i © 2021 The International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street NW Washington, DC 20433 USA All rights reserved Photos: Shutterstock and Unsplash. HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM Overview iii iv CONTENTS 1 Introduction .................................................................................... 1 2 Background and Context ............................................................ 4 3 The Economic Geography of the HoA ................................... 15 3.1 Density: Spatial Distribution of Economic Activity and Living Standards ................................ 15 3.2 Distance: Access to markets and integration ... 20 3.3 Division: Weak cross-border connectivity compounded by fragility and thick borders ..... 24 4 Regional economic integration for job creation and structural transformation ............................................................ 28 4.1 Building capabilities and connecting workers to jobs ......................................................................... 31 4.2 Creating and connecting to markets ................. 35 5 Key actions for jobs and economic transformation in the HoA ............................................................................................. 40 5.1 Institutions: Building human capital ................... 40 5.2 Infrastructure: Trade facilitation and harmonization and selective investments in physical infrastructure ........................................... 41 5.3 Interventions: Metropolitan regions and selected secondary cities ...................................... 41 5.4 Additional considerations ...................................... 42 v HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM: OVERVIEW SECTION 1 Introduction The Horn of Africa (HoA)1 is most commonly of off-farm employment in rural areas.3 There is known for its long history of fragility, strong demand from the Gulf for some products catastrophic droughts, and seemingly produced in the region, in particular, meat and intractable conflicts. A complex set of historical, other agricultural products. There is considerable ideological, political, economic, geographical, potential for enhanced cross-border trade, which territorial, and environmental factors have created is intrinsically transnational but mainly small-scale tensions within and between states, at times (informal) in the Horn, with significant potential boiling over into inter- and intra-state violent to leverage entrepreneurship and greater private conflicts. These conflicts have, in turn, further sector activity to generate jobs. For example, the weakened states’ capacity to provide public positive impact of digital innovation to open up services, social cohesion, and further increased new economic opportunities and boost incomes the vulnerability of its population (particularly in has been demonstrated in Kenya, and the wider historically marginalized border areas), thereby region could build on this experience. increasing the risk of future conflicts. Climate change is now exacerbating an already difficult The economic priorities in the region are situation, leading to increased tension over consistent with the jobs and economic natural resources. Successive shocks of different transformation (JET) agenda, which recognizes kinds have led to record numbers of displaced the role of connecting to markets in generating job- people and increased migration, mostly irregular, creating private investment, and the importance to Gulf countries.2 of building capabilities and linking workers to jobs. In the Horn of Africa, there is an opportunity to Yet the true picture of the HoA is more nuanced, drive the JET agenda through deeper regional and HoA countries could unlock significant integration and investments in connectivity common opportunities. Cities and towns across to improve access to markets across borders, the HoA are growing, increasing the demand realizing common opportunities to diversify, for food grown elsewhere in the region. The deliver jobs and reduce poverty and vulnerability. growth in urban incomes is increasing demand At the same time, however, implementing the for higher-value foods, including processed JET agenda in the HoA will require dramatic foods and animal products, with the potential to improvements in education and other investments benefit rural producers and create jobs across to build capabilities and overcome significant the agri-food value chains. The growth of cross-border and interconnected risks, including secondary cities attracts service providers that currently poor human capital outcomes, large and can provide jobs for those migrating from rural rapidly growing numbers of youth, fragility, and areas and deliver services supporting the growth environmental degradation and climate change. 1 For the purposes of this report, the Horn of Africa is defined as a subset of the larger Horn and comprises Djibouti, Eritrea, Ethiopia, Kenya, and Somalia. The map in the Annex shows the smaller Horn of Africa and its main cities. 2 In 2018, over 315,000 migration movements were observed between the HoA and the so-called Eastern Route (towards Yemen and the Arabian Peninsula), increasing to 469,000 in 2019 (DTM, 2019). 3 In Ethiopia, for example, the number of manufacturing firms in Addis Ababa grew by 48% (to 1,120) from 2010 to 2016, compared with growth of 81% (to 850) in the main secondary cities according to manufacturing census data. 1 SECTION 1 | INTRODUCTION The objective of this Regional Economic conflict and the region’s complex and evolving Memorandum (REM) is to strengthen the political economy. This overview synthesizes economic analysis available to policymakers the key findings of the analysis conducted for on the challenges and opportunities for the HoA REM, full details of which are presented regional economic integration to support job in a series of Background Papers. The overview creation and economic transformation in the briefly describes key aspects of the region’s Horn of Africa. It assesses the current state of economy and development progress (Section regional economic integration, how policies and 2). Next, in Section 3, it presents features of the investments can deepen this integration, and how economic geography of the region and some key this could help to address the opportunities and results from economic modeling and transport challenges confronting the region. The analysis connectivity analysis. The findings demonstrate applies both an economic geography perspective the salience of the JET agenda in the Horn, and (based on the 3Ds framework of the 2009 WDR this and its implications are discussed in Section – density, distance, and division)4 and the lens 4. Finally, Section 5 concludes by highlighting the of the jobs and economic transformation (JET) main policy messages which emerge from the agenda,5 whilst taking into account fragility and REM’s regional-level analysis. 4 World Bank (2009), World Development Report 2009: “Reshaping Economic Geography”. 5 Special Theme: Jobs and Economic Transformation (English). IDA19 Washington, D.C.: World Bank Group. 2 HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM: OVERVIEW 3 SECTION 2 | BACKGROUND CONTEXT SECTION 2 Background & Context Positive growth and poverty trends from a low base Between 2014 and 2019, growth in the HoA growth. Despite the recent positive trends, outpaced growth in Sub-Saharan Africa (SSA) income levels in the Horn remain below the as a whole. While real per capita GDP levels SSA average. Average GDP per capita of the four remained essentially flat in SSA (excluding high- HoA countries amounted to about US$1,100 in income economies), they grew at an annual rate 2019, compared to US$1,600 for SSA (excluding of four percent in the HoA (Figure 1).6 Ethiopia, high income). There is substantial variation Djibouti and Kenya recorded the highest per within the HoA itself, with per capita income capita growth rates, while in Somalia economic levels being 10 times higher in Djibouti than in growth hardly outpaced estimated population Somalia (Figure 2). Figure 1: Economic growth in the HoA was Figure 2: Though income levels remain low mostly solid in comparison (Per capita GDP growth, 2014-2019) (Per capita GDP levels, 2019, current US$) 8,0% 4,000 GDP per capita (current US$) GDP per capita growth 3,500 6,0% 3,000 4,0% 2,500 2,000 2,0% 1,500 1,000 0,0% 500 -2,0% 0 2014 2015 2016 2017 2018 2019 Djibouti Ethiopia Kenya Somalia HoA SSA Ethiopia Kenya Djibouti Somalia SSA (excl. high income) Source: WDI, 2020; Somalia Economic Update, 2020. World Bank Source: WDI, 2020. Somalia Economic Update, 2020. World Bank staff calculations. staff calculations. Though trends in poverty have also been positive, years later.7 Given rapid population growth, the the pace of poverty reduction has been too slow, absolute number of people living in poverty rose and population growth too rapid, to reduce the by 9 percent over the same period. Poverty rates number of people living in extreme poverty. For range from 17 percent in Djibouti to 69 percent in the three countries where data are available, poverty Somalia (Figure 3).8 Overall, an estimated 57 million rates based on the US$1.9 a day line decreased people across the HoA live below the international from around 40 percent in 2005 to 33 percent 10 US$1.9 a day poverty line (Figure 4). 6 This does not include Eritrea for which there are no publicly available and reliable GDP data. 7 For Ethiopia, poverty surveys were conducted in 2004/5 and 2015/16. For Kenya, 2005/6 and 2015/16. For Djibouti, 2002 and 2017. Poverty is based on the international US$1.9 poverty line. 8 Somalia’s poverty rate is estimated based on areas that were safe enough to survey. It does not include areas which were inaccessible due to insecurity. 4 HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM: OVERVIEW Human development outcomes have improved remain particularly poor, even among the in line with economic growth, though remain younger generation. Between half and two- low. Between 2014 and 2019, the Human thirds of children finish primary school (except in Development Index (HDI), a summary measure Kenya, where completion is universal) and gross of average achievements in health, education, enrolment in secondary remains relatively low and living standards, improved in all countries (Figure 6). Based on the most recent household except for Eritrea (Figure 5). However, except living standards surveys, 45 percent of youngsters for Kenya, the HDI remains below the average between 15 and 24 had completed primary school for Least Developed Countries (LDC). With the or more at the time of the survey (between 2015 exception again of Kenya, education outcomes and 2017, depending on the country).9 Figure 3: One third of the population lives below Figure 4: The bulk of the poor live in Ethiopia US$1.9 a day and Kenya (US$1.9 a day poverty rate) (Estimated number of people below US$1.9, millions) 100 50 Number of poor (in millions) Poverty headcount rate 69 40 75 31.9 30 50 37 18.2 31 20 25 17 7.2 10 0.2 0 0 SOM* KEN ETH DJI SOM* KEN ETH DJI Source: World Bank staff calculations based on the latest survey Source: World Bank staff calculations based on the latest survey available (2017 for Djibouti, 2015 for Kenya, 2015/16 for Ethiopia, available (2017 for Djibouti, 2015 for Kenya, 2015/16 for Ethiopia, and 2017/18 for Somalia). The dash line corresponds to the regional and 2017/18 for Somalia). poverty headcount rate for the HoA (except Eritrea). Notes: * indicates that Somalia’s poverty rate is estimated based on areas that were safe enough to survey. It does not include areas which were inaccessible due to insecurity. 9 This should not be confused with the primary school completion rate. Given the prevalence of late school enrolment and high repetition rates, a substantial share of 15-24-year-olds are still in primary school. 