Inter- and High commodity prices, amplified by market intra-island transport-related costs account for inefficiency, are creating an additional burden for approximately 10 percent to 30 percent of the price urban consumers. Comoros imports significant differentials for agricultural produce between agricultural crops and products every year. In 2016, production and consumption areas. For example, the approximately USD110 million worth of crops and farmgate price for 50 kg of bananas of KMF2,500 at food products was imported—approximately 18 Nioumachoua increases to KMF8,500 at Moroni percent of the country’s GDP. By enhancing domestic (Figure 52). Land transport costs are KMF1,500, which market integration, urban consumers and rural 47 farmers could benefit. Figure 51. Market Access Index, 2021 Figure 52. Banana value chain analysis (50 kg) Niumachoua, Moheli to Volo Volo, Moroni Mremani to Mutsamudu, Anjouan Source: World Bank survey. Figure 53. Market sales as a share of production Source: World Bank estimates based on INSEED (2014). 48 Agricultural Production Comoros also produces valuable export crops such as vanilla, ylang ylang, and cloves. According to the Agriculture is an important sector in Comoros. Enquete 1-2-3 survey (INSEED 2014), substantial Approximately 62 percent of the population is agricultural production activities take place in the engaged in agricultural production. Major food crops northern areas of Grande Comore, Anjouan, and produced include coconuts, cassava, and bananas. Moheli (Figure 54). Figure 54. Main crop production according to village Source: World Bank estimates based on INSEED (2014). Agricultural productivity is low. The vast majority of (Figure 55). Not only transport connectivity, but also agricultural production is subsistence farming. other complementary interventions are important for Because of the poor condition of the road network and agricultural growth. In particular, intermediary service lack of transport services, many farmers do not have providers, such as agricultural cooperatives and crop good access to input or output markets, so the use of collectors, play an important role in increasing advanced inputs is minimal. Only 12 percent of farmers’ crop production (Iimi 2019). households used fertilizer, and 11 percent used Figure 55. Estimated elasticity of crop production values improved seeds. Intermediary services are unavailable in most places. Crops are not often collected by cooperatives or traders, and many farmers do not have any transportation (Iimi 2019). Among others, accessibility to Moroni, the primary market in the country, is of particular importance to stimulate agricultural growth. A recent study predicted that, although a 10 percent increase in accessibility to the Moroni market could increase agricultural production by 8.5 percent, it could enhance farmers’ market participation by 19 percent Source: Iimi 2019. 49 Climate Resilience 𝐼𝑛𝑓𝑟𝑎𝑠𝑡𝑟𝑢�𝑡𝑢𝑟𝑒 𝑟𝑖𝑠𝑘 = 𝐶𝑙𝑖𝑚𝑎𝑡𝑒 𝑒𝑥�𝑜𝑠𝑢𝑟𝑒 Economic infrastructure risk is determined by × 𝑉𝑢𝑙𝑛𝑒𝑟𝑎�𝑖𝑙𝑖𝑡𝑦 exposure to climate events and vulnerability of infrastructure. In Comoros, the road network is 𝐸�𝑜𝑛𝑜𝑚𝑖� 𝑟𝑖𝑠𝑘 𝑜𝑓 𝑒𝑥𝑡𝑟𝑒𝑚𝑒 𝑒𝑣𝑒𝑛𝑡𝑠 particularly exposed to flood risk. According to a flood = 𝐼𝑛𝑓𝑟𝑎𝑠𝑡𝑟𝑢�𝑡𝑢𝑟𝑒 𝑟𝑖𝑠𝑘 map based on past events, the southern parts of the × 𝐸�𝑜𝑛𝑜𝑚𝑖� 𝑒𝑥�𝑜𝑠𝑢𝑟𝑒 islands are generally more exposed to climate risk. Anjouan is exposed to the highest economic risk of Vulnerability depends on the technical strength and extreme climate events (Figure 56). Roads with high resilience of the roads. Poorly maintained roads tend exposure to climate risk are concentrated on Anjouan to be more vulnerable. The economic risk of extreme and Moheli and, to a certain extent, on the southern climate events is determined according to part of Grande Comore. There are 53 road sections infrastructure risk and the degree of economic where climate risk is particularly high (score >0.43). activities that disrupted infrastructure services could The total length of these roads is approximately 165 potentially affect (economic exposure). To assess km, of which 106 km is in Anjouan. Approximately 37 underlying economic activities, village-level crop km of roads is identified as being at high economic risk production data can be used, as shown above. in Grande Comore and 22 km in Moheli. These roads are a priority from a climate resilience point of view. Figure 56. Economic risk of extreme climate events Source: World Bank estimation. explicit and implicit costs, such as emergency response Road Safety costs, medical and rehabilitation costs, long-term care Road safety is an important negative externality that costs for the disabled, permanently or temporarily lost must be factored in (Figure 57). Globally, it is income and disruption to business and work, police, estimated that the socioeconomic cost of road judicial system and jail costs (for those found guilty in accidents is typically 3 percent to 10 percent of GDP crashes), property damage, lost time, and fuel and air (World Bank 2019a). Costs of crashes include various pollution costs from crash-generated congestion. 50 These costs could retard long-term economic growth A web-based application, a simplified road safety risk by approximately 1 percentage point for some low- to assessment tool using automated imagery analysis, middle-income countries (World Bank 2017b). was used to understand existing road safety risk in Comoros (World Bank 2019f). The methodology of the Figure 57. Road safety risks in Comoros and other application follows a traditional risk assessment countries formula (Appendix F): Road safety risk is a function of three factors: danger (likelihood that a crash will happen), vulnerability (risk of injury of road users if a crash occurs), and exposure (extent to which a road user is exposed to a risk): 𝑅𝑜𝑎𝑑 𝑠𝑎𝑓𝑒𝑡𝑦 𝑟𝑖𝑠𝑘 = (𝐷𝑎𝑛𝑔𝑒𝑟 × 𝑉𝑢𝑙𝑛𝑒𝑟𝑎�𝑖𝑙𝑖𝑡𝑦) × 𝐸𝑥�𝑜𝑠𝑢𝑟𝑒 In Comoros, road safety risk is high, especially in urban areas. The assessment was focused on main roads; sufficient data were collected for analysis of approximately 384 km of main roads. The analysis indicates that 59 percent of the 100-m road sections Source: World Bank estimates. surveyed are high and very high risk , 37 percent are medium risk, and 3 percent are low risk (Figure 58). By For Comoros, the World Bank (2019a) estimated the comparison with other countries in Africa where the cost of fatalities and serious injuries to be USD90 same methodology has recently been applied, overall million, or 8.8 percent of GDP, in 2016. A holistic risk in Comoros is worse than in Mozambique and approach is needed following the principles of the Safe comparable with Liberia. System Approach (e.g., OECD 2008; World Bank 2020b). See more detail in Appendix E. Figure 58. Road safety risks according to region or island Urban areas Rural areas Source: World Bank estimates. By region and island, Moheli is most dangerous, designed, with wide sidewalks, road crossings, and especially in its urban zones, followed by Grande well-structured roundabouts, although they can be Comore. High-risk areas are concentrated around improved with features such as speed humps to major intersections (Figure 59). Some roads are well manage speed, but others do not have such 51 structures: Many pedestrians walk close to traffic implementation of any road program. Global without any separation. Rural roads are less experience shows that some road engineering dangerous, although still medium or high risk. In measures, such as road humps and narrowing, as well Moheli, rural roads are very dangerous because of the as some speed management devices, such as speed lack of road structures, difficult road alignments, and cameras, are cost-effective for managing speed, which poor road conditions. In mountainous areas, there are is fundamental to road safety improvements. many blind spots without any safety features. Vehicle International studies show that a 1 percent decrease speeds are relatively high. Many people in rural in speed results in a 4 percent decrease in deaths. communities walk along roads without sidewalks. Management of speed is a major, low-cost road safety Rural roads often pass through the middle of villages. opportunity for the country. Box 3 describes how other small island states have approached road safety Sound, evidence-based (not common sense–based) challenges. road safety interventions should be integrated into Figure 59. Road safety risks and black spots Source: World Bank estimates. Box 3. Road Safety Approaches in Small Island States In the Pacific, some small island states, such as Samoa and Solomon Islands, have recently made particular efforts to increase road safety by revising or developing a national road safety action plan. Small countries must take a holistic approach to various challenges but with a focus on particularly weak areas, using existing mechanisms and capacity wherever possible. Samoa updated its National Action Plan for Road Safety for 2021-30. To improve interministerial coordination, the functions and roles of the National Road Safety Committee will be strengthened, and a ministerial road safety council will be established and meet with the committee quarterly. The road safety strategy will be updated and target a 50 percent reduction in road crash fatalities and serious injuries by 2030. Road crash data systems will also be upgraded, with crash investigation training implemented. The committee and the Ministry of Finance will calculate an agreed-upon value of life and serious injury avoided. Vehicle safety requirements for imported vehicles will be reviewed. 52 Priority actions across the Safe System Pillars will be pursued vigorously, with an emphasis on road signage and markings, pedestrian protection platforms, bus lay-bys, phase out of left hand–drive vehicles, requirement for and enforcement of wearing rear seatbelts, establishment of a graduated driver licensing system inclusive of a learner permit, introduction of drug-impaired driving legislation and enforcement, multimedia community education on road safety, enforcement of helmet-wearing legislation, increase in number of people trained in first aid and paramedic care, installation of speed signs and lower speed limits in higher pedestrian activity areas, and enforcement of public bus speed limit. Solomon Islands has also embarked upon various policy actions to increase road safety. A secretariat will be established as a road safety lead agency within the Ministry of Infrastructure Development. A national road safety committee will be established as a governance and decision-making structure to facilitate coordinated, targeted actions across the relevant ministries. Appropriate road safety legislation must be developed and implemented in Solomon Islands, and an adequate funding mechanism must be established to increase the road safety capacity of relevant ministries. A national road safety vision and strategy will be developed. The government envisages a 50 percent reduction in road crash fatalities and serious injuries by 2030. Similar to Samoa, road crash data systems must be strengthened. Priority actions include safer road infrastructure, vehicle fleet, and road users; better emergency post-crash care; and safer speeds on the road network. Source: World Bank 2020c; 2020d. Strategic Prioritization—Multicriteria depending on a country’s development objectives Analysis (Table 10). It is important to reach a consensus across all economic and social sectors (e.g., agriculture, Prioritization can be accomplished based on a health, education, private sector development). Based conventional multicriteria analysis, for which a wide on agreement, priority roads can be systematically and range of socioeconomic variables can be used, objectively selected based on actual data. Table 10. Potential indicators for prioritization Development objective Possible measurement Poverty reduction District-level poverty rate Public transport Routes where public transport (minibus) operates Agricultural growth Agricultural production values along each road section, evaluated at market prices Health care access Number of health facilities along each road section Education access Number of schools along each road section Climate resilience Economic risk of climate events based on proximity to flood-prone areas, road vulnerability, and economic activities exposed to risk Road safety Number of black spots identified in road safety assessment General mobility Traffic count on each road section (average daily traffic in 2013) To capture different types of development different types of indicators representing different objectives, diverse indicators must be selected. Some development objectives must be selected. potential parameters are correlated with one another. For illustration purposes, a simple framework is Major bus routes are consistent with roads with high developed based on the following five criteria: traffic. To avoid high collinearity among indicators, 53 agricultural production, access to social services, used for simplicity. The total score is once again climate risk and vulnerability, road safety, and traffic. normalized from 0 to 1 for presentation purposes. High-priority roads (indicated in red) are different among the indicators (Figure 60). Each parameter is normalized between 0 and 1, and equal weights are Figure 60. Five key socioeconomic criteria for road prioritization: Example Agricultural production Health access Economic risk of climate events Road safety Traffic volume The prioritization result is shown in Figure 61. The according to lower design standards. By better detailed list of priory roads is presented in Appendix balancing high- and low-priority roads, Comoros can A. The overall priority score is broadly consistent with increase fiscal sustainability in the road sector while current road programs. Based on the above maintaining overall connectivity in the country. prioritization framework, national roads are the most Any prioritization tool must be applied strategically important to maintain (Figure 62). There are 187 km to make road investments selective because of the of top-priority (5th quantile) national roads. continued tight fiscal space, growing uncertainty in Approximately 46 percent of national roads are the global economy, and unforeseen external shocks. considered to be top priority (Figure 63). Many priority There is much uncertainty related to the outbreak of roads are located in Anjouan and Moheli (Figure 64). COVID-19, the outlook for global commodity markets Approximately 42 percent of roads in Anjouan and 70 is unclear, and climate-related uncertainty is high. As percent in Moheli are considered to be top priority new data and technologies become available, the (Figure 65). The government should focus available prioritization framework must be kept updated, and resources on those priority roads. adjusted if needed, to use limited available resources Based on the prioritization results, approximately effectively to achieve long-term, inclusive growth. 