5 SECTION 2 | BACKGROUND CONTEXT Figure 5: Human development improved Figure 6: But education outcomes remain weak (ex. Kenya) (Human Development Index, 2014-2019) (Primary completion rate and gross secondary enrolment) 0.7 100 Human development index LDC average 0.6 90 80 0.5 70 0.4 60 50 0.3 40 0.2 30 20 0.1 10 0 0 DJI ETH ERI KEN SOM DJI ERI ETH KEN SSA (excl. high 2014 2015 2016 Primary school completion rate income) 2017 2018 2019 Gross secondary enrolment Source: UNDP, 2020. Source: WDI, 2020. KNBS, 2018. Employment characterized by self-employment and own-account agriculture, with substantial inequities in participation Labor market outcomes differ across countries systems, rather than strong labor demand. In but remain relatively poor, especially for youth Djibouti and Somalia, labor force participation and women. In countries where the population is low, mainly driven by social norms regarding is predominantly rural (Ethiopia, Eritrea, and the labor market participation of women and, in Kenya), labor force participation rates are high Djibouti, a pattern of job-poor growth centered (Figure 7). High labor force participation is an around the port. Youths’ participation in the labor empirical regularity in largely agrarian low or market is low in Djibouti, Kenya, and Somalia lower middle-income countries, reflecting the (Figure 8). For Djibouti and Kenya though, this prevalence of unpaid family labor and the need is partly explained by the high share of young to work, given the lack of robust social assistance people who are in school. Figure 7: Labor force participation is high Figure 8: Youth are generally less likely to except for Djibouti and Somalia participate in the labor force (Labor force participation rate, %) (Labor force participation rate, %) 100 100 90 91 91 90 83 90 78 78 80 80 77 73 80 76 70 70 69 70 63 60 60 49 50 44 50 44 42 41 40 40 30 30 20 19 20 15 10 10 0 0 ETH DJI KEN SOM ERI ETH DJI KEN SOM ERI Male Female Youth Non-Youth Source: WDI, 2020. KNBS, 2018. Note: Youth defined as people between 15-24 years of age. Source: WDI, 2020. KNBS, 2018. 6 HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM: OVERVIEW A considerable share of youth across the HoA percent in Kenya, 30 percent in Ethiopia, and 46 are Not in Employment, Education, or Training percent in Djibouti. Speeding the school-to-work (NEET).The share of young people who are transition is a legitimate policy concern in these NEET captures the extent to which youth are countries, but should be placed in the context of not substantially engaged in productive activity young urban workers accounting for only a small or activity that can increase their productivity in share of the labor force (two percent and seven the future (e.g. education), which represents a percent of the labor force in Ethiopia and Kenya). waste of human resources. As a result, NEET is often considered a measure of exclusion from Structural transformation is incipient - for most productive activities. Around half (49 percent) workers in the Horn, a job still comes down of youth in Djibouti, 32 percent in Somalia, 13 to self-employment or work in household percent in Kenya and 10 percent in Ethiopia are activities. Self-employment and unpaid family NEET (Figure 9). Young women are particularly work dominate employment in Ethiopia, Kenya, at risk of being NEET. Open unemployment is Somalia, and Eritrea, while Djibouti has higher relatively low in the HoA (except for Djibouti). As rates of wage-employment (Figure 10). Women with participation rates, however, this does not are far more likely to be engaged in unpaid reflect a strong labor market, but rather the rural family work then men and are underrepresented nature of most countries in the HoA and the need in wage-employment, though the pattern is for people to work in the absence of formal social more equitable for the younger generations assistance systems. Unemployment is significantly (Figure 11). The incidence of wage employment higher in urban areas, especially among youth: 17 is closely related to the sectoral composition of employment. In Ethiopia, where agriculture Figure 9: Youth NEET and unemployment are accounts for the bulk of employment, wage- substantial across the HoA employment is particularly rare (Figure 12). In countries with a larger employment share in (Share of youth 15-29 Not in Employment, Education services, wage-employment is more common or Training or unemployed) (Djibouti, Kenya, Somalia), though even in the 0.60 services sector self-employment tends to be at 0.49 0.50 0.46 least as important as wage-employment. There 0.45 are few differences in the sector of employment 0.40 0.32 between youth and older adults, but youth are 0.30 0.30 much more likely to work as unpaid family labor. 0.20 0.18 0.17 0.13 0.08 As a result of it being at only an early stage, 0.10 0.10 0.07 0.04 0.07 0.05 0.04 0.04 0.03 accelerating the region’s economic transformation 0.02 0.00 has the potential to generate large income gains. Ethiopia Djibouti Kenya Somalia Eritrea Even in Kenya, where structural transformation Non-youth unemployment Youth unemployment and industrialization have progressed the most, Urban youth unemployment Youth NEET cross-sectoral earnings differentials remain large: the average earnings of self-employed workers in Note: Youth defined as people between 15-24 years of age. the informal and modern sectors are 2.1 and 4.0 Source: WB staff calculations on latest household surveys. times those of self-employed agricultural workers. 7 SECTION 2 | BACKGROUND CONTEXT Figure 10: Self-employment and unpaid work Figure 11: Women are under-represented in account for the bulk of employment in the HoA wage employment (Share of employment type in overall employment) (Share of women in total wage employment, %) 100% 12% 12% 60% 90% 27% 33% 80% 50% 47% 28% 70% 46% 40% 60% 50% 43% 30% 43% 40% 40% 30% 20% 50% 20% 41% 29% 10% 10% 24% 19% 0% 0% Ethiopia Djibouti Kenya Somalia Eritrea Djibouti Ethiopia Kenya Somalia Unpaid/family Self-employed/employer Young workers (15-24) Adult workers (25-65) Wage employed Other All workers Source: UNDP, 2019. Source: WDI, 2020. KNBS, 2018. Figure 12: Agriculture and services account for the bulk of employment (Sectoral distribution of employment) 100% 7 10 13 28 7 25 80% 6 18 8 5 8 16 7 60% 14 44 11 17 8 40% 71 11 20% 47 44 17 29 5 0% Ethiopia Djibouti Kenya Somalia Eritrea Agriculture Industry Commerce Transport & Communication Finance & Business Public Administration Personal services and other Source: UNDP, 2019. Rapid population growth The employment challenge in the HoA is average number of dependents supported by compounded by its rapidly growing population. their incomes falls sharply. However, this is not The working-age population in the HoA is the case in the HoA because fertility rates and projected to grow from 107 million in 2020 dependency ratios remain high and are declining to 143 million by 2030, leading to a surge in at only a moderate pace. While the share of the demand for jobs and economic opportunities. working-age population is projected to increase A rapidly growing working age population can modestly from 57 percent in 2020 to 60 percent be a boon for economic growth through the by 2030, each successive age cohort will still “demographic dividend” when working age remain bigger than the previous one, meaning that adults are economically productive and the a HoA-wide demographic dividend is not around 8 HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM: OVERVIEW the corner (Figure 13).10 Lowering fertility rates and for the HoA,11 along with generating productive strengthening the institutions necessary to realize employment opportunities for a burgeoning and the demographic dividend should be a priority relatively poorly educated labor force. Figure 13: The demographic transition is still incipient in the HoA (Population pyramid of the HoA, 2020 and 2030) 80+ 70-74 60-64 50-54 40-44 30-34 20-24 10-14 0-4 20,000 15,000 10,000 5,000 0 5,000 10,000 15,000 20,000 Male 2020 Female 2020 Male 2030 Female 2030 Source: WPP, 2019. WB staff calculations. Weak trade performance Trade is underperforming as a driver of growth, manufactured exports are typically more important job creation and poverty reduction. The ratio that agricultural exports. For both Ethiopia and of (recorded) trade to GDP decreased between Kenya, the share of manufactures in exports to 2013 and 2019 in Ethiopia, Kenya, and Djibouti, HoA countries is almost double that in trade with and for countries in the region trade plays a the rest of the world. Almost all of Kenya’s exports role in the economy considerably below what to the region are industrial products. The limited would be expected based on their economic size data suggest that Ethiopia is the major market in the and location. In general, HoA countries export region for HoA countries’ exports of manufactures agricultural products to the global economy and and the main exporter of agricultural products to primarily import manufactures. Official statistics the region. Somalia, the main import market for suggest that (except for Somalia) there is little intra-HoA trade, primarily purchases vegetable intra-HoA trade. For recorded intra-HoA trade, products and processed food from the region. 10 A common misperception is that a large and growing youth population itself indicates a coming demographic dividend. It does not. The first step is a transition from high to low birth and death rates (Population Reference Bureau, 2012). Associated development impacts are strongest when this happens quickly, and the age structure shifts abruptly. In the HoA, however, the demographic transition is taking place more slowly. 11 Research shows that the demographic transition will not generate a demographic dividend if quality institutions are not in place. These institutions include basic health care and schooling, infrastructure, rule of law and efficiency of the bureaucracy, etc. (Lee, Lee and Mason, 2006; Bloom, Canning and Sevilla, 2003). 9 SECTION 2 | BACKGROUND CONTEXT Figure 14: Recorded Trade-to-GDP Ratios and Export Structures in the HoA A. Trade (% of GDP) B. Export Structure HoA, 2018 0% 3% 3% 400 367 100% 12% 6% 348 1% 350 319 9% 27% 22% 298 80% 34% Trade (% of GDP) 300 11% 14% 61% 250 210 60% 200 165 150 130124 40% 68% 72% 59% 11% 63% 100 4129 53 50 50 20% 50 33 24% 0 0% Djibouti Ethiopia Kenya Somalia Vietnam Cambodia Djibouti Ethiopia Singapore Eritrea Kenya Somalia 2013 2019 Raw materials Intermediate goods Consumer goods Capital goods Source: staff calculations using data from United Nations Comtrade Database. Unrecorded cross-border trade is an important both within the region and with third countries. For phenomenon in the borderland regions of the example, Ethiopia has the largest cattle inventory Horn of Africa. Beyond trade that is recorded in Africa and the 5th largest globally. Kenya and at the border by Customs Authorities, typically Somalia also have considerable herds of cattle undertaken by formal and larger enterprises, there and Somalia is home to the world’s largest camel is a vast amount of cross-border trade undertaken population. Map 1 provides a snapshot of the by small-scale traders and micro-enterprises that is main border crossings and documented trade not systematically recorded.12 Such trade supports routes. Since there is no systematically collected as many as 17 million people along a variety of information on cross-border trade and the different value chains, including crop farmers, conditions at border towns, analysis must currently brokers, crop traders, livestock-keepers, fodder be based on the collection of available but ad suppliers, ranch owners, itinerant traders, large hoc studies. Investments in data collection are livestock traders and transporters. Unrecorded therefore essential. However, surveys at the border cross-border trade may be as much as 20 times and novel techniques such as analysis of satellite greater than officially recorded intra-regional trade imagery and cell phone data records may provide in the Horn.13 Yet, there is considerable potential critical information by which to better understand to increase regional trade including in livestock, the nature of cross-border trade (see Box 1). 12 In the existing literature, many sources refer to this as ‘’informal cross-border trade (ICBT)’’. However, this often carries a negative and unwarranted connotation as ‘’informal’’ can be easily confused with ‘’illegal’’. It also inaccurately reflects the reality of trade flows on the ground, as traders may indistinctly use both formal and informal crossing channels depending on a variety of factors, such as the nature of the goods, the value of their consignment, the length of the queue at the border, or the mood of the individual official on duty. The preferred terminology here is “unrecorded trade” as this commerce is either missed or under-represented in official (customs) collected data at the border. 