400 km of roads could be downgraded, or treated 54 Figure 61. Overall prioritization score for key road investment and maintenance Sources: World Bank survey. Figure 62. Length of priority roads according to road class Figure 63. Share of priority roads according to road class 55 Figure 64. Length of priority roads according to island Figure 65. Share of priority roads according to island Chapter Summary transportation must be improved selectively. • Agriculture is an important sector in Comoros. Key takeaways from this chapter are as follows: Productivity is low. Greater market • Strategic prioritization is critical to increase accessibility can stimulate agricultural growth. road connectivity sustainably with limited • Road traffic crashes are an important negative resources. externality of road network development. • To prioritize roads systematically, a Road safety risk is especially high in urban prioritization framework based on areas. Moheli is the most dangerous. multicriteria analysis must be prepared based • When implementing any road program, sound on the country’s development objectives. evidence-based road safety interventions • To have a good prioritization framework, it is should be used. Management of speed is important to understand wider economic fundamental. benefits of road connectivity in the Comorian • Taking all the above into account, a strategic context: prioritization framework has been developed • Rural accessibility is important. The poor tend based on five criteria: agricultural production, to suffer more from limited road accessibility, access to social services, climate risk and especially in remote areas in Moheli and vulnerability, road safety, and traffic. Anjouan. • National roads are generally high priority; • Affordability is important. Land transportation many high-priority roads are located in costs account for a substantial share of Anjouan and Moheli. The government should household consumption in some areas, such focus available resources on those high- as Moroni and Moheli. priority roads. • Public transportation services are particularly • Prioritization identified 400 km of low-priority important to support the poor. roads that should be downgraded (treated using lower design standards). • Opportunities (e.g., electric buses) exist to improve the quality of public transportation while increasing road safety and ameliorating climate change. • Market accessibility is important. Moroni is the most important primary market for everyone. Inter- and intra-island 56 VI. Conclusion Comoros has a well-established road network. network requires reconstruction, Although network coverage is comprehensive, rehabilitation, and periodic maintenance to connecting most major towns and villages, road upgrade its overall condition and overcome conditions are poor and have deteriorated in recent the backlog of deferred maintenance. This years. Half of the road network is in poor or very poor would require greater financial commitment condition. Comoros is also highly vulnerable to climate than under a steady-state scenario, under events such as tropical cyclones, flooding, and which network deterioration reaches a landslides, which increase challenges in the road relatively constant level year after year, and sector. future interventions can be anticipated and reasonably taken care of with existing With the assistance of external development partners, resources. the government of Comoros has been making particular efforts to rehabilitate major national roads • Critically review road standards. To ensure in recent years, but maintenance expenditures tend to long-term sustainability in road sector be omitted. Including internal and external resources, development, at least USD12.4 million would Comoros spent on average USD5.4 million per year on be needed per year for the next 5 years. Under road sector development for the period of 2016 to this practical solution, available resources 2019, of which approximately 70 percent was used for would be focused on priority roads, allowing capital investment. Available resources are highly lower design standards for low-volume roads variable from one year to another, and road (~400 km). Half of the network has been maintenance projects tend to be compromised when significantly damaged and is in need of full there are shortfalls. reconstruction. More cost-effective standards Comoros relies heavily (~60 percent) on external should be applied to feeder roads with limited resources for road sector funding. Although Comoros traffic. has a road funding mechanism, internal resources • Strategically prioritize roads. A prioritization have diminished over time because of institutional framework based on multicriteria analysis can flaws. The current allocation mechanism is highly be used to prioritize roads to accommodate discretionary and unpredictable, preventing the road different development objectives and fill the fund and road authority (DRTR) from planning and financing gap. implementing road programs in a timely, consistent • The list of priority roads that has been manner. prepared based on multicriteria analysis Current practice is financially and operationally (Appendix A) should be used to determine unsustainable. To maintain the current road network which road programs should be implemented optimally under current design standards, USD17.5 next. million per year would be needed, which is far from affordable under current budgetary availability. With • Road asset management must be enhanced. current available internal and external resources, even Current road programming practice is not the do-minimum scenario cannot be implemented. supported by evidence. To use available resources wisely, it is essential to establish a To achieve sustainability in road sector spending while comprehensive road asset management maintaining overall road connectivity, the main system, although it does not have to be very recommendations are as follows: data intensive. As underlying data are • Gradually move toward steady-state updated, the national road program must be condition of the road network. The road updated based on the road asset management analysis. 57 • Regular road condition and traffic surveys contracts and community-based maintenance should be conducted. To implement good contracts. road asset management, it is fundamental to • More external resources should be update underlying road condition and traffic mobilized. Even if all the above policy data regularly. Required resources should be measures were instituted, it is likely that a secured over the medium to long term. financing gap of USD5 million per year would • Cost-effective data collection tools should be remain. Subject to the country’s fiscal and used. Various new cost-effective technologies external balances, external development are available, such as mobile phone-based partners may be able to fill such a gap, but it is software. essential to implement the above-mentioned institutional reforms to ensure long-term • More than half of available resources should sustainability of the sector. be dedicated to periodic and routine maintenance if USD12 million is available per • Additional tradeoffs must be considered to year. Current resource allocation is build resilience and safety in the road inconsistent with optimal distribution.. network. Under a constrained budget scenario, the government must make • To restore stability and predictability of road informed decisions and consider tradeoffs in fund operations, the managerial and terms not only of road standards, but also of financial autonomy of the Fonds Routier building back better to reduce the climate must be reconfirmed. A transparent decision- vulnerability of priority roads and increase making process must be established to safety, which would add to investment needs transfer a full amount of dedicated resources. but bring overall positive economic returns in • To ensure sustainability of Fonds Routier the long run. revenue, the fuel levy must be linked to • Because there are no resilient design actual fuel consumption. The current fixed standards currently available in the country, amount transfer mechanism is not and it is not possible to determine with further sustainable. precision what the specific additional costs • RUCs should be increases and diversified. The would be, a key next step will be to consider current fuel levy is minimal in Comoros. investing in development of design Although limited, there are other possible specifications and estimate unit costs for sources of revenue such as vehicle license and different work items. registration fees, axle load taxes, parking fees, and driver’s license fees. • Planning and implementation capacity must be increased. A more professional, autonomous road agency or entity equipped with financial and implementation capacity must be established, possibly by transforming the current DRTR and DRP into a road management unit. • Cost efficiencies in delivery of road assets should be achieved by providing high- powered incentives following results and user-oriented approaches through, for instance, performance-based maintenance 58 References EC (European Commission). 2015. Etude d'Elaboration du Plan Directeur National du Transport Routier en Union de Comores ADB (Asian Development Bank). 2018. Guide to pour la Periode 2015-2025. EC. Performance-Based Road Maintenance GRSF (Global Road Safety Facility). 2019. Contracts. Mandaluyong, Philippines: ADB. “Road Safety Country Profile: Comoros.� AFCAP (Africa Community Access Program). World Bank, Washington, DC. 2013. Improved Maintenance Systems for Gwilliam, Ken. 2011. “Africa's Transport District Roads in Tanzania. AFCAP. Infrastructure: Mainstreaming Maintenance AfDB (African Development Bank). 2014. Study and Management.� World Bank, on Road Infrastructure Costs: Analysis of Washington, DC. Unit Costs and Cost Overruns of Road Haazen, Dominic, and Voahirana Hanitriniala Infrastructure Projects in Africa. Abidjan, Rajoela. 2016) “Sub-National Analysis of Côte d’Ivoire: AfDB. Systematic Differences in Health Status and ———. 2015. Seychelles Infrastructure Action the Access to and Funding of Health Plan. Abidjan, Côte d’Ivoire: AfDB. Services: An Example from Comoros.� Health, Nutrition & Population Global Bell, Clive, and Susanne van Dillen. 2012. “How Practice Discussion Paper. World Bank, Does India's Rural Roads Program Affect the Washington, DC. Grassroots? Findings from a Survey in Orissa.� World Bank Policy Research Iimi, Atsushi. 2006. “Auction Reforms for Working Paper No. 6167. World Bank, Effective Official Development Assistance.� Washington, DC. Review of Industrial Organization 28: 109– 128. Benmaamar, Mustapha. 2006. Financing of Road Maintenance in Sub-Saharan Africa: ———. 2019. “Hidden Treasures in the Reforms and Progress Towards Second Comoros: The Impact of Inter-Island Generation Road Funds. Discussion Paper Connectivity Improvement on Agricultural No. 6. Sub-Saharan Africa Transport Policy Production.� Policy Research Working Paper Program (SSATP). 8960. World Bank Group, Washington, DC. Briceno-Garmendia et al. 2008. “Financing ———. 2020. “Performance-Based Road Public Infrastructure in Sub-Saharan Africa: Contracts in Zambia.� Review of Industrial Patterns and Emerging Issues.� AICD Organization 57: 107–129. Background paper 15. Iimi, Atsushi, and Kulwinder Rao. 2018. “Spatial Comoros. 2019. Plan Comores Emergent. Analysis of Liberia's Transport Connectivity Government of Comoros. and Potential Growth.� World Bank, Washington, DC. Danida. 2010. Impact Evaluation of Danida Support to Rural Transport Infrastructure in IMF (International Monetary Fund). 2020. Nicaragua. “Union of the Comoros: 2019 Article IV Consultation – Staff Report; Press Release.� Dercon, Stefan, Daniel Gilligan, John Hoddinott, IMF Country Report No. 20/198. IMF. , and Tassew Woldehanna. 2008. “The Impact Washington, DC of Agricultural Extension and Roads on Poverty and Consumption Growth in Fifteen INSEED. 2014. Enquête sur les dépenses de Ethiopian Villages.� American Journal of consommation des ménages aux Comores Agricultural Economics 91(4): 1007–1021. (EDMC). Institut National de la Statistique, 59 des Etudes Economiques et Sante, de la Solidarite, de la Protection Démographiques, Comoros. Sociale et de la Promotion du Genre. The Union of Comoros. Job, R. F. S. and L. W. Mbugua. 2020. “Road Crash Trauma, Climate Change, Pollution Mu, Ren, and Dominique van de Walle. 2011 and the Total Costs of Speed: Six Graphs “Rural Roads and Local Market That Tell the Story.� GRSF Note 2020.1. Development in Vietnam.� Journal of Global Road Safety Facility, World Bank, Development Studies 47(5): 709–734. Washington DC. Nilsson, G. 2004. “Traffic Safety Dimensions and Khandker, Shahidur, Zaid Bakht, and Gayatri the Power Model to Describe the Effect of Koolwal. 2009. “The Poverty Impact of Speed on Safety.� Bulletin 221, Lund Rural Roads: Evidence from Bangladesh.� Institute of Technology, Department of Economic Development and Cultural Technology and Society, Traffic Change 57(4): 685–722. Engineering, Lund, Sweden. http://www.dissertations.se/dissertation/a995 Kuhudzai, Remeredzai Joseph. (2020). “We Hope 2d343f/. to See More Electric Buses on the Roads as More African Cities Move to Bus Rapid OECD (Organisation for Economic Co-operation Transit Systems.� CleanTechnica. and Development). 2008. Towards Zero: https://cleantechnica.com/2020/08/06/we- Ambitious Road Safety Targets and the Safe hope-to-see-more-electric-buses-on-the- System Approach. Paris, France: OECD. roads-as-more-african-cities-move-to-bus- Pinard, Michael. 2012. Progress on rapid-transit-systems/ Commercialized Road Management in Sub- Lall, Somik V, and Taye Mengistae. 2005. “The Saharan Africa. Sub-Saharan Africa Impact of Business Environment and Transport Policy Program Working Paper Economic Geography on Plant-Level No. 92. Productivity: An Analysis of Indian Roberts, Peter, K. C. Shyam. and Cordula Industry.� World Bank Policy Research Rastogi. 2006. “Rural Access Index: A Key Working Paper No. 3664. World Bank, Development Indicator.� Transport Papers Washington, DC. TP-10. World Bank Group, Washington, DC. Lancelot, Eric. 2010. “Performance Based Small, Martin, and Justin Runji. 