13 Official data on cross-border trade is only available for Ethiopia and Kenya. These suggest that intra-HoA trade amounts to just under one percent of total trade. Taking currently unrecorded cross-border trade into account gives a very different perspective on the importance of regional trade in the HoA. 10 HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM: OVERVIEW Map 1: Commodities and livestock trade however, is to enhance state capacity in ways routes in the HOA which enable trade to flourish, including by clarifying, harmonizing and consistently applying the rules affecting how goods and services flow across the borders in the region. Further, convergence in macroeconomic policies in the region over time, especially reducing exchange rate rigidities which may distort incentives and lead to deadweight losses, would boost gains from trade. Cross-border trade in the Horn plays an important role in food security, improvements in smallholder income and poverty reduction. Results from a survey of 200 traders at two Trade flow Capital border crossings (between Ethiopia and Somalia Market ADM2 boundary and between Ethiopia and Kenya) show that cross-border trading activities provide the main Source: Brenton & Edjigu (2021). source of income for 78% of the traders at the Trade, including the high volume of unrecorded Kenya border and 98% of those at the Somalia trade, is affected by macroeconomic and border. Previous studies have found that cross- trade policies, and differences in their border livestock trade enables herders to provide implementation, amongst the HoA countries. for their children’s education, housing and other Exchange rate arrangements differ widely across basic needs and also increases the incomes of Djibouti (currency board), Eritrea (fixed peg), traders, trekkers, fodder produces, brokers and Ethiopia (soft peg), Kenya (managed float) and other services, including veterinary services, Somalia (free floating, but with a high degree hotels and restaurants among others.15 Cross of de facto dollarization).14 There are wide border trade enhances food security through the disparities in tariff rates and customs duties, movement of grain and other food items from and in the extent to which they are enforced. surplus areas to deficit areas and by increasing The general lack of systematic regulatory the price incentives to production in those enforcement in Somalia, in particular, has led areas where efficiency and yields are highest.16 to the country acting as an entrepôt for the Income from cross border trade can also be an importation of goods into surrounding countries. important source of saving and capital to startup The interplay between lax regulations and the new businesses or strengthen livelihood activities importance of small-scale trade in supporting which can provide additional sources of income, livelihoods fuels perceptions that a lot of trade employment and security for the household and is illicit (cf. footnote 12). The real challenge, a more diversified local economy. 14 Source (except for Eritrea): IMF Annual Report on Exchange Arrangements and Exchange Restrictions (2019). 15 Tesfaye, A. and N Amaha (2018) ‘A Review on Cross-Border Livestock Trade Across Dry Land Borders of Ethiopia: The Trends and Implications’, Journal of Scientific and Innovative Research 7(2): 36-42 16 There is some evidence from Ethiopia that higher crop prices are associated with reductions in child stunting, with long-term development implications in terms of attainment at school and productivity in the workplace. (Brenton, P and M. Nyawo (2019). Food Prices, Access to Markets and Child Undernutrition in Ethiopia. Policy Research Working Paper 8823, World Bank, Washington, DC) 11 SECTION 2 | BACKGROUND CONTEXT Box 1: Monitoring border infrastructure and trade flows through satellite imagery and anonymized cell phone records A pilot analysis at two border crossings between market was open, typically spent most of the day Ethiopia and Somalia shows how, in the absence at the market, and spent multiple days in a row at of systematically collected data on trade and the market. trade-related infrastructure, satellite imagery and cell phone data can be used to identify the Figure 15 provides an example of the application size and quality of relevant structures in border of this approach to Tog Wajaale. This is a key towns and monitor trade flows. The analysis can border crossing between Somalia and Ethiopia that identify market locations, market size, the total appears to have been fully open and functioning number of stalls (uncovered and covered), the during the 2019/2020 period, based on human extent of available lighting, the total number of movement analysis and satellite imagery. This is market buildings, WASH facilities (well, reservoir, consistent with reporting by IOM, which monitors water tower, or water tank) and functionality and 14 border crossings between Ethiopia and Somalia, key roads (including type, maintenance status that all but two (Tog Wajaale and Dolo Ado) are and flood risk). The structural information can be designated as “closed for entry and exit” as of complemented by anonymized cell phone data 2020/2021. The human movement analysis from to analyze foot traffic around markets and border cell phone records suggests that between 2019 crossings. This can identify market-site activity, the and 2020, travel through the Tog Wajaale border change in volume of market participants between crossing increased by 29%, and over half the people different periods, routes traveled by traders, the who visited the Jijiga markets made multiple (2+) volume of cross-border traffic, changes in trade border crossing trips through Tog Wajaale. Hence, over time based on cell phone locations, pings, and it appears that with border closures elsewhere, at unique travel paths. Traders are identified as those least a portion of the trade through those routes who were within or around the market while the was re-directed through Tog Wajaale. Figure 15: Infrastructure and activity at Tog Wajaale border point Source: IPSOS, 2021. 12 HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM: OVERVIEW Tog Wajaale is a split city, with some of the urban occurred during Al Shabaab’s control and ultimate area lying in Ethiopia and some in Somalia. Analysis ousting from the area between 2006 and 2011. This of anonymized cell phone data shows that visitors damage may act as a security countermeasure in to the Somali side of Tog Wajaale tend to cluster a climate of persistent insecurity with occasional and spend most of their time in the main market clashes between rival clan militias along the border. area. This commercial market area may also be a As a result of the damage, the road is effectively a key draw to this town (for purchase/sale of goods single lane and conditions are severe enough to beyond cattle as well). Previous studies have shown prevent a consistent flow of travel, especially of that Tog Wajaale is a key node for cattle trade, with vehicles transporting goods. many cattle traded in its market originating from the Oromia region of Ethiopia, and that the construction Nevertheless, satellite imagery analysis identifies of quarantine stations has facilitated such trade. The what appears to be likely informal trading of cell phone data point to a potential quarantine area cattle and livestock along the riverbank. Clearly identified through imagery analysis, confirming that defined man-made watering holes are apparent in cattle trade infrastructure is important in attracting the riverbed as well as foot paths from one bank to additional flows. another and goats and cattle are visible gathering around a man-made watering hole .75 km south of In contrast, the border between Ethiopia and the Ferfer bridge, with what appear to be livestock Somalia at Ferfer-Beledweyne shows much less vehicles close by. Human mobility data volume cross-border activity. Ferfer is a small town, with is very low for this area, but it does demonstrate a bridge across a dry riverbed connecting it to a limited activity taking place across the bridge and cluster of buildings. This bridge appears to have within surrounding areas, with a slight uptick in likely border crossing infrastructure on either side 2020. This is consistent with the imagery findings, (“likely” since there is a lack of other reporting to that demonstrate difficulty to transport a heavy cross-verify this). While the bridge is intact, imagery flow of goods over the bridge, but a low level of analysis (Figure 16) reveals that there is noticeable ongoing trade taking place around the dry riverbed. damage – enough to inhibit the passage of heavy In sum, this crossing is currently not along a primary commercial vehicles. On the Somalia side, the corridor for flow of goods and trade, though there analysis identified significant road damage that likely does appear to be informal livestock trade. Figure 16: Little infrastructure and activity at Ferfer border point Source: IPSOS, 2021. 13 SECTION 2 | BACKGROUND CONTEXT Hence, this pilot approach has shown how analysis over time and information on the type and extent of satellite imagery and cell phone records can of market related infrastructure. Looking forward, if help improve understanding of where important some of the processes can be automated then the cross-border markets are situated, provide an approach could be useful for a wide area assessment indicator of trade volume and changes in trade of market locations and catchment areas. The livestock value chain is one of the most for reselling to the butchers. Exporters purchase important in the Horn of Africa and has developed animals and further fatten to export to overseas into a complex structure involving a wide markets. However, while this summarizes the range of stakeholders. These include producers broad structure of the value chain, processes and (herders), brokers, feed and water suppliers, traders, interactions between actors in the value chain are transporters/trekkers, processors, exporters and casual and change to suit the nature of the border. consumers. The general value chain for livestock starts with the collection of animals from farm A key feature of the livestock value chain in gates moving on to local primary markets the region is the importance of clan-based (collection markets), and then to secondary and trust-based trading. Cross-border trade in markets (regrouping markets) where livestock are the Horn of Africa is sustained by high levels of regrouped and sorted into different classes based social capital among key market actors. Livestock on appearance, size, color and conditions and sales are transacted mainly based on ethnic and then on to terminal markets and final buyers. If familial ties reflecting the uncertain business a primary market is near a border, cross-border environment, absence of strong institutional or traders purchase animals from farmers for onward judicial intermediation and formal systems of sale in a terminal market, which may be in an credit enforcement, weak infrastructure, limited overseas country. Secondary markets are relatively market support services and prevalent insecurity larger than primary markets, usually they are the in pastoral areas. In the absence of ethnic ties, a main trading center in a district. The main actors in trust-based network of pastoralists, intermediates, these markets include pastoralists, local butchers, traders and final buyers is important. Though clan- middlemen and export traders. Pastoralists may based and trust-based marketing facilitate trade, directly sell their animals at secondary markets or they can also distort the livestock market if the purchase breeding stock. Brokers purchase animals result is a clan-controlled monopoly. 14 HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM: OVERVIEW SECTION 3 The economic geography of the HoA 3.1 Density: Spatial Distribution of Economic Activity and Living Standards Economic activity in the HoA is spatially and stretching westwards towards Kisumu and concentrated. In countries with a substantial the shores of Lake Victoria. It is also in these agricultural base (Ethiopia, Kenya and, to a lesser agricultural breadbaskets that the countries’ extent, Eritrea), economic activity tends to be biggest cities are located. In Djibouti and Somalia, concentrated in places with better agricultural both arid and hot countries, economic activity potential (Figure 17 and Figure 18): The central is concentrated in places with a comparative highlands of Ethiopia crossing north into Eritrea advantage for trade through seaports: Djibouti towards Asmara and crossing east into Somalia city in Djibouti and Mogadishu (and to a lesser towards Hargeisa, and the area around Nairobi extent Kismayo) in Somalia. Figure 17: Economic production in the HoA is Figure 18: And largely reflects cropland availability highly concentrated (Estimate of GDP per squared kilometer, 2006) (Cropland area in km² per 5 arc minutes grid cell) Sources: Ghosh et al. (2010). Sources: HYDE3.2. Population density and economic density largely territory (World Bank, 2019). In Kenya, 20 percent overlap. In Ethiopia, close to 90 percent of the of counties – which are densely populated and population live in the highlands, despite the concentrated spatially in the southwestern part highlands accounting for less than half of Ethiopia’s of the country – account for over 40 percent of 15 SECTION 3 | THE ECONOMIC GEOGRAPHY OF THE HOA population17 while in Eritrea, it is estimated that The spatial distribution of economic density is the top 10 most populated districts, which are reflected in spatial differences in living standards. located mostly in the northeast, in and around the Using the US$1.9 a day poverty line, poverty country’s capital, account for around 40 percent rates are generally lower in the economically of population.18 In Somalia, high densities of dense areas than in the economically-lagging population are observed in Mogadishu, Kismayo, ones (Figure 19). There are several ways to Baidoa and Hargeysa and along major road classify economically dense (or “leading”) areas networks that are connected to these major cities. vs. lagging areas. Poverty rates in rural areas of the HoA (39 percent) are close to three times The vast and sparsely populated border areas are a higher than in urban areas (15 percent), while salient feature of the HoA. The HoA border areas,19 poverty in the border areas is 12 percentage the area within the red lines in Figure 17, account for points higher than poverty in the non-border or over 44 percent of the combined territory of the core areas. Given the sparsely populated nature five countries but make up only 13 percent of total of the border areas, the density of poverty is not population and 9 percent of economic output. A situated in the border areas; most of the poor large swath of the borderlands is classified pastoral/ live in the wealthier (and hence more populated) grazing land (see Figure 18). A large majority of parts of these countries (Figure 20). Other social people in the borderlands engage in pastoralism, indicators show large spatial disparities as well. agropastoralism and trade. The border areas are The share of 7-14-year-olds attending school, the also where many of the Horn’s challenges come share of youth aged 15-24 that have completed together: a legacy of economic, social, and primary education, and the share of youth 15-20 political marginalization; poor service delivery attending secondary education are substantially and infrastructure; conflict, violence, and forced lower in rural areas and in border areas for all displacement; and environmental degradation, all three countries, as opposed to urban areas and spilling across national boundaries. While there are core areas, respectively (Figure 21 and Figure 22). differences across countries, the reach of the state in Indicators of access to health care and health these areas is generally weak, and informal and illicit outcomes also tend to be substantially worse activities flourish. Yet these border areas are bound in the economically lagging areas, translating to play a key role if trade and cooperation across into multi-dimensional poverty rates that are the HoA is to increase and deepen, highlighting their substantially higher in the border areas than in importance in the regional integration agenda. the economically dense areas. 17 Out of 47 counties, the most populated 10 accounted for 41 percent of the population in the 2019 population census. 18 The World Bank’s estimates based on population data from WorldPop and GADM ADM2 shapefile. 19 Border areas are defined here as second-order administrative divisions (ADM2) that share a land border with another country, with the exception of the regions of Tigray and Djibouti which have distinctly different demographic and environmental characteristics and are thus not characterized as border areas. 16 HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM: OVERVIEW Figure 19: Poverty rates tend to be higher in the Figure 20: Though most of the poor live in the sparsely populated border areas economically-dense areas (Share of population below the US$1.9 poverty line) (Number of people below the US$1.9 poverty line) Source: World Bank staff calculations using the latest household Source: World Bank staff calculations using the latest household survey available in each country. survey available in each country. Figure 21: Education outcomes lag in rural areas… (Selected education indicators, Djibouti, Ethiopia, Kenya) 100 90 80 70 60 50 40 30 20 10 0 Attending Completed Attending Attending Completed Attending Attending Completed Attending school primary school secondary school primary school secondary school primary school secondary (7-14) (15-24) school (15-24) (7-14) (15-24) school (15-24) (7-14) (15-24) school (15-24) Djibouti Ethiopia Kenya Rural Urban Source: GMD, 2020; World Bank staff calculations. 17 SECTION 3 | THE ECONOMIC GEOGRAPHY OF THE HOA Figure 22: …and also in the border areas (Selected education indicators, Djibouti, Ethiopia, Kenya) 100 90 80 70 60 50 40 30 20 10 0 Attending Completed Attending Attending Completed Attending Attending Completed Attending school primary school secondary school primary school secondary school primary school secondary (7-14) (15-24) school (15-24) (7-14) (15-24) school (15-24) (7-14) (15-24) school (15-24) Djibouti Ethiopia Kenya Border areas Core areas Source: GMD, 2020; World Bank staff calculations. As countries develop from a low base, economic the same pattern, with places that were “medium concentration tends to further increase, widening built-up” to begin with experiencing faster growth the economic gap between leading and lagging in built-up area compared to unbuilt and highly areas. The Horn has also experienced this pattern, built-up places (Figure 24). Formally testing for though in a more nuanced way. In the absence of beta-convergence shows that economic activity regional output data, proxy indicators suggest that has diverged since the early 1990s, with more economic activity in the Horn has grown fastest developed places growing faster.20 Border areas for places in the middle of the baseline economic have experienced the slowest growth, even activity distribution. Night-time lights, a commonly after controlling for their lower levels of initial used proxy for economic production, increased luminosity.21 The non-linear pattern in Figure 23 strongest in places that were in the middle of and Figure 24 was partly driven by faster growth in the nightlights distribution in 1992, while both secondary cities than in large cities and towns (Box unlit and highly lit places (roughly corresponding 2). Data from Kenya and Ethiopia suggest that the to rural areas and big cities, respectively) grew divergence in economic activity was nevertheless slower (Figure 23). The growth in built-up area, associated with modest spatial convergence in another proxy indicator for development, shows living standards.22 20 Beta-convergence occurs when poor geographical units grow faster than richer geographical units. Divergence occurs when poor places grow slower than rich places. We estimate beta-convergence based on a regression of annual growth in NTL based on initial NTL. The coefficient on initial NTL is positive and statistically significant, indicating that places that were better lit in 1992 grew significantly faster. The same result is obtained when using built-up area instead of NTL. 21 The border dummy in the convergence regression is negative and statistically significant. 22 Kenya and Ethiopia are the only countries in the HoA with at least 2 comparable household poverty surveys since the early 2000s. In both countries, the share of total inequality in household consumption that can be explained by welfare gaps between regions decreased between 2005 and 2015, pointing towards narrowing spatial welfare gaps during this period. 18 HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM: OVERVIEW Figure 23: Luminosity grew fastest in places Figure 24: Growth in built-up area was higher that were medium lit in 1992 in places that were more built-up to begin with (Growth in night-time lights between 1992 and 2013 (Growth in built-up area between 1990 and 2014 by by intensity of night-time lights in 1992) density of built-up area in 1990) .15 .1 Annual growth in built-up Annual nighttime light densities (1990-2014) growth (1992-2013) .08 .1 .06 .04 .05 .02 0 0 0 1 2 3 4 0 1 2 3 4 Initial level of nighttime light in 1992 Initial level of built-up densities in 1990 Non-border areas Border areas Quadratic fit Non-border areas Border areas Quadratic fit Source: World Bank staff calculations on DMSP nighttime light data. Source: World Bank staff calculations on GHS built-up dataset. Box 2: Urbanization, but not metropolization Most countries in the HoA are urbanizing rapidly data is hard to come by, an analysis based on from a low base. The estimated urban population nighttime luminosity suggest that secondary in the HoA increased from 20 million in 2000 to 50 cities have been growing faster than capitals and million in 2020 and is expected to increase to 90 other large cities and, especially, towns and small million by 2035. In parallel, the urbanization rate cities, contributing to the observed non-linear is projected to increase to 35 percent, compared association between initial levels of economic to 26 percent in 2020.23 Though comparable development and subsequent growth (Figure 25). Figure 25: Growth in nighttime lights was Figure 26: Towns and secondary cities will faster in secondary cities and large cities account for the bulk of Ethiopia’s urban than in smaller towns population growth (Growth in night-time lights between 1992 and (Projected increase in urban population by city 2013 by city size) size, Ethiopia) 52.6 .06 50 5.81 Annual nighttime light growth (1992-2013) Population (million) 40 31.1 20.242 .04 4.561 30 17.5 9.225 5.593 .02 20 11.9 3.273 3.488 3.284 2.74 1.782 20.929 2.276 14.002 10 1.139 8.916 0 5.708 Small Secondary Primary/ large 0 2007 2015 2025 2035 cities cities cities Less than 50,000 50,000 to 100,000 100,000 to 500,000 Addis Ababa Source: Source: World Bank staff calculations on GHS built- Source: Schmidt et al., 2019. up dataset. 23 Figures based on World Urbanization Prospects (UN, 2018). 19 SECTION 3 | THE ECONOMIC GEOGRAPHY OF THE HOA Projections for the HoA’s biggest economies urban population growth will take places in cities suggest that, going forward, most of the urban of between 100,000 and 1 million people, while population increase will take place in towns the urban share of Nairobi is projected to remain and secondary cities. In Ethiopia (Figure 26), constant. These projections call for much-needed the bulk of urban population growth will happen investments in infrastructure and public service in small towns (less than 50,000 people) and provision in smaller cities and towns across the secondary cities (100,000-500,000 people). region, where infrastructure and service provision Similar projections for Kenya show that most of the are currently lagging. 3.2 Distance: Access to markets and integration The concentration of economic activity in dense places, such as Addis Ababa, Nairobi and certain favored places is a normal phenomenon. the southeastern part of Kenya, and the dense The extent to which other, less-favored, places cross-border corridor between Jijiga (Ethiopia) benefit from the dense clustering of economic and Hargeisa (Somalia). In contrast, market activity in the favored places depends on the access is low, and economic distance high, in economic distance that separates them.24 In the most of Somalia and Eritrea and in the peripheral HoA, economic density and economic distance lowlands of Ethiopia and Kenya (though even are strongly inversely correlated (Figure 27): in Kenya’s northeast market access, while low market access, an inverse proxy for economic compared to the rest of Kenya, is still better than distance, is high in and around economically in most other countries). Figure 27: Economic distance is high in the Figure 28: Places in the HoA with high economic border areas and in most of Somalia and Eritrea distance are poorer 1 .8 Poverty rate .6 .4 .2 0 0 .2 .4 .6 .8 Market access index (MAI) Poverty rate Lowess line Notes: Sample correlation amounts to -0.24 and the linear relationship is statistically significant at the 1% level and robust to outliers (such as Nairobi on the far right). The unit of analysis is the mixture of the first- and second-administrative areas at which household survey was Sources: World Bank staff calculation based on OpenStreetMap representative. and WorldPop (2015 population estimates). Source: World Bank staff calculations on countries’ latest poverty surveys. 