2014. “Managing Contracts in the Road Sector: Towards Road Safety in Africa: A Framework for Improved Efficiency in the Management of National Lead Agencies.� Sub-Saharan Maintenance and Rehabilitation.� Transport Africa Transport Policy Program Working Papers TP-31, The World Bank Group. Paper 101. Africa Transport Policy Lokshin, M., and R. Yemtsov. 2005. “Has Rural Programme, Washington, DC. Infrastructure Rehabilitation in Georgia Turner, B. M., R. F. S. Job, S. Mitra, M. H. M. A. Helped the Poor?� World Bank Economic Uddin. 2021. Guide for Road Safety Review 19(2): 311–333. Interventions: Evidence of What Works and Lozano-Gracia, Nancy, and Cheryl Young. 2014. What Does Not Work (English). Washington, “Housing Consumption and Urbanization.� DC: World Bank Group. Policy Research Working Paper 7112. World http://documents.worldbank.org/curated/en/2 Bank, Washington, DC. 06691614060311799/Guide-for-Road- Ministry of Health, Comoros. 2015. Annuaire Safety-Interventions-Evidence-of-What- Works-and-What-Does-Not-Work Statisticque: 2014-2015. Ministere de la 60 UNIDO (United Nations Industrial Development Safety-Opportunities-and-Challenges-Low- Organization). 2019. “Accelerating the and-Middle-Income-Country-Profiles.pdf Transition to Sustainable Mobility and Low ———. 2019b. Scaling Up Private Sector Carbon Emissions in Panama City.� UNIDO, Participation in Road Asset Management in Vienna. Sub-Saharan Africa. Washington, DC: World Bank. 2013. “Impact Evaluation for World Bank Group. Infrastructure: General Guidance and ———. 2019c. “Analyse Spatiale de la Existing Evidence.� World Bank, Connectivité du Transport et des Washington, DC. Potentialités de Croissance Economique aux ———. 2015. “Smartphone App Innovation Comores.� World Bank Group, Washington, Fund – Big Data for Development DC. Revamping Road Condition and Road Safety ———. 2019d. “Morocco Community-Based Monitoring.� Report No. 98194, prepared by Rural Roads Maintenance.� Project Paper ROAD LAB. World Bank, Washington, DC. Report No. PP2722. World Bank Group, ———. 2016. “Measuring Rural Access: Using Washington, DC. New Technologies.� World Bank, ———. 2019e. “Green Your Bus Ride: Clean Washington, DC. Buses in Latin America: Summary Report.� ———. 2017a. “Enhancing the Climate World Bank Group, Washington, DC. Resilience of Africa’s Infrastructure: The ———. 2019f. “Consulting Service for Applying Road and Bridge Sector.� World Bank, Simplified Road Safety Assessment Washington, DC. Methodology Using Automated Image ———. 2017b. “The High Toll of Traffic Analysis and Pilot Assessment Study of Injuries: Unacceptable and Preventable.� A National Highways in Liberia.� World Bank World Bank Study. Advisory Services and Group, Washington, DC. Analytics Technical Report P155310. World ———. 2020a. “Country Partnership Framework Bank, Washington, DC. ———. 2018a. for the Union of Comoros for the Period “Madagascar: Spatial Analysis of Transport FY20-24.� World Bank Group, Washington, Connectivity and Growth Potential.� World DC. Bank Group. ———. 2020b. Road Safety Interventions: ———. 2018b. “Desenvolvimento de uma Evidence of What Works and What Does Not Metodologia Simplificada de Avaliacao da Work. Washington, DC: World Bank Group. Seguranca Rodoviaria Atraves de uma Analise Automatizada de Imagens e um ———. 2020c. Road Safety Management Estudopiloto de Avaliacao das Auto-Estradas Capacity Assessment for Samoa. Nacionais de Mocambique.� World Bank Washington, DC: World Bank Group. Group, Washington, DC. ———. 2020d. Road Safety Management ———. 2019a. Guide for Road Safety Capacity Assessment for the Solomon Opportunities and Challenges: Low- and Islands. Washington, DC: World Bank Middle-Income Countries Country Profiles. Group. Washington, DC: World Bank. ———. 2021. “Investing in Mozambique’s http://documents.worldbank.org/curated/en/4 Connectivity: Toward Efficient, Sustainable 47031581489115544/pdf/Guide-for-Road- Road Spending.� World Bank Group, Washington, DC. 61 Zeufack, Albert G., Ibrahim Elbadawi, and Taye Mengistae. 2006. “Market Access, Supplier Access, and Africa’s Manufactured Exports: An Analysis of the Role of Geography and Institutions.� Policy Research Working Paper No. 3942. World Bank, Washington, DC. 62 Appendix A. Road Prioritization Results Current Prioritization score: Class NUM_SECT Length condition Agriculture Health Climate Safety Traffic Overall RN N1_01 2.25 Fair 0.00 0.25 0.00 0.00 1.00 0.47 RN N1_02 1.03 Fair 0.29 0.00 0.15 0.00 1.00 0.54 RN N1_03 1.05 Fair 0.36 0.00 0.17 0.00 1.00 0.57 RN N1_04 3.98 Poor 0.36 0.00 0.25 0.45 0.80 0.69 RN N1_05 2.34 Fair 0.50 0.00 0.23 0.00 0.80 0.57 RN N1_06 2.74 Fair 0.50 0.00 0.23 0.00 0.80 0.57 RN N1_07 2.8 Fair 0.36 0.25 0.16 0.00 0.80 0.59 RN N1_08 3.33 Fair 0.36 0.25 0.15 0.00 0.80 0.58 RN N1_09 5.42 Good 0.00 0.00 0.00 0.00 0.74 0.28 RN N1_10 2.29 Good 0.45 0.25 0.06 0.00 0.74 0.56 RN N1_11 2.31 Good 0.63 0.25 0.06 0.00 0.74 0.63 RN N1_12 2.19 Good 0.58 0.00 0.04 0.00 0.74 0.51 RN N1_13 2.06 Good 0.41 0.00 0.02 0.00 0.74 0.44 RN N1_14 3.18 Good 0.45 0.25 0.01 0.00 0.74 0.54 RN N1_15 0.71 Good 0.26 0.25 0.00 0.00 0.74 0.47 RN N1_A_01 0.34 Fair 0.00 0.25 0.00 0.00 0.00 0.09 RN N1_A_02 0.73 Fair 0.00 0.25 0.00 0.00 0.00 0.09 RN N1_A_03 2.8 Fair 0.29 0.25 0.15 0.00 0.00 0.26 RN N2_01 0.75 Good 0.00 0.25 0.00 0.00 0.00 0.09 RN N2_02 1.38 Fair 0.34 0.25 0.21 0.00 0.96 0.66 RN N2_03 2.17 Good 0.00 0.00 0.00 0.00 0.91 0.34 RN N2_04 1.94 Good 0.34 0.00 0.12 0.00 0.91 0.51 RN N2_05 2.71 Good 0.49 0.00 0.17 0.00 0.63 0.48 RN N2_06 3.57 Good 0.66 0.25 0.22 0.00 0.63 0.66 RN N2_07 0.84 Good 0.18 0.25 0.06 0.00 0.63 0.42 RN N2_08 3.93 Poor 0.48 0.25 0.50 0.62 0.63 0.92 RN N2_09 3.93 Fair 0.23 0.25 0.14 0.40 0.63 0.62 RN N2_10 4.05 Fair 0.56 0.00 0.30 0.00 0.63 0.56 RN N2_11 3.72 Poor 0.00 0.00 0.00 0.40 0.63 0.38 RN N2_12 3.15 Poor 0.36 0.25 0.28 0.00 0.65 0.57 RN N2_13 3.05 Poor 0.11 0.00 0.08 0.00 0.65 0.31 RN N2_14 2.33 Fair 0.00 0.00 0.00 0.00 0.65 0.24 RN N2_15 3.87 Poor 0.35 0.25 0.27 0.22 0.65 0.65 RN N2_16 1.81 Fair 0.25 0.25 0.12 0.22 0.65 0.56 RN N2_17 3.68 Poor 0.00 0.00 0.00 0.00 0.65 0.24 RN N2_18 2.35 Poor 0.42 0.25 0.38 0.00 0.65 0.63 RN N2_19 0.35 Poor 0.60 0.25 0.62 0.33 0.65 0.91 RN N3_01 0.81 Poor 0.00 0.00 0.00 0.00 0.59 0.22 RN N3_02 4.05 Poor 0.00 0.00 0.00 0.22 0.59 0.30 RN N3_03 4.36 Poor 0.38 0.00 0.00 0.33 0.59 0.48 RN N3_04 2.77 Poor 0.00 0.00 0.00 0.40 0.59 0.37 RN N3_05 2.66 Fair 0.00 0.00 0.00 0.33 0.59 0.34 RN N3_06 1.74 Poor 0.59 0.25 0.11 0.22 0.59 0.66 RN N3_07 0.6 Poor 0.00 0.25 0.00 0.22 0.59 0.40 RN N3_08 1.79 Fair 0.00 0.25 0.00 0.22 0.59 0.40 RN N3_09 3.27 Fair 0.60 0.50 0.16 0.00 0.69 0.72 RN N3_10 0.74 Poor 0.58 0.25 0.24 0.00 0.69 0.66 RN N3_11 1.68 Poor 0.58 0.00 0.27 0.00 0.69 0.58 RN N3_12 1.26 Fair 0.32 0.25 0.11 0.00 0.69 0.51 63 Current Prioritization score: Class NUM_SECT Length condition Agriculture Health Climate Safety Traffic Overall RN N3_13 0.79 Fair 0.32 0.25 0.11 0.00 0.59 0.47 RN N3_14 3.17 Fair 0.32 0.25 0.14 0.33 0.59 0.61 RN N3_15 5.59 Fair 0.53 0.25 0.29 0.69 0.59 0.88 RN N3_16 0.12 Fair 0.53 0.25 0.30 0.55 0.59 0.83 RN N3_17 2.85 Fair 0.53 0.50 0.32 0.58 0.44 0.89 RN N3_18 2.83 Fair 0.00 0.25 0.00 0.33 0.44 0.38 RN N3_19 2.92 Good 0.00 0.00 0.00 0.00 0.44 0.16 RN N3_20 2.54 Fair 0.61 0.00 0.36 0.00 0.44 0.52 RN N3_21 3.13 Fair 0.69 0.25 0.37 0.00 0.44 0.65 RN N3_22 2.27 Fair 0.40 0.00 0.19 0.22 0.44 0.47 RN N3_23 4.16 Fair 0.00 0.00 0.00 0.33 0.44 0.29 RN N3_24 1.53 Poor 0.00 0.00 0.00 0.00 0.44 0.16 RN N3_25 2.86 Poor 0.38 0.50 0.31 0.22 0.44 0.69 RN N3_26 5.03 Poor 0.38 0.75 0.38 0.00 0.44 0.73 RN N3_27 1.89 Poor 0.55 0.25 0.58 0.00 0.44 0.68 RN N3_B 4.15 Very poor 0.55 0.00 0.78 0.00 0.61 0.72 RN N4_01 1.01 Poor 0.00 0.00 0.00 0.00 0.00 0.00 RN N4_02 0.56 Poor 0.19 0.00 0.17 0.00 0.66 0.38 RN N4_03 3.44 Good 0.23 0.00 0.06 0.22 0.66 0.44 RN N4_04 3.9 Fair 0.23 0.00 0.12 0.22 0.69 0.47 RN N4_05 2.31 Fair 0.43 0.00 0.23 0.33 0.69 0.62 RN N4_06 2.63 Poor 0.43 0.00 0.37 0.22 0.57 0.59 RN N4_07 3.31 Fair 0.00 0.00 0.00 0.22 0.57 0.29 RN N4_08 3.48 Fair 0.00 0.00 0.00 0.22 0.57 0.29 RN N4_A_01 1.71 Fair 0.00 0.00 0.00 0.00 0.57 0.21 RN N4_A_02 3.21 Fair 0.53 0.25 0.33 0.58 0.57 0.84 RN N4_B_01 1.01 Fair 0.35 0.25 0.23 0.00 0.61 0.54 RN N4_B_02 0.83 Fair 0.35 0.25 0.24 0.49 0.61 0.72 RN N4_B_03 3.46 Poor 0.35 0.25 0.36 0.58 0.61 0.80 RN N4_B_04 0.77 Poor 0.00 0.00 0.00 0.00 0.61 0.23 RN N5_01 2.63 Poor 0.36 0.25 0.29 0.00 0.54 0.54 RN N5_02 2.77 Fair 0.37 0.75 0.20 0.00 0.54 0.70 RN N5_03 3.68 Fair 0.00 0.50 0.00 0.00 0.54 0.39 RN N5_04 3.38 Fair 0.44 0.25 0.28 0.00 0.54 0.56 RN N5_05 3.36 Good 0.60 0.25 0.21 0.00 0.54 0.60 RR R101_01 0.77 Good 0.36 0.00 0.08 0.00 0.79 0.46 RR R101_02 2.11 Good 0.24 0.25 0.06 0.00 0.79 0.50 RR R101_03 2.4 Good 0.00 0.25 0.00 0.00 0.79 0.39 RR R101_04 2.28 Good 0.43 0.00 0.12 0.00 0.79 0.50 RR R102_01 1.13 Very poor 0.50 0.00 0.48 0.00 0.00 0.36 RR R102_02 1.99 Poor 0.50 0.00 0.38 0.00 0.00 0.33 RR R103_01 1.45 Poor 0.36 0.25 0.24 0.00 0.00 0.32 RR R103_02 5.34 Poor 0.58 0.00 0.41 0.00 0.00 0.37 RR R104 1.73 Very poor 0.00 0.00 0.00 0.00 0.00 0.00 RR R105 7.76 Poor 0.72 0.50 0.29 0.00 0.60 0.79 RR R107 4.31 Poor 0.71 0.25 0.11 0.00 0.72 0.67 RR R108_A_01 3.57 Poor 0.72 0.00 0.09 0.00 0.00 0.30 RR R108_A_02 1.63 Very poor 0.71 0.00 0.14 0.00 0.00 0.32 RR R109 6.04 Fair 0.56 0.00 0.39 0.00 0.00 0.36 RR R109_A_01 1.11 Good 0.46 0.00 0.16 0.00 0.00 0.23 RR R109_A_02 1.57 Fair 0.53 0.00 0.36 0.00 0.00 0.33 64 Current Prioritization score: Class NUM_SECT Length condition Agriculture Health Climate Safety Traffic Overall RR R109_B_01 1.49 Poor 0.41 0.00 0.40 0.00 0.00 0.30 RR R109_B_02 0.52 Good 0.46 0.00 0.15 0.00 0.00 0.23 RR R109_C 0.24 Poor 0.41 0.00 0.40 0.00 0.00 0.30 RR R110_01 0.95 Very poor 0.61 0.00 0.77 0.00 0.00 0.51 RR R110_02 1.22 Poor 0.44 0.00 0.42 0.00 0.00 0.32 RR R110_A 0.9 Very poor 0.44 0.00 0.54 0.00 0.00 0.36 RR R111 1.63 Poor 0.31 0.25 0.33 0.00 0.00 0.33 RR R111_01 0.36 Poor 0.00 0.00 0.00 0.00 0.00 0.00 RR R112 2.37 Poor 0.27 0.25 0.28 0.00 0.00 0.30 RR R113 1.23 Fair 0.46 0.00 0.30 0.22 0.00 0.37 RR R114 1.47 Poor 0.11 0.00 0.10 0.55 0.00 0.29 RR R115_01 2.44 Fair 0.35 0.25 0.21 0.40 0.00 0.45 RR R115_02 2.55 Poor 0.37 0.50 0.32 0.00 0.00 0.44 RR R116 6.76 Poor 0.62 0.50 0.50 0.00 0.65 0.85 RR R117 1.94 Poor 0.11 0.50 0.09 0.00 0.62 0.49 RR R118_01 1.33 Fair 0.00 0.00 0.00 0.00 0.00 0.00 RR R118_02 1.04 Fair 0.46 0.00 0.27 0.00 0.00 0.27 RR R118_03 1.23 Fair 0.46 0.00 0.27 0.00 0.00 0.27 RR R118_04 1.5 Fair 0.46 0.00 0.27 0.00 0.00 0.27 RR R118_05 2.24 Poor 0.00 0.00 0.00 0.00 0.00 0.00 RR R118_A_01 1.36 Fair 0.00 0.25 0.00 0.00 0.00 0.09 RR R118_A_02 1.52 Poor 0.00 0.25 0.00 0.00 0.00 0.09 RR R118_B 0.68 Poor 0.00 0.00 0.00 0.00 0.00 0.00 RR R119 3.99 Poor 0.42 0.50 0.04 0.00 0.00 0.36 RR R120_01 2.27 Very poor 0.68 0.00 0.05 0.22 0.00 0.35 RR R120_02 1.71 Poor 0.68 0.00 0.07 0.00 0.00 0.28 RR R120_B 2.5 Poor 0.68 0.00 0.05 0.22 0.00 0.35 RR R121_01 1.74 Poor 0.53 0.25 0.04 0.22 0.00 0.39 RR R121_02 1.02 Poor 0.53 0.25 0.05 0.00 0.00 0.31 RR R121_03 0.35 Poor 0.53 0.25 0.06 0.00 0.00 0.31 RR R121_B_02 1.04 Poor 0.53 0.25 0.05 0.00 0.00 0.31 RR R122_01 3.56 Poor 0.66 0.50 0.14 0.22 0.00 0.57 RR R122_02 5.59 Poor 0.53 0.25 0.09 0.00 0.00 0.32 RR R122_03 4.01 Poor 0.71 0.25 0.10 0.00 0.00 0.39 RR R122_04 1.06 Poor 0.71 0.25 0.11 0.00 0.00 0.40 RR R122_05 2.95 Poor 0.81 0.25 0.29 0.00 0.00 0.50 RR R122_06 2.07 Poor 0.78 0.25 0.31 0.00 0.00 0.50 RR R122_07 3.52 Poor 0.63 0.25 0.35 0.00 0.00 0.46 RR R123 2.22 Poor 0.42 0.25 0.08 0.00 0.00 0.28 RR R124_01 2.84 Poor 0.27 0.00 0.09 0.00 0.00 0.13 RR R124_02 3.13 Poor 0.60 0.25 0.28 0.00 0.00 0.42 RR R124_A_01 4.09 Poor 0.32 0.25 0.17 0.00 0.00 0.28 RR R124_A_02 1.18 Very poor 0.58 0.25 0.33 0.00 0.00 0.44 RR R124_A_03 1.62 Poor 0.00 0.25 0.00 0.00 0.00 0.09 RR R125_01 1.97 Poor 0.00 0.50 0.00 0.00 0.00 0.19 RR R125_02 1.12 Poor 0.00 0.00 0.00 0.00 0.00 0.00 RR R125_03 3.51 Poor 0.18 0.00 0.16 0.00 0.00 0.13 RR R125_04 2.39 Poor 0.45 0.25 0.44 0.00 0.00 0.43 RR R125_05 7.05 Poor 0.44 0.25 0.43 0.00 0.00 0.42 RR R125_A 0.49 Poor 0.36 0.25 0.30 0.00 0.00 0.34 RR R126_1 0.34 Very poor 0.19 0.00 0.23 0.00 0.00 0.16 65 Current Prioritization score: Class NUM_SECT Length condition Agriculture Health Climate Safety Traffic Overall RR R126_2 1.34 Fair 0.52 0.25 0.33 0.00 0.00 0.41 RR R127 4.82 Very poor 0.54 0.25 0.65 0.00 0.00 0.54 RR R127_A_1 1.71 Very poor 0.54 0.25 0.66 0.00 0.00 0.54 RR R127_A_2 1.28 Poor 0.53 0.50 0.49 0.00 0.00 0.57 RR R128E 4.7 Very poor 0.60 0.50 0.31 0.00 0.53 0.72 RR R128W_01 1.14 Fair 0.00 0.00 0.00 0.00 0.00 0.00 RR R128W_02 2.67 Fair 0.00 0.00 0.00 0.00 0.00 0.00 RR R129 4.01 Fair 0.32 0.25 0.11 0.00 0.00 0.25 RR R130 3.69 Poor 0.35 0.50 0.35 0.33 0.00 0.57 RR R131 5.9 Fair 0.77 0.50 0.41 0.00 0.00 0.63 RR R131_A 2.83 Poor 0.53 0.25 0.42 0.00 0.00 0.45 RR R132 2.46 Fair 0.00 0.25 0.00 0.22 0.00 0.18 RR R133_01 3.16 Poor 0.00 0.00 0.00 0.00 0.00 0.00 RR R133_02 2.12 Poor 0.00 0.00 0.00 0.00 0.00 0.00 RR R134_01 2.06 Very poor 0.51 0.25 0.66 0.00 0.00 0.53 RR R134_02 1.25 Very poor 0.51 0.50 0.64 0.00 0.00 0.62 RR R135 2.51 Good 0.41 0.00 0.13 0.00 0.00 0.20 RR R136 2.04 Good 0.45 0.50 0.12 0.52 0.00 0.60 RNC RNC01NO 0.94 Poor 0.00 0.00 0.00 0.00 0.00 0.00 RNC RNC02NO 0.66 Poor 0.00 0.25 0.00 0.00 0.00 0.09 RNC RNC03NO 0.34 Poor 0.00 0.00 0.00 0.00 0.00 0.00 RNC RNC05NE 1.65 Poor 0.42 0.50 0.08 0.00 0.00 0.37 RNC RNC06NE 2.61 Fair 0.00 0.25 0.00 0.22 0.00 0.18 RNC RNC07NO 0.85 Fair 0.00 0.00 0.00 0.00 0.00 0.00 RNC RNC08NO 2.81 Fair 0.71 0.25 0.12 0.00 0.00 0.40 RNC RNC09NE 1.05 Fair 0.00 0.00 0.00 0.00 0.00 0.00 RNC RNC10NO 1.22 Poor 0.58 0.00 0.15 0.00 0.00 0.27 RNC RNC11NE 0.88 Fair 0.58 0.50 0.16 0.00 0.00 0.46 RNC RNC12NE 0.72 Very poor 0.58 0.25 0.33 0.00 0.00 0.44 RNC RNC13NO 0.89 Fair 0.58 0.00 0.17 0.00 0.00 0.28 RNC RNC16CO 0.29 Fair 0.36 0.25 0.14 0.00 0.00 0.28 RNC RNC17NO 1.05 Fair 0.45 0.25 0.10 0.00 0.00 0.30 RNC RNC19CE 1.9 Poor 0.53 0.25 0.44 0.55 0.00 0.66 RNC RNC21CE 2.48 Poor 0.00 0.00 0.00 0.00 0.00 0.00 RNC RNC22CO 0.63 Fair 0.00 0.00 0.00 0.00 0.00 0.00 RNC RNC24CO 2.18 Fair 0.00 0.00 0.00 0.00 0.00 0.00 RNC RNC25_01CO 1.16 Poor 0.00 0.00 0.00 0.00 0.00 0.00 RNC RNC25_02CO 1.73 Poor 0.50 0.00 0.36 0.00 0.00 0.32 RNC RNC26CE 1.32 Poor 0.35 0.25 0.36 0.49 0.00 0.54 RNC RNC27CE 2.54 Poor 0.43 0.25 0.44 0.58 0.00 0.63 RNC RNC29CO 0.71 Poor 0.24 0.00 0.17 0.00 0.00 0.16 RNC RNC30CO 0.25 Fair 0.36 0.00 0.17 0.00 0.00 0.20 RNC RNC31CO 1.45 Poor 0.00 0.00 0.00 0.00 0.00 0.00 RNC RNC34CO 0.19 Fair 0.45 0.25 0.25 0.52 0.00 0.55 RNC RNC36CO 0.08 Fair 0.19 0.00 0.11 0.00 0.00 0.11 RNC RNC37SE 1.58 Poor 0.00 0.00 0.00 0.33 0.00 0.12 RNC RNC38SE 1.5 Poor 0.00 0.00 0.00 0.00 0.00 0.00 RNC RNC39CO 0.72 Poor 0.34 0.00 0.36 0.00 0.00 0.26 RNC RNC40SE 0.93 Poor 0.00 0.00 0.00 0.00 0.00 0.00 RNC