24 “Distance” here does not refer to physical distance, but to the ease or difficulty for goods, services, labor, and capital to move between the leading places and the other places. It measures how easily capital flows, labor moves, goods are transported, and services are delivered between two locations. Distance here is an economic concept, not just a physical one. 20 HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM: OVERVIEW Economic distance is strongly correlated with to opportunities. Access to basic opportunities poverty rates. The farther it is from density, the for children in the HoA is low and uneven, more likely an area is to be lagging because of its with substantial horizontal inequalities when it lack of integration in the economy of the leading comes to completing primary school, attending areas. This is often reflected in poverty rates secondary school, and having access to clean being higher in remote and poorly connected water and electricity (Figure 30 and Box 3). parts of a country. The HoA is no different: market Significant shares of the inequity in access to key access, an indicator of economic distance, is opportunities can be explained by location – that strongly correlated with poverty rates. Places is, economic distance. Rural vs urban location is with better market access have lower poverty a major contributor to access to opportunities rates and vice versa (Figure 28). Rural areas in the in the Horn, with children in rural areas, where Horn that are better connected, as measured economic distance is high, having far lower by the Rural Accessibility Index,25 also achieve access to important opportunities relative to higher economic growth rates (proxied by children in urban areas (Figure 29). Living in a changes in nighttime light) than places that are lagging border area has a sometimes significant poorly connected.26 residual effect: living in a border area explains 37 percent of the inequity in attending primary Economic distance is not just a theoretical school. Reducing economic distance is key to construct; in the Horn of Africa, it shapes access expanding opportunities in the HoA. Figure 29: Significant gaps in coverage of basic Figure 30: Location and wealth shape access to opportunities opportunities in the HoA (Coverage rate and human opportunity index, Djibouti, (Contribution to inequity in access to a given Ethiopia, Kenya) opportunity, Djibouti, Ethiopia, Kenya) 100 90 Electricity (0-14) 80 70 Improved water source 60 (0-14) 50 40 Attending secondary 30 school (15-24) 20 10 0 Completing primary DJI ETH KEN DJI ETH KEN DJI ETH KEN DJI ETH KEN DJI ETH KEN school (15-24) Attending Completing Attending Improved Electricity Attending school (7-14) school (7-14) primary secondary water source (0-14) school school (0-14) 0 20 40 60 80 100 (15-24) (15-24) Sex of child Sex of HH head Wealth HOI Coverage Rural vs urban Border Sources: World Bank staff calculations on countries’ latest Sources: World Bank staff calculations based on countries’ latest poverty surveys. poverty surveys. 25 See Roberts, Peter; KC, Shyam; Rastogi, Cordula. 2006. Rural Access Index: A Key Development Indicator. Transport paper series; no. TP-10. World Bank, Washington, DC. 26 See Bundervoet & Masaki (2021). 21 SECTION 3 | THE ECONOMIC GEOGRAPHY OF THE HOA High economic distance in the border areas can and Djibouti-Dire Dawa connections, which are be explained by long physical distances to the mainly driven by air and rail connectivity.27 The domestic centers of density and weak cross- lack of regional economic integration results in, border connectivity. Transport connectivity in and also results from, the lack of connectivity the HoA can best be described as a collection between the HoA’s major cities. Despite domestic of national spaces. Within a network analysis connectivity being a lot stronger than cross- framework that takes account of road, rail, and country connectivity, domestic connectivity air connections, within-country connectivity in remains weak in Somalia and Eritrea. In Eritrea, the the HoA is more than 15 times stronger than city of Assab is presently isolated on all possible between-country connectivity. In the top 10 fronts in the face of poor connections between of most connected city-pairs in the Horn, only Assab and the major centers in northern Eritrea, two straddle national borders: the Djibouti-Addis in spite of its potential importance as a port. Box 3: The Human Opportunity Index The Human Opportunity Index (HOI) is widely level of a basic service or “opportunity” (for example used to measure inequality of opportunity. The whether a child is enrolled in primary education) HOI captures both (i) the overall access to basic and combines this with the extent to which that services, such as education, water and electricity opportunity is determined to be beyond the control (the coverage rate); and (ii) inequality in access of the child (for example being born in a rural rather (Barros, Ferreira et al. (2008)). If access to a basic than urban area or being a girl rather than a boy). service is perfectly equal, then the HOI is the same Ideally, random circumstances should play no role as the coverage rate. As access becomes more in determining access to opportunities. and more unequal, so the HOI becomes lower and lower. The D-index measures dissimilar access rates to a given basic opportunity for groups of children The extent of inequality of opportunity is where groups are defined by circumstance measured using the D-index (the dissimilarity characteristics (for example, area of residence, index). This index calculates how much access or gender) compared to the average access rate to services varies by birth characteristics, such to the same service for the population as a whole. as socio-economic status of a households and To formulate groups the sample is stratified into the location of the household. A D-index of zero groups or “cells,” so that all individuals in any given indicates perfect equality (no gaps in access to cell have the same combination of circumstances. services across circumstance groups), whereas a The resulting subgroups are known in the literature D-index of one indicates perfect inequality as “types” (Barros, Ferreira et al. (2008)). These cells are then compared to one another. The difference The central question behind the HOI is to what in outcomes between cells can be attributed to extent circumstances beyond one’s control inequality of opportunity, while the differences influence one’s access to a set of important basic within cells can be considered the result of effort services. Simply put, the HOI takes the coverage or luck. 27 The Nairobi-Mombasa connection is the strongest link in the HoA. 22 HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM: OVERVIEW The D-index summarizes all the gaps into a single the D-index is be zero. If average access is denoted measure by weighting them according to the by p ̅, the specific access rate of group i is pi , and population share in each circumstance group. The the share of group i in the population is given by D-index generates a value between 0 and 1. In a βi then the D-index is: society in which there is no inequality of access 1 n D= βi| pi - p ̅ | 2p ̅ i=1 The HOI can then be calculated as: HOI = p(1 - D) The measure is also decomposable so that the extent to which specific opportunity sets contribute to the dissimilarity can be assessed. This means that the contribution of different circumstances to overall inequality of opportunity can be determined. Figure 31: Transport connectivity in the HoA Figure 32: Road transport connectivity in the HoA (Strength of connectivity based on road, rail, and (Strength of connectivity based on road only) air connections) Source: Kunaka & Derudder (2021). Source: Kunaka & Derudder (2021). 23 SECTION 3 | THE ECONOMIC GEOGRAPHY OF THE HOA Cross-border connectivity in the HoA is mostly largest economies, requires attention. Beyond driven by air, which is far too costly for most the flights between Addis Ababa in Ethiopia and of the population. Focusing on road transport Nairobi and Mombasa in Kenya, and the Addis- only (Figure 32), the Djibouti-Dire Dawa-Hargeisa Nairobi road link via Moyale, there are essentially triangle is the only part of the HoA where no viable connections between the countries, multiple cities are reasonably well connected which share one single border post for an 861km- across borders. In contrast, there are only minor long border. The weakest connections in the HoA connections between Djibouti and Eritrea, Ethiopia are mainly found in the borderlands between and Eritrea, Kenya and Somalia, and Kenya and Ethiopia, Djibouti, and Eritrea and the borderlands Ethiopia. The low average connectivity between between Ethiopia, Somalia, and Kenya, reflecting cities in Ethiopia and Kenya, the region’s two and reproducing patterns of peripheralization. 3.3 Division: Weak cross-border connectivity compounded by fragility and thick borders Division refers to the restrictions on the flow of damaged roads and overloaded trucks that are goods, capital, people, and ideas across space. not specifically designed for carrying live animals, Division can both be domestic and international. At lead to stress and injury to animals, resulting in a the domestic level, nations can be internally divided rejection rate of up to 5% upon inspection at the across linguistic, ethnic, religious, or cultural lines. point of export. In the HoA countries for which At the international level, divisions mainly arise price data are available, a one percent increase from “thick” borders: the many restrictions some in travel time between towns is associated with a countries impose on the flow of goods, capital, nine percent increase in the difference in prices people, and ideas with other countries. Thick of homogenous agricultural products between borders limit trade and the flow of factors of towns.28 If these towns are located on different production. Interstate conflict creates the thickest sides of a border, price differentials increase by a borders. While borders in the rich world have further 12 percent, all else equal. This price effect become increasingly thin, thereby facilitating trade is substantial given the undifferentiated nature of and the movement of people and capital, borders the products and the high food budget shares in many poor countries remain thick. of households across the region. Cross-border price differences are highest between Djibouti The main sources of division in the Horn are and Somalia, and Djibouti and Ethiopia. related to weak transport infrastructure and connectivity, thick borders, and pervasive Cross-border economic integration is further conflict and fragility. Domestic transport constrained by the relative lack of border connectivity is weak in Eritrea and Somalia, infrastructure. Kenya and Ethiopia have one leading to poorly integrated markets and high border post for an 861-km-long border. Ethiopia within-country price differentials even for basic and Somalia share a 1,600-km border, served by commodities such as maize and sorghum. only 2 operational official border posts. In addition Poor road quality as well as lack of suitable to the relative dearth of operational border vehicles increase the cost of trade and limit the crossings, border infrastructure is often poor. As exploitation of scale economies. For example, shown in Figure 16, the border between Ethiopia 28 Maize and sorghum. 