RNC41CO 0.47 Poor 0.42 0.00 0.42 0.00 0.00 0.31 RNC RNC42CO 1.34 Poor 0.00 0.00 0.00 0.00 0.00 0.00 66 Current Prioritization score: Class NUM_SECT Length condition Agriculture Health Climate Safety Traffic Overall RNC RNC45SE 1.59 Fair 0.00 0.00 0.00 0.22 0.00 0.08 RNC RNC47S 0.55 Poor 0.18 0.00 0.16 0.00 0.00 0.13 RNC RNC48SE 0.4 Fair 0.35 0.25 0.24 0.49 0.00 0.49 RNC RNC49S 0.41 Poor 0.00 0.00 0.00 0.00 0.00 0.00 RNC RNC50CS 3.56 Poor 0.58 0.00 0.54 0.00 0.00 0.42 RNC RNC51SE 0.44 Fair 0.42 0.25 0.25 0.00 0.00 0.34 RNC RNC52S 0.59 Fair 0.00 0.00 0.00 0.00 0.00 0.00 RNC RNC53SE 0.64 Fair 0.00 0.00 0.00 0.00 0.00 0.00 RNC RNC54S 0.51 Fair 0.00 0.00 0.00 0.00 0.00 0.00 RNC RNC55S 2.27 Fair 0.25 0.25 0.13 0.22 0.00 0.32 RNC RNC57CO 0.83 Very poor 0.51 0.50 0.64 0.00 0.00 0.61 RN N21_01 2.44 Good 0.57 0.50 0.15 0.00 0.82 0.76 RN N21_02 2.96 Good 0.67 0.25 0.17 0.00 0.81 0.71 RN N21_03 2.51 Poor 0.65 0.25 0.50 0.00 0.76 0.81 RN N21_04 0.54 Poor 0.00 0.00 0.00 0.00 0.76 0.28 RN N21_05 7.62 Poor 0.61 0.25 0.55 0.00 0.76 0.81 RN N21_06 2.01 Fair 0.00 0.00 0.00 0.00 0.76 0.28 RN N21_07 6.26 Fair 0.64 0.25 0.42 0.00 0.76 0.77 RN N21_08 2.59 Good 0.64 0.25 0.20 0.00 0.66 0.65 RN N21_09 4.24 Fair 0.00 0.25 0.00 0.00 0.66 0.34 RN N21_10 0.77 Good 0.59 0.25 0.14 0.00 0.66 0.61 RN N21_A 1.26 Good 0.65 0.25 0.16 0.00 0.81 0.70 RN N22_01 4.8 Fair 0.57 1.00 0.39 0.00 0.72 1.00 RN N22_01a 3.05 Fair 0.00 0.00 0.00 0.00 0.72 0.00 RN N22_02 4.97 Fair 0.43 0.25 0.30 0.00 0.72 0.64 RN N22_03 5.85 Poor 0.63 0.25 0.61 0.00 0.72 0.82 RN N22_A 5.93 Poor 0.59 0.50 0.49 0.00 0.53 0.79 RN N23_01 12.36 Poor 0.72 0.50 0.76 0.00 0.48 0.92 RN N23_02 9.31 Fair 0.65 0.50 0.45 0.00 0.48 0.78 RN N23_03 1.87 Good 0.57 0.25 0.20 0.00 0.48 0.56 RN N23_04 3.35 Good 0.61 0.50 0.21 0.00 0.48 0.67 RN N23_05 9.6 Fair 0.43 0.25 0.30 0.00 0.59 0.58 RN N23_06 2.49 Fair 0.00 0.00 0.00 0.00 0.59 0.22 RN N23_07 2.96 Poor 0.78 0.25 0.67 0.00 0.59 0.85 RN N23_08 8.95 Poor 0.81 0.25 0.66 0.00 0.59 0.86 RN N23_09 0.86 Poor 0.59 0.00 0.41 0.00 0.59 0.59 RN N23_10 0.43 Fair 0.59 0.25 0.27 0.00 0.59 0.63 RN N23_A 3.75 Fair 0.45 0.25 0.28 0.00 0.00 0.36 RN N23_B 0.08 Fair 0.45 0.25 0.27 0.00 0.00 0.36 RN N24_01 2.33 Fair 0.48 0.25 0.28 0.00 0.00 0.37 RN N24_02 9.66 Good 0.59 0.50 0.17 0.00 0.00 0.47 RN N24_03 3.56 Very poor 0.00 0.00 0.00 0.00 0.00 0.00 RR R201_01 1.59 Good 0.53 0.50 0.14 0.00 0.00 0.44 RR R201_02 0.29 Good 0.53 0.50 0.14 0.00 0.00 0.44 RR R202A 2.26 Fair 0.84 0.50 0.51 0.00 0.00 0.69 RR R202S 3.61 Good 0.59 0.50 0.20 0.00 0.00 0.48 RR R202_01 2.61 Good 0.79 0.50 0.23 0.00 0.00 0.56 RR R202_02 3.16 Good 0.84 0.50 0.24 0.00 0.00 0.59 RR R202_03 2.54 Fair 0.69 0.00 0.36 0.00 0.00 0.39 RR R203_01 4.88 Fair 0.72 0.50 0.50 0.22 0.00 0.73 RR R203_02 6.57 Fair 0.68 0.00 0.43 0.33 0.00 0.54 67 Current Prioritization score: Class NUM_SECT Length condition Agriculture Health Climate Safety Traffic Overall RR R203_03 2.95 Poor 0.54 0.00 0.44 0.00 0.00 0.37 RR R203_04 3.06 Fair 0.41 0.25 0.21 0.00 0.00 0.32 RR R204 1.16 Fair 0.64 0.25 0.39 0.00 0.00 0.48 RR R204b 0.36 Fair 0.64 0.25 0.39 0.00 0.00 0.48 RR R205_01 1.47 Fair 0.00 0.00 0.00 0.00 0.00 0.00 RR R205_02 7.6 Fair 0.98 0.00 0.56 0.00 0.00 0.58 RR R205_03 5.84 Fair 1.00 0.00 0.53 0.00 0.00 0.57 RR R206 2.62 Good 0.65 0.25 0.16 0.00 0.00 0.39 RR R206_A 0.88 Very poor 0.59 0.25 0.54 0.00 0.00 0.51 RR R206_B 1 Fair 0.59 0.25 0.27 0.00 0.00 0.41 RR R207 3.3 Very poor 0.76 0.25 1.00 0.00 0.00 0.75 RR R208 1.24 Very poor 0.59 0.25 0.66 0.00 0.00 0.56 RR R209 0.92 Poor 0.00 0.25 0.00 0.00 0.00 0.09 RR R210_01 1.45 Very poor 0.53 0.25 0.59 0.00 0.00 0.51 RR R210_02 3.73 Very poor 0.56 0.50 0.60 0.00 0.00 0.62 RR R210_03 2.37 Very poor 0.44 0.25 0.44 0.00 0.00 0.42 RR R211_01 0.96 Fair 0.00 0.00 0.00 0.00 0.00 0.00 RR R211A 2.15 Poor 0.00 0.25 0.00 0.00 0.00 0.09 RR R212_01 0.62 Very poor 0.50 0.00 0.45 0.00 0.00 0.36 RR R212_02 1.08 Very poor 0.58 0.00 0.53 0.00 0.00 0.41 RR R212_03 1.05 Very poor 0.58 0.00 0.52 0.00 0.00 0.41 RR R213 3.36 Very poor 0.67 0.25 0.68 0.00 0.00 0.60 RNC SR201 1.22 Good 0.65 0.25 0.17 0.00 0.00 0.40 RNC SR202 1.17 Fair 0.00 0.00 0.00 0.00 0.00 0.00 RNC SR203 0.83 Fair 0.00 0.00 0.00 0.00 0.00 0.00 RNC SR204 0.88 Fair 0.53 0.00 0.37 0.00 0.00 0.34 RNC SR205 0.6 Fair 0.00 0.00 0.00 0.00 0.00 0.00 RNC SR206 0.89 Fair 0.37 0.25 0.25 0.00 0.00 0.32 RNC SR207 1.01 Poor 0.00 0.00 0.00 0.00 0.00 0.00 RNC SR208 0.29 Poor 0.53 0.50 0.42 0.00 0.00 0.54 RNC SR208B 0.94 Good 0.53 0.50 0.14 0.00 0.00 0.44 RN N31_01 1.39 Good 0.56 0.50 0.20 0.60 0.71 0.96 RN N31_02 1.03 Good 0.23 0.25 0.08 0.64 0.71 0.71 RN N31_03 2.49 Good 0.31 0.00 0.10 0.77 0.61 0.67 RN N31_03b 5.68 Good 0.38 0.25 0.12 0.60 0.61 0.73 RN N31_04 2.45 Good 0.41 0.50 0.12 0.73 0.61 0.88 RN N31_05 4.17 Poor 0.29 0.25 0.26 0.84 0.41 0.76 RN N31_06 3.67 Poor 0.00 0.00 0.00 1.00 0.41 0.53 RN N31_07 8.83 Poor 0.57 0.25 0.46 0.75 0.41 0.91 RN N31_08 0.79 Poor 0.56 0.25 0.45 0.33 0.41 0.75 RN N32_01 2.68 Poor 0.55 0.50 0.58 0.22 0.65 0.93 RN N32_02 3.45 Poor 0.38 0.00 0.40 0.22 0.65 0.61 RN N32_03 0.2 Poor 0.40 0.00 0.42 0.00 0.65 0.55 RN N32_04 1.57 Poor 0.24 0.25 0.25 0.00 0.65 0.52 RN N32_05 1.35 Poor 0.24 0.25 0.25 0.33 0.65 0.64 RN N32_05b 5.07 Fair 0.45 0.25 0.31 0.74 0.65 0.89 RN N32_05c 4.48 Poor 0.42 0.00 0.40 0.69 0.65 0.80 RN N32_05d 4.22 Fair 0.00 0.00 0.00 0.73 0.65 0.51 RN N32_05e 2.79 Fair 0.34 0.25 0.23 0.83 0.65 0.86 RN N32_06 3.27 Fair 0.34 0.25 0.24 0.66 0.44 0.72 RN N32_06a 2.35 Fair 0.35 0.00 0.25 0.90 0.44 0.73 68 Current Prioritization score: Class NUM_SECT Length condition Agriculture Health Climate Safety Traffic Overall RN N32_06b 2.88 Poor 0.52 0.25 0.54 0.87 0.44 0.98 RN N32_07 1.34 Poor 0.47 0.25 0.45 0.49 0.44 0.79 RN N32_08 1.96 Poor 0.47 0.25 0.45 0.49 0.39 0.76 RN N32_09 4.96 Poor 0.47 0.25 0.44 0.58 0.39 0.79 RN N32_10 2 Poor 0.56 0.25 0.48 0.60 0.39 0.85 RR R301 1.01 Very poor 0.34 0.25 0.37 0.22 0.57 0.66 RR R302_01 2.85 Poor 0.45 0.25 0.42 0.58 0.64 0.87 RR R302_02 4.14 Fair 0.39 0.00 0.27 0.00 0.34 0.37 RR R302_03 4 Poor 0.36 0.25 0.38 0.00 0.34 0.50 RR R302_A_01 1.93 Very poor 0.57 0.25 0.71 0.00 0.32 0.69 RR R302_A_02 1.5 Poor 0.54 0.25 0.50 0.00 0.32 0.60 RR R303 1.11 Fair 0.24 0.25 0.17 0.33 0.00 0.37 RNC RNCCN 0.59 Good 0.00 0.00 0.00 0.00 0.00 0.00 RNC RNCCS 0.7 Very poor 0.56 0.25 0.60 0.00 0.00 0.53 RNC RNCSO 0.39 Fair 0.47 0.25 0.30 0.40 0.00 0.53 RNC RNCSO 0.39 Fair 0.47 0.25 0.30 0.40 0.00 0.53 69 Appendix B. Road Condition Data Road condition data in Comoros are limited. The European Commission (2015) conducted one of the most detailed road inventory surveys in 2014 and found that 57 percent of classified roads were in poor or very poor condition (Figure B.1). In 2018, the World Bank (2019c) conducted a geographical analysis to examine the country’s transport constraints and growth potential, in which the 2014 data were updated manually based on recent road programs that had been implemented or were under implementation. In April 2019, Cyclone Kenneth devastated the road network. A high-level post-disaster visual assessment of the road network found that approximately 90 km of roads, or more than 10 percent of the total network, was damaged, so the share of roads assessed as being in poor and very poor condition was increased to 49 percent.10 Figure B.1. Road condition, 2014-2021 Sources: European Commission 2015; World Bank surveys. To update the road condition data with greater granularity, the World Bank conducted a road condition survey using a mobile phone-based application, RoadLab, in 2021. Thus, there are at least four sets of road condition data from 2014 to 2021 (FigureB.2). The granularity, accuracy, and underlying methodologies of these data are different, but the trend in road conditions has been consistent. The World Bank designed RoadLab, a free application available from the Google Play Store, in collaboration with Beldor Center, Softteco, and Progress Analytics LLC. The app runs on all Android smartphones and tablets. Similar to other road assessment applications, which use accelerometers in smartphones, this app automatically records road conditions, such as roughness, speed, and GPS coordinates of the starting and ending points of each road segment (e.g., at 100-m intervals). All that is needed is to have the RoadLab app running on a phone while driving. Users can also manually link other types of data: road safety hazards, pictures of potholes, black spots, road crashes. When an Android device is connected, data can be exported in the form of comma-separated values, which can be converted and used in other types of software, such as ArcGIS. Measured roughness (International Roughness Index (IRI)), is an estimate based on related parameters, such as vehicle speed, vehicle suspension type, vertical acceleration, and phone position. There may be measurement and statistical errors in the output data, because the app is designed to calculate roughness based on the tested regression equation. In the pilot project, it was confirmed that the roughness estimated by the smartphone app correlated with the actual IRI at a reasonable significance level (correlation = 0.57). 10 The underlying road condition data were collected based on different methodologies. Thus, the figures should be compared with caution. 70 Figure B.2. Road conditions, 2014-2021 2014 2018 2019 after Kenneth 2021 Sources: European Commission 2015; World Bank surveys. In Comoros, 594 km of 815 km of official roads were surveyed; roughness was measured for 449 km. Where the roads were not assessed this time, road conditions are assumed to be the same as in 2018. Because of the statistical confidence, the app is designed to be turned on only when the vehicle reaches 30 km per hour. When the app is not activated, road conditions are most likely to be so poor that the vehicle cannot go fast. Based on RoadLab’s algorithm (World Bank 2015), the IRI is estimated depending on measured speed (SP) and the standard deviation of the accelerometer (SD): 1.879 + 5.852 ∗ 𝑆𝐷 − 0.0114 ∗ 𝑆𝑃 − 0.0393 ∗ (𝑆𝐷 ∗ 𝑆𝑃) 𝑖𝑓 𝑓𝑖𝑥𝑒𝑑, ℎ𝑎𝑟𝑑 2.511 + 6.177 ∗ 𝑆𝐷 − 0.0338 ∗ 𝑆𝑃 − 0.0111 ∗ (𝑆𝐷 ∗ 𝑆𝑃) 𝑖𝑓 𝑛𝑜𝑡, ℎ𝑎𝑟𝑑 𝐼𝑅𝐼 = { 1.258 + 5.215 ∗ 𝑆𝐷 − 0.0285 ∗ 𝑆𝑃 − 0.0207 ∗ (𝑆𝐷 ∗ 𝑆𝑃) 𝑖𝑓 𝑓𝑖𝑥𝑒𝑑, 𝑚𝑒𝑑𝑖𝑢𝑚 1.4036 + 6.903 ∗ 𝑆𝐷 − 0.0236 ∗ 𝑆𝑃 − 0.0303 ∗ (𝑆𝐷 ∗ 𝑆𝑃) 𝑖𝑓 𝑛𝑜𝑡, 𝑚𝑒𝑑𝑖𝑢𝑚 Fixed indicates whether a mobile device is firmly fixed, and hard or medium indicates the strength of the vehicle’s suspension. (See World Bank (2015) for more detail.) Because estimated IRI based on the algorithm was still found to be biased, depending on island, the thresholds of 71 the road assessment are determined taking into account the differences in the vehicles used and driving behavior across the islands (Table B.1). The latest post-Kenneth visual assessment was used, which is available only at an aggregate level. Finally, by using the interval regression technique, measured IRI is adjusted to a common scale of roughness for all three counties. Table B.1. Thresholds of road conditions Condition Island Good Fair Poor Very poor Grande Comore < 4 [4, 5] [5, 6] >6 Anjouan < 2.5 [2.5, 3] [3, 3.5] > 3.5 Moheli < 11 [11, 13.5] [13.5, 16] > 16 Common scalea < 3.5 [3.5, 5.5] [5.5, 10.5] > 10.5 a. This is consistent with default values in the Road Network Evaluation Tools model. 72 Appendix C. RONET Analysis This report is a strategic evaluation of the Comoros road network using the Road Network Evaluation Tools (RONET) developed by the World Bank. 11 The objective of the evaluation is to assess rehabilitation and maintenance project requirements and effects of budget constraints on infrastructure, the economy, and road users. The evaluation was based on road network inventory, condition, and traffic data estimated in 2021 to determine needs for 2022 to 2026. Because of the lack of availability of refined road network data, the results of this evaluation should be considered a rough estimate of the needs of the network. RONET is an Excel-based model for assessing the performance of road maintenance and rehabilitation policies and the importance of the road sector for the economy. It assesses current network condition, computing the asset value of the network and road network monitoring indicators. It uses country-specific relationships between maintenance spending and road condition and between road condition and road user costs to assess the performance over time of the network under different road project standards. It also determines proper allocation of expenditures among recurrent maintenance, periodic maintenance, and rehabilitation. Finally, RONET determines the funding gap, defined as the difference between current preservation spending and required spending and the effect of underspending on transport costs. The RONET Performance Assessment module assesses the consequences of various budget scenarios that represent different levels of road preservation expenditures over time. The consequences for road project requirements, financial costs, road conditions, asset value, road user costs, and total transport costs are presented. This module evaluates the performance of each road class of the network under different road project standards over a 20-year period. Road classes are characterized according to functional classification, surface type, traffic, and condition. The very high standard represents a policy without budget constraints and a high frequency of road projects. The high, medium, low, and very low standards represent decreasing frequencies of road projects and corresponding annualized road project expenditures. The do-minimum standard represents a policy under which the only capital road work performed over the evaluation period is reconstruction at a very high roughness. RONET calculates, for all standards applied to a road class, the corresponding stream of road agency costs (rehabilitation and maintenance), road user costs, and total transport costs (sum of road agency and user costs) over the evaluation period. RONET them computes the present value of total transport costs for each standard, at the given input discount rate, and determines the optimal standard for the road class, which is the one that yields the lowest present value of total transport costs. Once the performance of each road class is evaluated under different standards, RONET presents the results for a series of budget scenarios. The optimal scenario applies the optimal standard for each road class to all road classes in the network. Once RONET determines the optimal scenario, it evaluates other budget scenarios using lower road agency expenditures than those needed for the optimal scenario. Road Network Data The evaluation was based on Comoros’ road network traffic data from 2013, recent multicriteria analysis traffic data, and road condition data (roughness) collected in 2020/21, using the RoadLab Pro Cellphone App that the World Bank developed, which estimates road roughness. The rural road network consists of 761 km of paved roads divided into national roads, regional roads, and unclassified roads (Table C.1). In addition, there are 54 km of urban roads that were not included on this study. 11 http://www.ssatp.org/en/page/road-network-evaluation-tools-ronet 73 Table C.1: Network length according to network type Road Length Percent Type (km) (%) National 404.0 53.1 Regional 296.0 38.9 Unclassified 61.0 8.0 Total 761.0 100.0 Roads were classified into road condition classes based on their estimated roughness (Table C.2). Table C.2: Road classification according to International Roughness Index Road condition International Roughness Index (m/km) Very good <2.5 Good 2.5–3.5 Fair 3.5–5.5 Poor 5.5–10.5 Very poor >10.5 Table C.3 shows the network distribution according to roughness range (condition of pavement); 49 percent of the road network is in sustainable condition (very good, good, fair). Table C.3: Network length according to network type and roughness Road type International Roughness Index (m/km) Total < 2.5 (very 2.5–3.5 3.5–5.5 5.5–10.5 >10.5 (very good) (good) (fair) (poor) poor) km National 44.0 31.8 162.6 155.1 10.5 404.0 Regional 2.6 25.1 78.6 142.7 46.9 296.0 Unclassified 0.0 2.8 25.4 30.6 2.3 61.0 Total 46.7 59.7 266.7 328.3 59.7 761.0 % National 10.9 7.9 40.3 38.4 2.6 100.0 Regional 0.9 8.5 26.6 48.2 15.9 100.0 Unclassified 0.0 4.5 41.7 50.1 3.7 100.0 Total 6.1 7.8 35.0 43.1 7.8 100.0 Table C.4 shows the network distribution according to average annual daily traffic (vehicles per day). Traffic is low, with 90 percent of the road network carrying fewer than 1,000 vehicles per day and 60 percent carrying fewer than 300. 74 Table C.4: Network length according to network type and traffic Road type Annual average daily traffic (vehicles per day) Total <300 300–1,000 1,000–3,000 >3,000 km National 132.8 199.7 61.9 9.6 404.0 Regional 263.6 24.8 7.6 0.0 296.0 Unclassified 61.0 0.0 0.0 0.0 61.0 Total 457.4 224.5 69.6 9.6 761.0 % National 32.9 49.4 15.3 2.4 100.0 Regional 89.1 8.4 2.6 0.0 100.0 Unclassified 100.0 0.0 0.0 0.0 100.0 Total 60.1 29.5 9.1 1.3 100.0 Table C.5 shows network distribution according to terrain type; 92 percent is flat or hilly (vs mountainous) terrain. Table C.5: Network length according to road and terrain type Road type Terrain Total Flat Hilly Mountainous km National 254.5 118.9 30.6 404.0 Regional 172.7 93.3 30.0 296.0 Unclassified 55.5 5.5 0.0 61.0 Total 482.8 217.6 60.6 761.0 % National 63.0 29.4 7.6 100.0 Regional 58.4 31.5 10.1 100.0 Unclassified 91.1 8.9 0.0 100.0 Total 63.4 28.6 8.0 100.0 The RONET evaluation considered network length distribution according to road type (national, regional, unclassified), traffic (average annual daily traffic—vehicles per day), and condition (very good, good, fair, poor, very poor) function of roughness ranges. Table C.6 show the distribution of the network according to network type, traffic, and condition. The road network carries approximately 160 million vehicle-km per year, of which 78 percent travel on national roads that represent 53 percent of the network length. Average annual daily traffic is approximately 576 vehicles per day, of which an average of 848 use national roads and 284 use regional roads. Fifty-nine percent of national roads and 36 percent of regional roads are in sustainable condition (very good, good, fair). Average network roughness, weighted per vehicle-km, is 5.3 m/km (Table C.7). National roads are in better condition, with an average roughness of 5.0 m/km. 75 Table C.6: Road network evaluated using RONET (km) Average Road type annual daily International Roughness Index (m/km) Total traffic < 2.5 IRI 2.510.5 National < 300 3.3 17.3 47.5 61.2 3.5 132.8 300–1,000 17.3 9.0 86.9 79.5 7.0 199.7 1,000–3,000 23.5 1.4 22.8 14.3 0.0 61.9 3,000–10,000 0.0 4.0 5.6 0.0 0.0 9.6 Total 44.0 31.8 162.6 155.1 10.5 404.0 Regional < 300 2.6 17.5 78.6 118.9 45.9 263.6 300–1,000 0.0 0.0 0.0 23.7 1.0 24.8 1,000–3,000 0.0 7.6 0.0 0.0 0.0 7.6 3,000–10,000 0.0 0.0 0.0 0.0 0.0 0.0 Total 2.6 25.1 78.6 142.7 46.9 296.0 Unclassified < 300 0.0 2.8 25.4 30.6 2.3 61.0 300–1,000 0.0 0.0 0.0 0.0 0.0 0.0 1,000–3,000 0.0 0.0 0.0 0.0 0.0 0.0 3,000–10,000 0.0 0.0 0.0 0.0 0.0 0.0 Total 0.0 2.8 25.4 30.6 2.3 61.0 Table C.7: Comparison of network types Road type Length Use Average annual daily traffic International Roughness Indexa (Million Very good, good, fair (km) (%) vehicle-km) (%) Vehicles/day (m/km) (%) National 404 53 125 78 848 5.0 59.0 Regional 296 39 31 19 284 6.7 35.9 Unclassified 61 8 4 3 200 6.5 46.2 Total 761 100 160 100 576 5.3 49.0 a. Weighted per vehicle-km Unit Road Preservation Project Costs Table C.8 presents estimated average unit costs of capital projects, according to road class. Capital projects to be applied on a road are a function of current road condition. According to the RONET logic, roads in good or fair condition (sustainable network) require periodic maintenance, and roads in poor or very poor condition require rehabilitation. Unit costs are based on the table “Tâbleau Recapitulatif des Travaux Routiers Realises de 2016 a 2020� and the document “Projet: Projet de Rehabilitation du Reseau Routier -Projet De Rehabilitation de la RN2 (Moroni-Foumbouni): Troncon Moroni Ouroveni�. Routine maintenance costs were estimated to be USD6,000/km per year. RONET was not used to optimize routine maintenance project requirements of the road network. RONET optimizes only capital road projects (periodic maintenance and rehabilitation). Table C.8: Unit costs of road projects Unit cost Road condition Road project class Road project type (USD/km) Good Periodic maintenance Preventive treatment 15,000 Fair Resurfacing (overlay) 150,000 Poor Rehabilitation Strengthening (overlay) 400,000 Very Poor Reconstruction 600,000 Note: Routine maintenance=USD6,000/km per year 76 Vehicle Fleet Characteristics Table C.9 presents estimated 2021 vehicle fleet characteristics and economic unit costs used in the RONET evaluation. The traffic growth rate was set to 4.0 percent per year, considering that the International Monetary Fund predicts that Comoros’ gross domestic product will grow at 4.0 percent per year from 2022 to 2025. From 2017 to 2020, fuel consumption in SONELEC has increased on average by 3.8 percent per year. The RONET evaluation considered that typical traffic composition is 50 percent cars and four-wheel drive vehicles, 11 percent trucks, and 39 percent buses. Table C.9: Vehicle Fleet Characteristics Pass. Pass. Non- New New Lubric. Maint. Crew Annual Work Work Vehicle Vehicle Tire Fuel Oil Labor Wages Overhead Time Time Description ($/vehicle) ($/tire) ($/liter) ($/liter) ($/hour) ($/hour) ($/year) ($/hour) ($/hour) Car Medium 20,000 100 1.09 4.00 4.00 0.00 0 4.00 1.00 Four-Wheel Drive 20,000 100 1.09 4.00 4.00 2.00 0 2.00 0.25 Truck Light 30,000 300 1.09 4.00 4.00 2.00 0 2.00 0.25 Truck Heavy 50,000 350 1.09 4.00 4.00 2.00 0 2.00 0.25 Bus Medium 30,000 200 1.09 4.00 4.00 2.00 0 2.00 0.25 Annual Annual Number Work Gross km Working Service Private of Pass. Vehicle Annual Cargo Vehicle Driven Hours Life Use Passengers Trips Weight Interest Time Description (km) (hours) (years) (%) (#) (%) (t) (%) ($/hour) Car Medium 18,000 450 10 100 2 75 1.20 6 0.00 Four-Wheel Drive 35,000 875 10 0 1 75 2.00 6 0.00 Truck Light 25,000 625 9 0 0 0 6.00 6 2.18 Truck Heavy 38,000 950 10 0 0 0 16.00 6 2.18 Bus Medium 50,000 1,250 9 0 20 75 6.00 6 0.00 77 Appendix D. Wider Socioeconomic Benefits from Transport Connectivity Greater transport access can bring a wide range of economic and social benefits in the short and long term (Figure D.1). Greater transport connectivity can lower transport costs and time, facilitating efficient mobility and creating more jobs and thereby alleviating poverty. In the short term, transport costs and travel time are often found to be reduced (e.g., Danida 2010; Lokshin and Yemtsov 2005). Access to facilities such as schools and hospitals also increases (e.g., Khandker, Bakht, and Koolwal 2009; Mu and van de Walle 2011). In the longer term, agricultural productivity increases, firms become more profitable, more jobs are created (e.g., Bell and Van Dillen 2012), and poverty is alleviated (Dercon et al. 2008). See World Bank (2013) for a more detailed discussion. Figure D.1. Results Chain Access to Access to Income and Physical Access to Local market education and input/ output welfare connectivity labor market development health markets increase 1-2 years 3-5 years 5-10 years Source: World Bank 2013. Rural Accessibility and Poverty Reduction Rural accessibility is one of the most important constraints on developing countries. The poor tend to suffer more from limited road accessibility. Globally, rural accessibility is highly correlated with poverty incidence; although causality is complex, correlation is high (~0.7) (Figure D.2). The poverty rate in Comoros was estimated to be 36 percent in 2014 (Haazen and Rajoela 2016). According to the conventional definition of 2 kmdistance from a road, the Rural Access Index (RAI) (Box 4) is 65.3 percent, which is relatively high according to regional standards. Small countries tend to have high rural accessibility because of high population density. Figure D.2. Rural Access Index (RAI) and poverty incidence in Sub-Saharan African countries Source: World Bank estimates. 78 Box 4. Measuring Rural Accessibility Using New Technology The Rural Access Index (RAI) is one of the most important global indicators of connectivity in the transport sector. Roberts, Shyam, and Rastogi (2006) developed the original RAI, which used household surveys, combined with some modeling when data were not available. It measures the share of rural population with access to an all-season road within an approximate 2-km walking distance. Although how to measure it is not always clear, according to the original RAI, 1 billion people (31 percent of the total rural population) do not have access to the road network globally. In Africa, 450 million people, or more than 70 percent of the rural population, are estimated to be unconnected. This compares unfavorably with many other developing countries. For instance, the RAI is estimated at 90 percent in East Asia. Despite its important contribution to transport operations, the original RAI has several limitations. First, it is difficult to update data in a timely manner, because household surveys are costly and are not available in all countries. Second, because it uses household surveys, geographic representativeness is not perfect, and the resultant index is available only at the national level and therefore has little operational relevance. The estimated RAI cannot be used to infer accessibility at a given location. In addition, the original RAI may or may not reflect actual efforts to construct or rehabilitate roads because of its incomplete geographic representation. If the sample of a household survey does not cover a project area, the index does not change even though accessibility increases. To overcome these disadvantages, a new RAI is being developed using new technologies, such as geo- referenced road network data and highly disaggregated population distribution data (e.g., WorldPop). With the quality of roads taken into account, the number of rural residents who live within 2 km of the nearest road in good condition is estimated.a This new RAI is expected to contribute to the ongoing discussion of the Sustainable Development Goals. For Comoros, the original RAI was estimated at 73 percent in 2006. This was inferred from a modeling approach. No household survey was used. It was estimated at 51 percent in 2014, 72 percent in 2018, and 65 percent in 2021—all lower than the previous estimate because the new RAI takes the quality of roads into account more explicitly. An advantage of the new RAI is that it can measure accessibility at a highly disaggregated level, such as according to district. It is useful for investment decisions. a. A road in good condition is tentatively defined as a paved road in good or fair condition or with an International Roughness Index (IRI) less than 5 or an unpaved road in good condition or with an IRI less than 10. Poverty is particularly persistent in Moheli and remote areas of Anjouan and Grande Comore. The poverty rate is highest in Moheli (53 percent) and lowest in Moroni (30 percent) (Figure D.3). According to division, poverty is highest in Nioumamchoi and Djando in Moheli, followed by Muremani in Anjouan and Hamahamet Mboikou in Grande Comore (Figure D.4). Poverty is relatively low around the national and regional capitals (Moroni, Mutsamudu, Fomboni). As expected, poverty tends to be high where rural accessibility is poor (Figure D.5). This merely shows correlation, not causality, although poverty is high where people do not have access to the road network. To alleviate poverty, it is essential to ensure rural accessibility, especially in remote areas. 79 Figure D.3. Poverty incidence according to island, 2014 Figure D.4. Poverty incidence according to division, 2014 Source: Haazen et al. 2016. Source: Based on Haazen et al. 2016. Figure D.5. Rural accessibility and poverty incidence according to town Sources: Haazen et al. 2016; World Bank estimates. Rural accessibility increased between 2014 and 2018 but has deteriorated since then. It may not be possible to compare results directly because data and methodologies differ, but the RAI improved significantly from 2014 (51.5 percent) to 2018 (72.0 percent) (Figure D.6). As discussed above, some of the major national roads were rehabilitated, along which people benefited from greater transport accessibility. It dropped slightly in 2019, to 65.3 percent, because of the devastating impact of Cyclone Kenneth and lack of road maintenance. All these estimates are lower than the original RAI in 2006, which estimated Comorian accessibility at 73 percent, but it was calculated based on statistical modeling without any actual data. 80 Figure D.6. Rural Access Index according to region Source: World Bank estimates. According to region (or island), Grande Comore has the lowest RAI (51 percent), followed by Moheli (77 percent) and Anjouan (78 percent). This may seem counterintuitive, but the RAI is concentrated on rural areas. In Grande Comore, many areas where the road condition is good and thus accessibility is high are excluded from the calculation. There is substantial heterogeneity in rural accessibility within each island. Accessibility increased in the northern divisions in Grande Comore and Anjouan, where road rehabilitation projects were implemented in recent years (Figure D.7), whereas it deteriorated on the eastern side of Grande Comore and the southern part of Moheli, largely because of Cyclone Kenneth. Rural accessibility is highly sensitive to climate events. Although the above- mentioned market accessibility is measured with market size taken into account, the RAI is a more direct measurement, depending on the condition of the nearest road. Thus, when road conditions worsen, estimated accessibility can drop sharply, as observed in Nioumachoua, Moheli. Figure D.7. Rural Access Index according to division 2014 2021 Source: World Bank estimates. 