24 HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM: OVERVIEW and Somalia at Ferfer is characterized by a lack that interviewed 200 traders at two main border of infrastructure, with a visibly damaged road and crossings, one between Ethiopia and Kenya and waterholes dug in the dry riverbed to compensate one between Ethiopia and Somalia, found that for a lack of water infrastructure. Remote sensing the exchange rate and access to foreign currency, analysis shows little activity at this border crossing corruption and security were problems at both (which, according to humanitarian monitoring sites. Access to finance, the availability of market reports, was not operational in 2020), which information and the ability to find transport were effectively limits the accessibility of Beledweyne, also reported as key problems at the Ethiopia- a fairly sizable city in Somalia’s urban system,29 Kenya border. In contrast, at the Ethiopia- with connections to Mogadishu and northern Somalia border, taxes and their administration, Somalia via Garowe covering large distances at trade permits and customs authorities were limited speeds. In contrast, where borders are important challenges. For every issue that created operational and some infrastructure has been challenges at the border, the proportion of female built, trade has tended to flourish (such as along traders reporting it as an issue was larger than the the Jijiga-Hargeisa corridor, where cross-border proportion of male respondents. mobility increased by 29 percent between 2019 and 2020 and revenues for traders are higher than Social unrest, violence, and conflict, both along other border posts). domestic and interstate, have at times also posed formidable barriers to the flow of goods In addition to poor cross-border connectivity and people within countries in the Horn. While and infrastructure, there is a lack of transparency this has been most pervasive in Somalia, recent and awareness of policies being applied at the years have seen various episodes of unrest, borders. Information on the policy requirements to political turmoil, and open armed conflict in export and import are not readily available. This is Ethiopia, severely restricting the domestic flow of reflected in reports of ad hoc application of border goods and factors of production. To illustrate the measures, including exports bans and export taxes division caused by conflict, food price inflation on agricultural products and products such as amounted to more than 38 percent year-on- hides and skins. For example, at the border town year in Ethiopia’s Tigray region in December of Moyale in Ethiopia, in addition to a municipality 2020, more than double the rate observed in tax, local officials collect fees at town markets the neighboring regions and the country as a when an animal is sold and an additional tax if whole, and disproportionally impacting low- the animal crosses the border to Kenya. A lack of income households. understanding of cross-border trade by officials leads to the misperception that all small-scale Conflicts in the HoA have been spatially clustered trade is illicit, resulting in widespread threats of in the border areas. The underlying drivers of conflict confiscation of products. In reality though, a large and fragility in the border areas are multi-facetted proportion of cross-border traders are undertaking and complex, ranging from historical grievances the simplest of private sector activities in a hostile about the political and economic marginalization policy environment and are not seeking to evade of the periphery to tensions over scarce resources legitimate taxation and regulation by authorities to opportunities for illicit enrichment. Statistical that are legally responsible for border management. analysis of conflict events in the Horn shows that conflict is clustered in the lagging border areas Challenges at the border appear to be more (Figure 33) and has significant spillovers, with a restrictive for women traders. A pilot survey high degree of probability that conflict in one 29 With an estimated population of over 200,000, Beledweyne is the fifth biggest city in Somalia. 25 SECTION 3 | THE ECONOMIC GEOGRAPHY OF THE HOA region, both within countries and between them, Facilitating trade can be a powerful force for is related to the incidents of conflict in another stability in border areas and the Horn at large (usually neighboring) region. Fragility is correlated by promoting development and increasing the with economic distance, with conflicts clustered opportunity cost of conflict.31 Trade provides new in sparsely populated border areas with low market opportunities and creates jobs related to exports accessibility.30 Pervasive fragility in the border which can offer alternative sources of income for areas has led to frequent and sometimes long- those otherwise drawn towards violence and armed lasting closure of several border posts and trade groups. However, trade can also have negative routes, negatively affecting cross-border trade and, impacts on jobs in sectors that are subject to greater indeed, market access. The lack of a stable and competition from imports, and without support the regulated environment for cross-border trade has people who bear these losses may turn to violence contributed to the growth of illicit activities, with to sustain their income. Trade can also affect conflict the sustainability of these opportunities depending when it increases the value of economic resources, on the continuing marginalization and peripheral usually commodities such as minerals and precious status of border areas. Addressing fragility in the stones and timber, but also livestock and hence border areas will require addressing the complex land, and the incentives to fight over their control. web of interests in lucrative activities in these areas, Finally, taxation of trade may be an important source both by state and non-state actors. of revenue for both the government and armed organizations and changes in the structure and value Figure 33: Conflict is clustered in the border areas of trade may affect the ability to sustain conflict. In general, the evidence across a range of countries suggests that increasing trade with neighboring countries reduces both the duration and the intensity of conflict. There is also a lower risk of conflict when both countries are members of the same regional trade agreement. Reducing barriers to cross-border trade has been found to promote economic activity in border regions, which is sorely needed in the Horn’s lagging borderlands.32 Cross-border trade typically has strong backward and forward linkages to local producers and distribution markets, creating job opportunities for vulnerable youth in production, transport and logistics in border areas. Moreover, facilitating local cross-border trade through trade policy reforms and trade facilitation can be an Note: Hot spots of conflict based on ACLED conflict data. Hot spots identify statistically significant spatial clusters of high values of important means of addressing food security conflict (as measured by total fatalities between 2000 and 2019), issues faced by poor populations, which can often based on the Getis-Ord Gi* statistic. Source: ACLED, 2000-2019. be a factor behind increasing tensions. 30 This bivariate correlation however disappears if country effects are controlled for. In other words, the incidence of conflict is more related to country-wide influences and legacies than to differences in economic distance within countries. 31 The following draws heavily from Cali, M (2015) Trading Away from Conflict: Using Trade to Increase Resilience in Fragile States, World Bank, Washington https://www.worldbank.org/content/dam/Worldbank/document/Trade/Trading%20Away%20 from%20Conflict.pdf 32 See Brulhart, M., Cadot, O., Himbert, A., 2019. Let There Be Light: Trade and the Development of Border Regions. CEPR Discussion Papers 13515, CEPR Discussion Papers. 26 HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM: OVERVIEW 27 SECTION 4 | REGIONAL ECONOMIC INTEGRATION FOR JOB CREATION AND STRUCTURAL TRANSFORMATION SECTION 4 Regional economic integration for job creation and structural transformation The predominant challenge for the countries in Strengthening development outcomes will hinge the Horn is to generate enough adequate-quality on the extent to which the quantity and quality jobs through economic transformation, which of jobs keeps up with the growth of the working- will require massive human and physical capital age population, from 107 million at present to 143 investments. The region’s population is growing million by 2030. While the scale of the challenge fast, but the fertility rate is not yet falling fast enough is daunting (Box 4), there are also significant to deliver a demographic dividend; for at least the economic opportunities that could be generated by next decade and likely beyond, dependency ratios accelerating structural transformation, supporting will remain high, and each successive cohort will be trade, and reducing critical gaps and distortions in larger than the previous one. Major improvements product and factor markets (for example, gender in education and healthcare will be needed to gaps in regional labor markets, which currently accelerate the region’s demographic transition suppress labor force participation to only 56% in and expand economic opportunities for its youth. Somalia and 33% in Djibouti). Box 4: Projected labor force growth dwarfs the existing stock of “good jobs” With youthful populations, the labor force in the 3% per year, translating in large net increases in the region is expected to grow rapidly in the coming number of people available to work. In Ethiopia, decade. Table 1 shows estimates of this evolution, by far the region’s most populous nation, nearly based on UN population projections, and assuming two million more workers may be looking for a job that age and gender-specific labor force participation each year over the coming decade. In Kenya, the rates remain as measured in the most recent survey net increase is close to a million potential workers data. With the exception of Djibouti, all countries per year. If labor force participation rates were to are expected to see labor force growth rates around increase, these increases would be larger still. 28 HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM: OVERVIEW Table 1: Projected labor force increase in HoA countries Labor force Average annual net increase Growth rate 2020 2025 2030 2020-2025 2025-2030 2020-2030 Djibouti 255,000 280,000 300,000 5,000 4,000 1.7% Ethiopia 55,578,000 60,930,000 70,173,000 1,832,000 1,995,000 3.0% Kenya 23,472,000 27,438,000 31,574,000 793,000 827,000 3.0% Somalia 4,373,000 5,180,000 6,098,000 161,000 184,000 3.4% Eritrea 1,562,000 1,806,000 2,057,000 49,000 50,000 2.8% Source: WPP, 2019. WB staff calculations. The projected labor force growth dwarfs employment in industry and 34 percent of current the existing stock of “good jobs” in the HoA private sector employment. In Ethiopia, the annual countries.33 Annual labor force growth is expected projected labor force growth amounts to over half to be the equivalent of all current formal private of all existing private sector wage employment. Even sector wage employment within one to four years. though Ethiopia managed to create a lot of jobs In each country, the labor force is projected to in the industrial sector through focused support, grow within two to three years by as many workers the manufacturing share of urban employment has as are currently employed in industry (Table 2). In fallen from 12% in 2010 to 9% in 2018 , illustrating Kenya, for instance, the projected annual growth of the difficulty of increasing the share of good jobs the labor force amounts to 29 percent of current when the labor supply is increasing rapidly. Table 2: Annual labor force growth projections as a share of existing “good jobs” Annual net increase as a share of… Absolute annual increase Employment Formal private sector 2020-2025 in industry wage employment Djibouti 5,000 37% 65%a Ethiopia 1,832,000 39% 52% Kenya 793,000 29% 34%b Somalia 161,000 34% 89% Eritrea 49,000 53% 25%b a Increase shown relative to all current private wage employment, including informal. b Increase shown relative to all current formal private employment, including self-employment. Source: 33 Due to data limitations, “good jobs” are defined as wage jobs in the private formal sector. The public sector provides good and stable jobs too, but growth in public sector employment is constrained by fiscal space and, hence, not sustainable. Though not all informal self-employment jobs are bad jobs, the comparable data needed to categorize self-employment into “good” and “bad” self-employment do not exist. 