81 Affordability and Public Transportation In general, demand for transport services increases with household income. Transportation is a normal good, meaning that people spend more on transportation as they become richer. The global experience supports this (Figure D.8). Comoros follows the same spending pattern. According to a recent household expenditure survey (Enquête sur les Dépenses de Consommation des Ménages aux Comores) in 2014, approximately 6 percent of poor people’s (first quantile) and more than 8 percent of rich people’s (fifth quantile) total spending is on transportation (Figure D.9). Figure D.8. Household spending on transportation according to quantile in African countries Sources: Lozano-Gracia and Young 2014. Figure D.9. Household spending on transport in Comoros Sources: World Bank estimates based on 2014 EDMC. In Comoros, the same trend is observed at the national level: The rich spend more on transportation as a whole, although the poor spend more on land transportation in some regions, such as Moroni and Moheli. There is a clear difference in transport modes that people choose. For interisland movement, poorer people use maritime transport more, and richer people use air transport more. Poor people spend more on land transportation services, particularly in Moroni and Moheli. There is a significant difference between regions. In Moroni, the poorest spend the most on land transportation, which accounts for 9.4 percent of their total consumption (Figure D.10). In Moheli, 3.3 percent of total spending of poor people is on land transportation, which is much 82 higher than spending by the richest (2.5 percent). In Moheli, therefore, transport infrastructure connectivity is generally poor, and the poor struggle more with relatively high transportation costs. Figure D.10. Household spending on transport according to region Moroni Other areas in Grande Comore Anjouan Mohel Sources: World Bank estimates based on 2014 EDMC. Better public transportation services could particularly benefit the poor. In Comoros, the public transport sector is active and well organized. There are 29 transport service operators connecting major cities and towns to the regional capitals. The network covers the entire country (Figure D.11). Approximately 500 buses are operating (Figure D.12). Based on normal operational cycles (before the COVID crisis),12 it is estimated that approximately 36,000 passengers use a public bus every day, or 13 million passengers per year. Although bus operations are well organized by cooperatives, fares are high. Average minibus fares range from KMF500 to KMF700 (~USD1.50) per passenger, depending on destination. When people take a bus every weekday, this amounts to approximately USD1,400 per year, or 26 percent of per capita gross domestic product (GDP) (Figure D.13). 12 Average bus capacity is 17 passengers. 83 Figure D.11. Minibus operating routes and travel time to regional capitals Source: World Bank estimates. Figure D.12. Minibus operations in Comoros Figure D.13. Bus fares and vehicle operating costs in Grande Comore Source: World Bank 2019c. Source: World Bank 2019c. Bus operations may not be efficient, particularly in Anjouan and Moheli, partly because of poor road conditions. Average operating speed of public buses is 51 km per hour In Grande Comore, 18 km per hour in Anjouan, and 13 km per hour in Moheli (Figure D.14), partly because of insufficient demand, because bus operators tend to wait until seats are filled, but also because of difficult terrain and poor road conditions. To increase the reliability and efficiency of bus operations, the main national roads must be maintained in good condition. 84 Figure D.14. Daily ridership and operating speed Sources: World Bank 2019c. Market Accessibility and Inter- and Intra-Island Connectivity Market accessibility is one of the most important aspects of improved road networks, bringing more economic opportunities. When the road network is improved, it is generally expected that more people and firms will have better access to economic opportunities. The literature is supportive of this: When rural roads were improved in Bangladesh, agricultural output prices increased between 2 percent and 22 percent (Khandker, Bakht, and Koolwal 2009). In Orissa, India, output prices increased 5 percent more in villages receiving all-weather roads than comparison villages (Bell and Van Dillen 2012). Better roads can open business opportunities. New services, such as women’s hair dressing, became available after a road rehabilitation project in Vietnam (Mu and van de Walle 2011). Market access generally increases with size of purchasing power or market capacity and decreases with degree of impediment between origin and destination. In practice, the Market Access Index is often defined as the sum of purchasing power or market capacity, y, inversely weighted by the degree of impediment, d, between two locations, i and k:13 𝑀𝐴𝐼𝑖 = (∑ 𝑦𝑘 �𝑑𝑖𝑘 )�max 𝑀𝐴𝐼𝑗 𝑘 𝑗 In Comoros, market accessibility differs considerably according to region. Estimated transport costs composed of vehicle operating costs and time costs are used to measure the impediment, d, from each location to a populated area. For each road section, vehicle operating and time costs are computed based on observed road conditions. Because of its low population density, lack of major cities, and poor condition of roads, Moheli is considered to be least connected to the domestic market (Figure D.15). In Grande Comore, market accessibility is relatively high along National Road 1 (RN1), RN2, and RN4 connecting Moroni and Kombani. Although Moroni and Bambao divisions are well connected, the northern and southern areas of the island are less connected to the market. In Anjouan, connectivity is good along RN21 connecting Mutsamudu to Bambao and Domoni, where many people are concentrated in the island. These regional inequalities remained unchanged from 2014 to 2021. Moroni is the most important primary market in Comoros. 13 This is based on a conventional gravity framework. The literature also uses the negative exponential weights. See for instance, Zeufack, Elbadawi, and Mengistae (2006) and Lall and Mengistae (2005). 85 Figure D.15. Market Access Index 2014 2021 Source: World Bank estimates. The increase in market accessibility has been heterogeneous and marginal. Market accessibility increased in the northern divisions in Grande Comore and many areas in Anjouan (Figure D.16), primarily because of recent investments in major national roads (e.g., RN1 from Moroni to the airport and Mitsamiouli, RN21 between Sima and Moya) financed by international partners and the Comorian government (Figure D.17). On Moheli, market accessibility has remained largely unchanged or has deteriorated despite some improvement in road condition between Fomboni and Miringoni, partly because of low population density on Moheli. By definition, the Market Access Index increases with the size of the market or the number of people, holding everything else constant. Moheli is least connected to the national market and has benefited least from recent road programs. Figure D.16. Changes in Market Access Index Figure D.17. Recent and ongoing road projects Source: World Bank estimates. Because of the poor road network and costly, unreliable interisland connectivity, many local farmers do not 86 have access to markets. Approximately 20 percent to 30 percent of major food crops are sold in the market. Cash crops, such as tomatoes and chili peppers, are more often sold. Despite Anjouan’s significant agricultural potential, market transactions are particularly limited (Figure D.18), largely because of lack of reliable, affordable interisland connectivity. The primary market, Moroni, is far from many parts of the country, especially Anjouan and Moheli. Limited connectivity results in inefficiency of the domestic crop market. In Comoros, price differentials are substantial. In general, crop prices in Moroni are more than twice as high as those in Anjouan and Moheli (Figure D.19). High commodity prices, amplified by market inefficiency, are placing an additional burden on urban consumers. Comoros imports significant agricultural crops and products every year. In 2016, approximately USD110 million worth of crops and food products was imported—approximately 18 percent of the country’s GDP. Figure D.18. Share of market sales out of production Figure D.19. Crop price differences between islands Source: World Bank estimates based on 2014 EDMC. Source: Iimi 2019. inter- and intra-island transport-related costs accounted for approximately 10 percent to 30 percent of these price differences. For example, the price for 25 kg of cassava is KMF5,000 in Nioumachoua and KMF15,000 in Moroni (Figure D.20). Land transport costs account for KMF4,000—approximately 27 percent of the final price. Interisland transport costs are KMF1,500. Similarly, land transportation costs increase banana prices at Moroni by about 16 percent. On Anjouan, crop prices are twice as much at the market as at the farm. Land transport costs account for 15 percent of banana prices at Mutsamudu (Figure D.21). Profit margins for retailers are high in all cases. The limited supply of agricultural commodities allows them to set their retail prices high. Figure D.20. Value chain analysis of cassava (25kg) Niumachoua, Moheli to Volo Volo, Moroni Dzindri to Mutsamudu, Anjouan Source: World Bank survey. 87 Figure D.21. Value chain analysis of bananas (50 kg) Niumachoua, Moheli to Volo Volo, Moron) Mremani to Mutsamudu, Anjouan Source: World Bank survey. Agricultural Production Agriculture is an important sector in Comoros, with approximately 62 percent of the population engaged in agricultural production (Figure D.22). Major food crops include coconuts, cassava, and bananas. Comoros also produces valuable export crops, such as vanilla, ylang ylang, and clove (Figure D.23). According to the 2014 Enquete 1-2-3 survey, substantial agricultural production takes place in the northern areas of Grande Comore, Anjouan, and Moheli (Figure D.24). These expenditure survey data are generally not very precise, but the distributional trend is clear. Figure D.22. Labor force composition according to sector Figure D.23. Major crops Source: World Bank estimates based on 2014 EDMC. Source: FAOSTAT. 88 Figure D.24. Main crop production according to village Source: World Bank estimates based on 2014 EDMC. Agricultural productivity is low. The vast majority of agricultural production is subsistence farming. Because of the poor condition of the road network and lack of transport services, many farmers do not have good access to input or output markets, so the use of advanced inputs is minimal. Only 12 percent of households used fertilizer, and 11 percent used improved seeds (Iimi 2019). Intermediary services are unavailable in most places. Crops are not often collected by cooperatives or traders. In addition, many farmers do not have any transport means in Comoros (Figure D.25). Figure D.25. Use of agricultural inputs and transport means by farmers Source: Iimi 2019. Crop production can be increased by improving transport connectivity to markets, especially Moroni. A recent study shows that accessibility to Moroni, the primary market in the country, is of particular importance in stimulating agricultural growth in Comoros. By reducing inter- and intra-island transport costs to Moroni, farmers can be encouraged to participate in market transactions. It is predicted that a 10-percent increase in accessibility 89 to the Moroni market could increase agricultural production by 8.5 percent and increase farmers’ market participation by 19 percent (Figure D.26). Not only transport infrastructure, but also complementary interventions are important for agricultural growth. In particular, intermediary service providers, such as agricultural cooperatives and crop collectors, play an important role in increasing farmers’ crop production (Iimi 2019). Figure D.26. Estimated elasticity of crop production values Source: Iimi (2019). Access to Healthcare Services As the COVID pandemic has revealed, healthcare accessibility, particularly to advanced healthcare services, is an important challenge in Comoros. The country has a well-established healthcare network, with 74 healthcare facilities (Table D.1). Many people have good access to basic healthcare services, if all kinds of healthcare facilities are taken into account, including health centers and health posts where only basic healthcare services are offered (Figures D.27 and D.28), but advanced healthcare services are available at only a few facilities on each island. Thus, despite its relatively well-connected road network, Grande Comoro has many areas where advanced healthcare services are not easily reachable from local communities (Figure D.29). In approximately half of Anjouan and Moheli, as well, it takes more than 30 minutes by car to visit an advanced healthcare facility. Table D.1. Health facilities in Comoros No. of Available services: Type facilities Emergency, surgery Medicines Centre Hospitalier National 1 Yes Yes Centre Médico-Chirurgical 4 Yes Yes Centre de Sante de District 12 No Yes Centre de Sante Familial 3 No Yes Center de Sante 2 No Yes Poste de Sante/Pole de Sante 52 No No Source: Ministry of Health, Comoros (2015). 90 Figure D.27. Access to healthcare services Source: World Bank 2019c. Figure D.28. Access to healthcare services Figure D.29. Access to high-end healthcare services Source: World Bank estimation. Source: World Bank estimation. Because of the variation in accessibility to healthcare services, health performance differs widely across the country. Although data are limited, the share of outpatients under 5 years old who visited a health facility was highest in Moheli according to the Annual Health Statistics in 2014/15, followed by Anjouan and Grande Comore (Figure D.30). The tuberculosis rate was lowest in Moheli. These are consistent with transport accessibility to all types of healthcare facilities measured above. It is likely that poor health is related to poor transport connectivity. Although many factors affect access to healthcare facilities and health status, there is a correlation between transport connectivity and health sector indicators, as district-level data confirm. Rural accessibility is negatively correlated with under-5 mortality, with mortality high where rural accessibility is low (Figure D.31). Similarly, more cases of severe wasting are observed where rural access is poor. It is likely that enhancing transportation access to healthcare facilities would improve the health of Comorians 91 Figure D.30. Healthcare access and performance indicators according to region Source: Ministry of Health, Comoros (2015). Figure D.31. Health performance indicators and rural accessibility Under-5 mortality Under-5 severe wasting Sources: World Bank estimates; Haazen and Rajoela 2016. 92 Appendix E. Road Safety Measures There is a wide range of road safety measures, some more cost-effective than others. Sound evidence-based (not common sense based) interventions deliver strong economic savings. For example, global experience shows that some road engineering measures, such as road humps and narrowing, and some speed management devices, such as speed cameras, are much more cost-effective measures for managing speed (Figure E.1). Some, although not all, behavioral change interventions, such as graduated driver training, may be cost effective. Management of speed is fundamental to road safety improvements. Even small reductions in speed can yield large safety benefits, whereas allowing even small increases will result in large increases in deaths and injuries. Based on a comprehensive synthesis of many international studies, a 1 percent decrease in speed results in a 4 percent decrease in deaths (Figure E.2). Management of speed is a major low-cost road safety opportunity. Figure E.1. Cost effectiveness of road, vehicle, and behavior-change interventions for speed management Source: Job and Mbugua 2020. Figure E.2. Relationship between speed and road crashes Source: Nilsson 2004. 93 Institutions and governance. Comoros has a lead agency for road safety, National Multisectoral Committee on Road Safety (NMCRS), but it is not funded in the national budget, nor does it function as a proper national road safety agency (GRSF 2019). Expected functions of such an agency generally include coordination, legislation, development, and monitoring and evaluation of road safety strategies, but the NMCRS does not have the capacity to develop, implement, or coordinate road safety policies. Fundamental requirements of a lead agency include sufficient road safety expert staff, funding, and resources to perform these multiple functions, legislated powers to guide and direct road safety activities of multiple government departments, and responsibility for road safety outcomes. Substantial capacity building to develop the agency’s capacity is urgently required. Collection, collation, sharing, and analysis capacity for road safety data (at a minimum including sound comprehensive crash data) are also required to guide and target road safety interventions. Without accurate, timely data, it is difficult to improve road safety performance sustainably. The government should prepare and implement a national road safety action plan. Comoros does not have a clear national road safety strategy or any known national road safety target. The NMCRS is expected to be responsible for reviewing traffic legislation, maintaining an up-to-date georeferenced road crash database, and elaborating the road safety strategy and policy. Based on these, the NMCRS should prepare a multisector national road safety action plan that defines short- to long-term tasks and responsibilities of all relevant government agencies and stakeholders, as well as resources for a functional lead agency. The NMCRS should coordinate all their policy actions and provide training programs for traffic enforcement police, road engineers and designers, national and regional policy makers, and the community. Interministerial collaboration and coordination are critical. The Ministry of Land Planning and Urbanization in charge of Land Affaires and Land Transport, together with the NMCRS and the Ministry of Education, should be responsible for road safety education and promotion of the action plan to improve on-road behavior and not just raise awareness. Promotion and education. Community road safety is important and requires a bottom-up approach whereby road users are required to be educated and motivated (primarily through the threat of enforcement) to obey road rules and increase safety on the road. Despite their common-sense value, general road safety education and driver training are consistently shown to be generally ineffective at increasing road safety (see Turner et al. 2021 for a summary of relevant international evidence), although certain elements can improve road safety, such as training children of sufficient age how and where to cross the road and extended hours of novice driver experience on the road with an experienced supervising driver—not necessarily a professional driver trainer—before being allowed to drive solo. Community promotion on road safety can also influence agencies and persons responsible for road safety to act more effectively if the community generates more demand for real road safety action. Targets of the community road safety program are local residents, including students, teachers, hawkers, shopkeepers, and commercial drivers. Comorian police statistics from 2017 to 2019 show that 70 percent of victims of road crashes are male, 30 percent are younger than 20, and 52 percent are aged 20 to 45, the most economically active age group. Vehicle occupants account for 58 percent of deaths and serious injuries, pedestrians 27 percent, and two-wheelers 16 percent. Some of the main reasons for pedestrian deaths and injuries are lack of strong management of speed and lack of safety facilities for pedestrians (e.g., footpaths, sidewalks, safe crossing facilities). Regulations and enforcement. There are no laws on vehicle seatbelt or motorcycle helmet use, child restraints, or use of mobile devices while driving. There is a speed limit, but enforcement is rated 3 on a scale of 1 to 10; there is a national drunk-driving law, although a systematic approach to a robust blood alcohol concertation limit is not apparent (GRSF 2019), but enforcement is rated 2; and there is a drug-impaired driving law. Significant efforts are required to improve national laws and effectively enforce them. Infrastructure. Together with training of relevant government staff, the Ministry of Land Planning and 94 Urbanization can adopt standards and regulations and a number of reactive and preventive engineering countermeasures that would increase road safety significantly. The NMCRS could possibly discharge most of its operational activities to a specific road agency in charge of traffic, which is the DRTR under the current setup, and a new mandate for a road agency, if created, could be developed. Most of these countermeasures could cover expanded use of traffic calming to decrease speeds, especially in urban areas; median separation; sidewalk protection; guardrails on curves, embankments, and high-risk locations with unforgiving roadside objectives and drops; intersections and road access improvement and signaling; bus parking and market truck unloading areas; road project signaling and fencing; black spot treatment; and horizontal and vertical signage and delineation. Detailed road safety surveys. The need for and location of these improvements can be identified, valued, and prioritized through a comprehensive road safety survey conducted along the entire road network. One well- known survey system is the International Road Assessment Programme, developed in a collaboration between the Global Transport Knowledge Partnership and the World Bank Global Road Safety Facility. A typical International Road Assessment Programme survey measures the safety level of vehicle occupants and pedestrians along road sections and rates them on a scale from 1 to 5. It also develops a safer road investment plan based on relative benefits of more than 90 countermeasures, ranging from low-cost road markings and pedestrian refuges to higher-cost intersection upgrades and full highway duplication. The investment plan gives countermeasure options that could maximize prevention of deaths and serious injuries within available budgets and focuses on reducing risk of head-on collisions, reducing risk associated with run-off crashes by improving lane markings and reducing the severity of roadsides, reducing risk at intersections, and improving pedestrian facilities. Given Comoros’ tight fiscal space, a reasonable way of funding road safety may be to seek external resources. There are a number of funding opportunities that governments, road safety organizations, and nongovernmental organizations can apply for to implement road traffic injury prevention projects during the Decade of Action for Road Safety (2011–20). The World Bank Global Road Safety Facility was launched in 2005 with the objectives of increasing funding and technical assistance for global, regional, and national initiatives to enable low- and middle- income countries to implement road safety programs. It can be activated in conjunction with a World Bank lending project. The United Nations Road Safety Fund was soliciting proposals until January 2021 and may extend the deadline. The Global Road Safety Partnership, an implementing partner of the Bloomberg Philanthropies Initiative for Global Road Safety, manages a grants program and supports projects to develop and deliver high-impact, evidence-based road safety interventions designed to strengthen road safety policies and their implementation. Tanzania has been a recipient of these funds. The above United Nations funding opportunities are limited in size, are restricted to eligible road safety activities, are competitive, require thorough preparation, and often involve a long approval process. Therefore, the government must create a national road safety fund that could use various national public and private resources for priority road safety activities following rules that the NMCRS will develop and propose. Beyond the international and national trust funds, the road maintenance budget generally includes a 2 percent component for road safety engineering improvement and countermeasures; a recent trend to integrate road safety into road asset management is developing. In addition, the road investment budget, primarily funded by development partners, can include a road safety investment component combining hard and soft interventions. Priority high- impact hard interventions often result from a comprehensive road safety countermeasures survey, and the Regional Centre of Excellence for Road Safety, sponsored by the African Development Bank, could provide soft interventions in the form of training 95 Appendix F. Road Safety Risk Assessment Conceptual Framework Each factor is calculated based on the concept of an accident modification factor or a crash modification factor (CMF), which is a multiplicative factor to compute the expected number of crashes when road characteristics at a particular site are taken into account (e.g., by assuming certain countermeasures). A CMF greater than 1 indicates greater likelihood or severity of a target crash type. Many CMFs are derived from experimental or before-and- after analysis in the literature. Following the literature, a wide variety of attributes are taken into account (Table F.1). For instance, the likelihood that a crash will happen depends on operating vehicle speed, existence of intersections and pedestrian crossings, and road conditions. Vulnerability depends on vehicle speed but also existence of road medians and shoulders. Exposure varies depending on how many pedestrians and bicycles pass and, in theory, should be related to vehicle traffic. This type of data (e.g., annual average daily traffic) is not always available or collected through a different survey from safety assessments. In the case of Comoros, no recent traffic data are available, so vehicle traffic is omitted or assumed to be constant across roads. Based on the three factors, the risk score si is calculated for each of the three road user categories ( 𝑖 ∈ {𝑀𝑉, 𝐶𝑌𝐶, 𝑃𝐸𝐷}): • Crashes involving only one or more motor vehicles (MV) • Crashes involving at least a bicyclist (CYC) • Crashes involving at least a pedestrian (PED) Motorcyclists are combined with the motor vehicle category. Safety risk may vary depending on type of vehicle (e.g., passenger car, bus, heavy truck), but there are few studies that distinguish CMFs from different types of vehicles. In addition, in Comoros, motorcycle traffic is generally low. Thus, only one category is considered for all motor vehicles. Taking into account pedestrian and bicycle flows, a global risk score (GRS) is calculated as a weighted average of the risk scores of motor vehicles, cyclists, and pedestrians: 𝐺𝑅𝑆 = (𝑠𝑀𝑉 + 𝑤𝐶𝑌𝐶 𝑠𝐶𝑌𝐶 + 𝑤𝑃𝐸𝐷 𝑠𝑃𝐸𝐷 )/(1 + 𝑤𝐶𝑌𝐶 + 𝑤𝑃𝐸𝐷 ) where wCYC and wPED are the weights calculated in proportion to the numbers of cyclists and pedestrians observed on a given section of road, respectively. The weighted average is calibrated according to the gamma distribution. If pedestrians or cyclists are not counted on a road, the GRS becomes close to the risk score for motor vehicles, sMV, whereas if there are many pedestrians or cyclists, the GRS becomes more correlated with their risk scores: sCYC and sPED. In general, a road is considered to be more dangerous for pedestrians and cyclists when proper road structures such as wide shoulders, raised medians, safe crossing points, and sidewalks are missing or vehicles drive faster. Table F.1. Attributes of observed road safety risk factors in the model Attribute Factor Danger Vulnerability Exposure [likelihood] [severity] Observed bicycle flow X Observed pedestrian flow X Operating speed X X Median type X X Intersection type X Area type X Access points X 96 Number of lanes X Lane width X Curvature X Grade X Road surface conditions X Delineation X Pedestrian crossing X Speed management/traffic calming X Shoulder width X Roadside severity - distance X X Sidewalk X Cycling facilities X Dedicated motorcycle lane X Detailed Definition of Parameters The attributes used in the simplified methodology are listed in Table F.2. It also includes references/citations for the CMFs used. Table F.2. Road attributes considered in the simplified methodology Attribute Road safety considerations Literature reference