29 SECTION 4 | REGIONAL ECONOMIC INTEGRATION FOR JOB CREATION AND STRUCTURAL TRANSFORMATION These simple figures highlight the multi-facetted will remain in agriculture and informal small- nature of the jobs challenge in the HoA. The scale business activities in services, highlighting accelerated creation of more private formal sector the policy importance of supporting broad- jobs is needed for structural transformation and to based productivity gains in these jobs, including accommodate the rapidly growing and youthful agriculture. Low participation rates and worse labor force, whose aspirations have been lifted employment outcomes for women call for policy by relatively higher education levels than the measures to address the structural and behavioral older generations, especially in urban areas. At barriers to women’s employment, further adding the same time, the bulk of employment is and to the complexity of the challenge. Achieving the economic transformation integration and investments in connectivity to required to generate sufficient higher improve access to markets across borders, productivity jobs and lift living standards in the realizing common opportunities to diversify, region will require a decisive upturn in the pace deliver jobs and reduce poverty and vulnerability. of private sector job creation. The economic At the same time, however, implementing the priorities in the Horn are consistent with the jobs JET agenda in the HoA will require dramatic and economic transformation (JET) agenda, improvements in education and other which recognizes the role of connecting to investments to build capabilities and overcome markets in generating job-creating private significant cross-border and interconnected investment, and the importance of building risks, including currently poor human capital capabilities and linking workers to jobs (Figure outcomes, large and rapidly growing numbers of 34). In the Horn of Africa, there is an opportunity youth, fragility, and environmental degradation to drive the JET agenda through deeper regional and climate change. Figure 34: The Jobs and Economic Transformation Framework Jobs & Sustainable, Inclusive Growth Job-creating Private Economic More Productive Workers Investment Transformation & Entrepreneurs Building Blocks Creating & Connecting Building Capabilities & (what) to Markets Connecting Workers to Jobs Address new technologies - Support the digital economy and green jobs Integration into Global Value Chains Approach Quality infrastructure to connect and power the economy (how) Maximizing Finance for Development - Whole of WBG Approach Adopt a jobs lens to WBG interventions and measure impact Source: World Bank, 2019. 30 HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM: OVERVIEW 4.1 Building capabilities and connecting workers to jobs A first-order priority in the HoA is to improve before, poor outcomes persist for the youth as human capital outcomes of the young and rapidly well (Figure 35). In Somalia, 43 percent of young growing populations. Returns to education in people have not had any education. In Ethiopia, the region are high, driven by the large earnings with a Human Capital Index of 0.38, less than 20 gains individuals are able to realize when equipped percent of young workers achieved more than with the skills and capabilities needed to move primary education.34 Raising worker productivity into higher-productivity activities and sectors. and improving access to improved earning While countries in the Horn have made progress opportunities will require strong improvements in on getting more children in school, completion basic education, especially given the high returns of primary school, progression to secondary to primary schooling in low-income countries. school, and quality of learning have lagged. Improvements in education for girls are especially While the young generation generally has better important if countries in the Horn hope to reap a educational outcomes compared to generations demographic dividend. Figure 35: Educational attainment remains low in the Horn (Educational attainment of the working-age population) 3% 5% 11% 9% 5% 5% 6% 6% 100% 16% 8% 9% 80% 18% 37% 25% 12% 29% 48% 14% 60% 57% 60% 26% 40% 73% 58% 59% 47% 10% 44% 43% 20% 25% 20% 3% 11% 0% Youth Non-Youth Youth Non-Youth Youth Non-Youth Youth Non-Youth Ethiopia Djibouti Kenya Somalia No education Primary (including incomplete) Secondary (including incomplete) Tertiary (including incomplete) Source: WPP, 2019. WB staff calculations. Improving human capital outcomes will require next, creating persistent pockets of poverty and enhancing the spatial and social equity in fragility. Provision of basic public services, including public service delivery through spatially blind education and health care, should be spatially blind institutions. Location and household wealth are and universal in coverage. In the border areas, key determinants of educational outcomes in where populations tend to be mobile, and selected the Horn (Figure 30), with children in rural areas, lagging rural areas, targeted programs to improve lagging border areas, and children from lower- access to education and other critical public income households faring far worse. As a result, services may be needed to facilitate convergence poverty is transmitted from one generation to the with the more developed areas. Better educational 34 Kenya is the exception in the Horn, with better educational outcomes and an HCI of 0.55 (the highest in continental sub- Saharan Africa). 31 SECTION 4 | REGIONAL ECONOMIC INTEGRATION FOR JOB CREATION AND STRUCTURAL TRANSFORMATION outcomes in the lagging areas would also spur this would generate. For low-income countries, labor mobility towards areas with better economic private returns to primary education have been opportunities, facilitating the integration of lagging estimated at 25.4 percent. Returns to complete and leading areas in the HoA. Such a scale-up in primary schooling amount to 32.5 percent in public services, and particularly investing heavily in Djibouti, 32.7 percent in Ethiopia and 17.6 percent education in lagging areas, while costly, would be in Kenya.35 Box 5 examines the associated public more than merited by the expected high returns revenue mobilization challenge for the region. Box 5: Revenue mobilization in the Horn of Africa The level of revenue mobilization varies widely in social and infrastructure spending needs, especially the HoA countries, ranging from 4% of GDP for if these are to be ramped up and made spatially Somalia to 20% for Djibouti as of 2019 (Figure 36). blind to improve development outcomes in the The (unweighted) HoA average of 15.1% is higher lagging areas. Deepening regional integration than that for low income countries in sub-Saharan will support this indirectly through its positive Africa (12.7%), but it has declined over the past five impact on economic growth and the tax base, but years, and is now similar to the global average effective policy reforms and enhanced revenue for low-income developing countries (Figure 37). administration will also be critical to meeting the Mobilizing more revenues will be essential to meet funding challenge. Figure 36: Revenue mobilization is Figure 37: And has decreased since 2016 relatively low (HoA countries’ revenue (excl. grants), % of GDP) (Regional average revenue (ex. grants), % of GDP) 25% 18% 20% 16% 15% 15,1% 14% 14,7% 10% 12,7% 5% 12% 0% 10% Djibouti Ethiopia Kenya Somalia 2015 2016 2017 2018 2019 2015 2016 2017 2018 2019 HoA (excl. Eritrea) LIDC LI-SSA Sources: World Bank staff calculations based on IMF data. Sources: World Bank staff calculations based on IMF data. With respect to revenue policy, a number of tax taxes on tobacco and beer. In the case of Djibouti, expenditures and special regimes in the region are implicit tax expenditure costs amount to 7% of GDP, costly and distortionary, and could be streamlined while in Kenya, rationalization of income tax and to minimize revenue leakage and broaden the tax VAT could potentially yield revenue of up to 3% of bases. Ethiopia could realize additional revenue of GDP. Furthermore, the countries should review 6.4% of GDP through sunsetting tax expenditures and “nuisance” taxes and fees (which can generate not granting new ones, removing VAT exemptions disproportionate administrative costs and weaken on some goods and services, and raising excise tax morale and compliance), and regularly evaluate 35 See Psacharopoulos and Patrinos (2018) on education returns in LICs, and Montenegro and Patrinos (2014) for the country estimates cited. 32 HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM: OVERVIEW the effectiveness of individual tax expenditures in would enable a better understanding of key achieving the intended policy objectives. Annual taxpayers, their needs and behavioral risks, and reporting on tax expenditures would increase improve the targeting of taxpayer service and accountability and transparency. compliance measures. Reforms in revenue administration could also Effective data management is critical for tracking improve collection, particularly in Djibouti goods and services, managing revenue risk, and Somalia, where institutional and technical designing tools to facilitate compliance, and capacity are weak. Business processes should enhancing tax policies. There is a critical need be simplified to reduce the cost of revenue for more and better revenues data, including on administration and taxpayers’ compliance trade-related revenues, which remain opaque in burden. This requires comprehensive review and most of the region. This agenda would benefit reengineering of processes, such as registration from increased automation of revenue systems, of taxpayers, assessment, payment, audit, and provisions for transparency, and more data dispute resolution. For example, in Somalia, the exchanges between countries’ revenue agencies operationalization of the large- and medium- and other government entities which can serve as taxpayer office is important, as such segmentation third-party data sources. For populations in the lagging border areas, relatively good cross-border connectivity has led connectivity to economic opportunity can be to flourishing trade and rapidly growing cities. improved through easier cross-border mobility. Using night-time lights as a proxy for growth, Jijiga, In many cases, border cities are closer to economic Hargeisa, Borama and Berbera grew substantially centers in other countries than they are to their faster than other cities in border areas (Figure 39). domestic centers of economic density. Cities in Coincidentally or not, poverty rates in Jijiga city are border areas have been growing slower due to thick the lowest among all Ethiopian cities.36 An example borders and large distances from their domestic of the other extreme is Assab in Eritrea, the only city economic centers. An exception to this general which experienced a large decrease in economic pattern and an example of the possibilities of activity (as measured by NTL) since the early 1990s improved cross-border connectivity has been the (Figure 39). Despite its strategic importance as a corridor between Jijiga in Ethiopia and Hargeisa in port, border closures with Ethiopia and Djibouti Somalia, in particular the Jijiga – Hargeisa – Borama isolated Assab on all possible fronts with close triangle through the border post of Tog Wajaale to zero market access, leading to a collapse in and onwards to Berbera port (Figure 38), where economic activity. 36 Based on the national poverty line, poverty in Jijiga decreased from 31.6 percent in 2005 to 5.9 percent in 2016. Overall, urban poverty decreased from 35.1 percent to 14.8 percent over the same period (FDRE, 2018). 