Observed Observed bicycle flow is the number of cyclists observed within a 100-m road iRAP (2013) bicycle flow section and is used as a proxy for exposure of cyclists to road hazards. It is used to calculate a GRS, taking into consideration bicycle and pedestrian flows plus the risk calculated for the three road user categories. Observed Observed pedestrian flow is the number of pedestrians observed (crossing and iRAP (2013) pedestrian flow walking along the road) within a 100-m road section and is used as a proxy for exposure of pedestrians to road hazards. It is used to calculate a GRS, taking into consideration bicycle and pedestrian flows plus the risk calculated for the three road user categories. Inspecting Inspecting vehicle maximum speed observed within a 100-m road section is iRAP (2013) vehicle speed used as a proxy for operating speed along a road section. (operating speed) Area type Some treatments have a different impact according to traffic conditions. A Elvik et al. roundabout in an urban context has a different safety effect than a rural (2009) roundabout. Area type attributes affect the value of specific CMFs. In general, number of crashes per million vehicle-kilometers of travel varies greatly between different traffic environments. Even if it would be better to identify specific CMFs for rural and urban areas, the literature is not sufficiently comprehensive for this. Increasing the risk in urban areas is a good compromise to differentiate rural from urban CMFs. Median type Medians have been found to reduce crashes in most situations. Elvik et al. Their CMF value is the expected reduction in crashes after installing a median. (2009) Intersection Studies mostly show that intersections with four or more legs are associated Elvik et al. type with more crashes than three-leg intersections. Intersections are associated (2009) with more crashes than roundabouts. Access points The number of access points has a major impact on collision rate. Elvik et al. (2009) Number of The impact of number of lanes on collision rates depends on whether the road Elvik et al. lanes is in an urban or a rural area. In rural areas, collision rate declines as road (2009) width (number of lanes) increases, whereas in urban areas, collision rate increases as road width (number of lanes) increases. Differences in speed and 97 Attribute Road safety considerations Literature reference mix of traffic may account for this difference in the effect of number of lanes. In general, the more lanes, the higher the pedestrian exposure to crashes when crossing the road. Lane width Widening lanes reduces the occurrence of run-off road crashes, head-on and AASHTO sideswipe collisions, and same- and opposite-direction crashes. (2010) Curvature Collision rate increases as curves are sharper. Elvik et al. (2009) Grade There are more crashes at intersections with steep gradients than at those with Elvik et al. no or small gradients. The opposite is true for crashes along the road. (2009) Road It is estimated that increasing unevenness and ruts increases accidents. AASHTO conditions (2010) iRAP (2013) Lane markings This attribute measures the presence of clearly visible markings along a road AASHTO or at an intersection (edge lines and center line). Placing edge lines and center (2010) line markings where no markings exist decreases collisions with injuries of all types. Pedestrian Appropriate pedestrian crossings may increase pedestrian safety, although Elvik et al. crossing motor vehicle collision rates appear to increase, possibly because of an (2009) increase in rear-end accidents. Some measures designed to increase the safety of a specific crash pattern or road user group may have an indirect positive effect on other situations and road users. According to the literature, this is the case for pedestrian crossings designed to increase vulnerable road users’ safety. Speed Traffic-calming and speed-reducing devices are generally found to reduce Elvik et al. management, number of crashes. (2009) traffic calming Shoulder width Shoulders wider than 0 meters are considered paved shoulders. Paving and AASHTO increasing shoulder width have been found to reduce the number of collisions (2010) with injury. This road attribute is considered for all three road users assessed. Wider shoulders can provide more space for motor vehicles to correct errors and for other road users such as pedestrians and cyclists to travel in. This space can be used to increase the distance between vulnerable road users and motor vehicle traffic, thus decreasing the crash risk with VRUs. Pedestrians often walk along roads in developing countries, where the level of motorization is lower than in other countries. Roadside Increasing the distance to fixed obstacles was found to decrease the number of Elvik et al. severity – collisions with injury. (2009) distance Roadside hazard risk should be calculated not only according to distance to the hazard, but also according to type of hazard. Technology limitations of automated video analysis make it impossible to recognize roadside features reliably, thus assessing the existing risks. Current algorithms for video analysis do not allow all possible roadside obstacles to be recognized, even the main obstacles (e.g., trees). Recognizing them would only be possible manually, which would be too complicated for the simplified methodology. Sidewalk Sidewalks carry pedestrian traffic and sometimes bicycle traffic in both Elvik et al. directions. Collision rates are lower for pedestrians and cyclists and higher for (2009) motor vehicles on roads with sidewalks than on other roads. Some measures designed to increase the safety of a specific crash pattern or road user group may also have an indirect positive effect on other situations and road users. This is the case for sidewalks, which are designed to increase the safety of vulnerable road users. 98 Attribute Road safety considerations Literature reference Cycling Cycle lanes and cycle tracks reduce the number of collisions with injury (for Elvik et al. facilities all vehicles). For bicycle crashes, the reduction in number of collisions is (2009) smaller than for other road users. More collisions for cyclists could be because there are more cyclists and greater speed differences between cyclists. Some measures designed to increase the safety of a specific crash pattern or road user group may also have an indirect positive effect on other situations and road users. This is the case for cycling facilities, which are designed to increase vulnerable road users’ safety. Dedicated No relevant CMF was found, it is assumed that a dedicated motorcycle lane iRAP (2013) motorcycle would have effects similar to those of a cycle lane. lane Note: GRS, global risk score; CMF, crash modification factor CMFs for each road attribute have been considered for the following main locations: • Along the road (when a road user is traveling or walking along the road) • At intersections (when a road user is crossing through an intersection while traveling along the road— valid for motor vehicles and bicycles) • While crossing a road (valid for pedestrians) These conditions lead to two possible cases for each road user category. Not all of the CMFs are defined for all possible conditions, because not all the attributes are used to calculate traffic crash risks for the three road user categories. Figures F.1 and F.2 list the road attributes to be considered for each crash type on a road section or at a road intersection or access point. Figure F.1. Road attributes considered for crash types at road sections MV crashes Pedestrian crashes Cycle crashes •Facilities of bicycling •Pedestrian crossing facility •Facilities of bicycling •Grade •Curvature •Curvature •Curvature •Delineation •Delineation •Median •Sidewalk •Shoulder width •Motorcycle dedicated lane •Speed management/ traffic •Road surface conditions •Shoulder width calming •Speed management/ traffic •Pedestrian crossing •Area Type calming •Delineation •Area Type •Roadside severity •Motorcycle dedicated lane •Sidewalk •Speed management/ traffic calming •Lane width •Number of lanes •Road surface conditions •Area Type 99 Figure F.2. Road attributes considered for crash types at road intersections MV crashes Pedestrian crashes Cycle crashes •Facilities of bicycling •Facilities of bicycling •Facilities of bicycling •Grade •Intersection type •Intersection type •Intersection type •Pedestrian crossing •Speed management/ •Access points •Speed management/ traffic calming •Pedestrian crossing traffic calming •Sidewalk •Speed management/ •Number of lanes •Motorcycle dedicated traffic calming •Shoulder width lane •Lane width •Access points •Access points •Motorcycle dedicated lane CMFs derived from the literature and used in the simplified methodology are shown in Table F.3. Table F.3. Crash modification factor values according to road attribute category, crash type, and location Attribute category Bicycle crashes Pedestrian crashes Motor vehicle crashes Intersection Along Crossing Along the Intersection Along the the road road road Area type Urban 2.00 2.00 2.00 Rural 1.00 1.00 1.00 Median type Physical median 0.85 Center line 1.00 Intersection type No intersection rural 0.00 0.00 0.00 urban 0.00 0.00 0.00 Roundabout rural 1.00 1.00 1.00 urban 1.00 2.50 2.50 3-leg urban 1.36 2.50 2.50 rural 1.36 2.50 2.50 4-leg urban 1.45 2.67 2.67 rural 1.64 3.00 3.00 ≥35-leg rural 1.82 3.33 3.33 urban 1.82 3.33 3.33 Access points ≥33 1.50 1.50 1.50 1 or 2 1.30 1.30 1.30 0 0.00 0.00 0.00 Number of lanes 1 1.00 1.00 2 1.90 0.90 ≥33 3.20 0.90 Lane width, m 100 Attribute category Bicycle crashes Pedestrian crashes Motor vehicle crashes Intersection Along Crossing Along the Intersection Along the the road road road <2.75 1.50 1.50 2.75–3.25 1.20 1.20 ≥3.25 1.00 1.00 Curvature, m <200 2.35 2.35 2.35 200–400 1.85 1.85 1.85 400–600 1.55 1.55 1.55 600–1,000 1.30 1.30 1.30 1,000–2,000 1.10 1.10 1.10 >2,000 1.00 1.00 1.00 Grade, % ≥7.5 1.10 0.80 <7.5 1.00 1.00 Road condition Poor 1.40 1.40 Poor-medium 1.30 1.30 Medium 1.20 1.20 Medium-mood 1.10 1.10 Good 1.00 1.00 Lane markings Adequate 1.00 1.00 1.00 Poor 1.25 1.25 1.25 Pedestrian crossing facility Yes 0.90 0.90 1.10 1.10 No 1.00 1.00 1.00 1.00 Speed management / Traffic calming Yes 0.90 0.90 0.90 0.90 0.90 0.90 No 1.00 1.00 1.00 1.00 1.00 1.00 Shoulder width, m 0 1.50 1.50 1.50 < 1.0 1.30 1.30 1.30 1.0–2.4 1.00 1.00 1.00 ≥ 2.4 0.80 0.80 0.85 Roadside severity – distance, m <1 1.45 1 –5 1.20 ≥5 1.00 Distance from sidewalk to roadway ≥1 0.95 0.95 1.15 <1 1.00 1.00 1.00 No sidewalk 1.10 1.10 1.10 Cycling facilities On-road lane 0.75 0.80 0.75 0.50 0.75 No facility 1.00 1.00 1.00 1.00 1.00 Dedicated motorcycle lane On-road lane 0.75 0.80 0.50 0.75 No facility 1.00 1.00 1.00 1.00 101 Results In Comoros, road safety risk is high, especially in urban areas. The assessment was focused on 384 km of main roads for which sufficient data were collected for the analysis. Approximately 3 percent of road sections (100 m) surveyed are considered to be low risk, 37 percent medium risk, 51 percent high risk, and 8 percent very high risk (Figure F.3). In comparison with other countries in Africa where the same methodology has recently been applied, the overall risk of Comoros is worse than Mozambique and comparable with Liberia. In Mozambique, of approximately 470 km of national highways (N1 and N10) surveyed, 37 percent were assessed as high or very high risk, although there are fewer low-risk roads than in Comoros (World Bank 2018b). In Liberia, of approximately 500 km of primary roads between major cities, such as Monrovia, Buchanan, Ganta, and Bo, examined, 64 percent were assessed as high or very high risk (World Bank 2019f). Figure F.3. Road safety risks in Comoros, Liberia, and Mozambique Source: World Bank estimates. In general, urban areas are much riskier because there are more intersections and more pedestrians and cyclists. This is consistent with the recent increasing trend of road crashes in urban areas. More than 90 percent of urban roads are at high or very high risk for vehicles, bicyclists, and pedestrians (Figure F.4). On average, roads in urban areas have more than twice as many pedestrians as those in rural areas (Figure F.5). Fifty-one percent or rural roads are at high or very high risk. The risk for vehicle users is lower in rural areas because there are few intersections. There are also fewer pedestrians and cyclists, but vehicle speed is faster (Figure F.6), so the small number of pedestrians and cyclists tends to translate into high overall road safety risk. Figure F.4. Road safety risks in urban and rural areas Source: World Bank estimates. 102 Figure F.5. Average number of pedestrians and bicycles Figure F.6. Average vehicle speed according to area along roads Source: World Bank estimates. Source: World Bank estimates. According to region and island, Moheli is most dangerous, especially in its urban zones, followed by Grande Comore (Figure F.7). As discussed, most of the urban roads are at high risk in Comoros. High-risk areas are concentrated around major intersections (Figure F.8). Some roads are well designed, with wide sidewalks, road crossings, and well-structured roundabouts, although they can still be improved with features such as speed humps to manage speed (Figure F.9). Others do not have such structures. Many pedestrians walk close to roads without any separation (Figure F.10). Rural roads are less dangerous, although still at medium or high risk. In particular, in Moheli, rural roads are very dangerous because of the lack of road structures, difficult road alignments, and poor road conditions. In mountainous areas, there are many blind spots without any safety features. In rural areas, vehicle speed is relatively high. Rural roads often pass through the middle of rural villages. Road conditions are generally poor (Figure F.11). Figure F.7. Road safety risks according to region and island Urban areas Rural areas Source: World Bank estimates. 103 Figure F.8. Road safety risks and black spots Source: World Bank estimates. Figure F.9. Areas with some road safety features in urban zones in Anjouan Sidewalk and road crossing Roundabout with wide sidewalk and street lights Source: World Bank. Figure F.10. High-risk areas in urban zones in Grande Comore Busy intersection with no traffic structures Many pedestrians without sidewalk Source: World Bank. 104 Figure F.11. High-risk areas in rural zones in Moheli Blind spot in mountainous areas with no guard rail Good road, meaning high speed, but no sidewalk Road passing through a rural village, without speed Rural roads in very poor condition management devices Source: World Bank. 105