33 SECTION 4 | REGIONAL ECONOMIC INTEGRATION FOR JOB CREATION AND STRUCTURAL TRANSFORMATION Figure 38: The Jijiga-Hargeisa link is currently Figure 39: And cities along this link have the densest cross-border economic corridor grown rapidly (GDP of cities in the HoA border areas) (Growth in nighttime lights, 1992-2013) Galkayo [SOM] Hargeisa [SOM] Borama [SOM] Semera [ETH] Berbera [SOM] Hart Sheik [ETH] Jijiga [ETH] Kebri Beyah [ETH] Beledweyne [SOM] Wajir [KEN] Mandera [KEN] Asaita [ETH] Moyale [ETH] Assab [ERI] -10% -5% 0% 5% 10% 15% Annual rate of growth in NTL Source: Source: Ghosh et al. (2010). Sources: WB staff calculations based on Li et al. (2020). Growth of towns and secondary cities also Digital infrastructure holds substantial plays an important role in connecting workers promise for growth and jobs, though, except to jobs and economic opportunities. Towns and for Kenya, its development is lagging in the secondary cities are growing fast across the HoA Horn. Poor regulatory quality, slow download and the bulk of urban population growth in coming speeds and relatively high costs result in low decades is projected to take place within these mobile penetration rates, predominantly smaller cities and towns. An urbanization pattern skewed towards the largest cities (Table 3). characterized by growth of towns and smaller The importance of digital connectivity has been cities has tended to be more poverty-reducing highlighted by the ongoing pandemic, with as compared to growth of mega-cities, given the everything from maintaining business operations closer linkages towns have with the surrounding and administrative government services to rural populations and the lower skills requirements providing basic services such as education and of jobs in towns and smaller cities.37 In the sparsely pandemic-response social transfers hinging on populated border areas, towns and secondary access to affordable and reliable connectivity. cities play a particularly important role in providing Improving digital infrastructure in the HoA essential services to generally underserved mobile in an inclusive way will be required to allow populations. Infrastructure and service provision businesses to be competitive in the digital is, however, lagging in towns and smaller cities economy, connect workers to jobs more as compared to large cities and national capitals, effectively, including through the development and substantial investments will be required to of the “gig economy”, and spur financial prepare towns and smaller cities in the HoA for inclusion, especially for vulnerable people in the projected large growth in their populations. remote and underserved areas. 37 Christiaensen, L., De Weerdt, J., & Todo, Y. (2013). Urbanization and poverty reduction: The role of rural diversification and secondary towns. Agricultural Economics, 44(4-5), 435-447. 34 HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM: OVERVIEW Table 3: Key telecom indicators for the HoA Country Fixed Broadband Mobile Number of Mobile Number of Fixed Mobile concentration Penetration % Penetration % Licensed Operators Licensed Operators index HHI* (GSMA) Djibouti 16.3 37.6 1 1 10,000 Eritrea 0 10.6 1 4 10,000 Ethiopia 1.3 46.9 1 1 10,000 Kenya 3.8 107.9 4 + 8 5,135 Somalia 0.6 50.8 8 8 1,886 Africa Average 8.3 84.7 n/a n/a 4,600 Notes: *The Herfindahl-Hirschman Index (HHI) is a measure of market concentration. A market with an HHI of less than 1,500 is considered to be a competitive marketplace, an HHI of 1,500 to 2,500 to be a moderately concentrated marketplace, and an HHI of 2,500 or greater to be a highly concentrated marketplace. + Kenya has also licensed three Mobile Virtual Network Operators (MVNOs) which can resell spectrum to offer more specialized financial services, for instance for the retail and financial services sectors. 4.2 Creating and connecting to markets Poor intra-regional connectivity in the Horn and procedure harmonization are promising, they emerges as one of the main manifestations must be grounded in the development of coherent of weak regional integration as well as an and comprehensive connectivity corridors. In impediment to increased trade within the that regard, targeted improvements in domestic region. Enhancing regional connectivity should connectivity can have substantial impacts on reduce the cost of access to factor and product improving overall regional connectivity. In Eritrea, markets and make connected economic centers a national policy strengthening connectivity in different countries more attractive by providing across the coastline from Assab to Massawa and productivity benefits. Improving connectivity in Asmara and connecting this to the links coming the Horn will require substantial and coordinated in from northern Ethiopia would have large effects investments in the development of physical on both domestic and regional market access, as transport corridors as part of a coherent strategy. would extending connections from Assab into To reap maximum benefits from these large Djibouti. In Somalia, upgrading the road from investments, the corridor investments must Kismayo to the Kenyan border at Liboi would have be accompanied by policy and procedure major effects on connectivity between Nairobi and harmonization and trade facilitation. southern Somalia. Unlocking the gains from the priority corridors and other initiatives to improve In the HoA, four priority corridors have been transport connectivity in the region would benefit put forward as levers of regional integration, from improved coordination between national, but improvements in domestic connectivity bilateral, region-wide, and border city connectivity will be required to maximize the impact of these plans. The need for international coordination is corridors. The four regional priority corridors are particularly critical in the case of connectivity (i) the Kismayo, Lamu and Mogadishu corridor, (ii) to port cities and major airports that serve as the Assab and Djibouti corridor, (iii) the Berbera gateways to the region and the rest of the world. and Djibouti corridor, and (iv) the Mogadishu, Berbera and Bossasso corridor. While these Investments in the four corridors are expected corridors and other envisaged or ongoing policy to have significant effects on structural ideas on infrastructure coordination and policy transformation and real household incomes, 35 SECTION 4 | REGIONAL ECONOMIC INTEGRATION FOR JOB CREATION AND STRUCTURAL TRANSFORMATION but only if accompanied by trade and border of construction, and based on a back-of-the- facilitation. Results from a spatial general envelope calculation to illustrate the strong likely equilibrium model suggest that upgrading the returns). The importance of improved trade and four priority corridors would result in a three border facilitation cannot be emphasized enough: percentage-point increase in the non-agricultural Reducing impediments to trade and thinning employment share and a modest one percent borders can be at least as important to connect annual increase in real household incomes. If the and spur the growth of markets as infrastructure physical corridor investments are accompanied spending that decreases travel time and cost. by better trade and border facilitation, estimated In the price dispersion analysis conducted for income effects are far larger, amounting to a this report, the price effect of crossing a border 4.3 percent annual increase in real household is equivalent to increasing the travel time by incomes over the medium term (Table 4). Under three hours. Substantial improvements in market conservative assumptions, this corresponds to an access can thus be attained though a coordinated internal rate of return on the corridor investments approach on border facilitation to alleviate physical of 15-18 percent (depending on the duration and institutional barriers (Figure 40). Table 4: Percentage change in real income from corridor investments, with and without trade facilitation Scenarios Total Ethiopia Djibouti Somalia Kenya Transport only 1.0 1.3 0.0 1.4 0.7 With trade facilitation 4.3 3.9 5.3 6.3 4.9 Source: Herrera-Dappe and Lebrand (2021). No data for Eritrea. Figure 40: Improved border facilitation would have a large effect on market accessibility (Current market accessibility index) (Market accessibility index when borders are opened) Source: World Bank background paper on HoA connectivity. The “borders opened” scenario models market accessibility if all existing border posts within the HoA would be open and operational. 36 HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM: OVERVIEW In general, the border areas would benefit most widespread, reaching the hinterlands of the from regional corridor investments, and the Horn as well (Figure 42). Corridor investments benefits would be higher and more widespread and improved border facilitation would also if accompanied by trade and border facilitation. lead to increased regional specialization, with Upgrading regional corridors would boost certain regions specializing in agriculture (due to household incomes in the border regions most comparative advantages) and others in non-farm (Figure 41). If combined with trade and border activities; Figure 43 shows the modeled results of facilitation, the benefits in terms of increased improved connectivity and border facilitation on household incomes would be higher and more the share of non-agricultural employment. Figure 41: Corridor investments would boost Figure 42: And would be higher and more incomes in the border areas widespread if combined with border facilitation (Welfare impacts of corridor investments vs. baseline, %) (Welfare impacts of corridor investments and border facilitation vs. baseline, %) Sources: Herrera Dappe and Lebrand, 2021. Sources: Herrera Dappe and Lebrand, 2021. Figure 43: More regional specialization in response Trade facilitation encompasses measures and to improved connectivity investments at the border, and also along trade (Change in the non-agricultural employment share, value chains to reduce trade costs. Whilst there transport and border facilitation) are still important knowledge gaps to fill on the key trade cost raising barriers and challenges, information from this REM, and experience from other regions in Africa, suggest a broad approach to trade facilitation that both reduces trade costs and improves the efficiency of regulation at the border. Every border crossing is unique in terms of the terrain, climate, the types of goods being traded and the challenges that traders and officials face. Nevertheless, effective trade facilitation will typically require: • Simplification and greater transparency of trade policies and border procedures Source: Herrera Dappe and Lebrand, 2021. (accompanied by automation where possible). 37 SECTION 4 | REGIONAL ECONOMIC INTEGRATION FOR JOB CREATION AND STRUCTURAL TRANSFORMATION • Investments in physical infrastructure at the border crossings and promote security. Poor border to improve security and working facilities at official crossings together with lack conditions for officials and traders (e.g., of awareness of relevant rules and regulations by electricity, water, buildings, shelter). traders, harassment by, and weak accountability of, border officials, may push some traders • Capacity-building for officials (especially towards more convenient and less costly informal gender awareness) and traders and traders’ crossings. However, this can put the traders at risk associations. from lack of security and they may be forced to make payments to local militias. Solutions to this • Developing cross-border markets in terms of challenge are undermined by a lack of cohesion facilities and access (including ensuring that and a common understanding between the women traders have equal access). federal government, which controls the official international border checkpoints, and the local • Investing in sanitary, phytosanitary and administrations. veterinary services and facilities to improve the quality of food and livestock trade, increase Increased volumes from trade facilitation returns to farmers and improve control over measures and reductions in corruption tend to pest and disease outbreaks. generate larger revenues from taxation of trade. In addition, greater trade activity leads to more • Addressing the challenge of access to foreign expenditures in the local economy on services exchange in certain countries. such as food and hotels, which in turn contributes to general revenue collection. If some of these • Improving access to, and the quality of, trade- revenues remain in the border areas, or are allocated related transport and logistics to reduce costs by the center to investments in those areas, then and damage and wastage. experience from other countries suggests this can support a gradual escape from a low level These reforms and investments will be most equilibrium of corruption, distrust of officials and effective when coordinated on both sides of the small-scale activities, to a more virtuous circle in border. This can be achieved through dialogue which improvements in policy and facilities further at the national level on policies, procedures, and contributes to higher trade volumes and revenues physical investments and also at the local level and broader private sector activity. on implementation through, for example, joint border committees. Monitoring of progress to Deepening trade integration in the Horn ensure the effective implementation of reforms of Africa will pave the way for the effective will require investments in data collection and implementation of the Africa Continental Free reporting mechanisms. For example, the ability of Trade Area and the significant economic benefits traders to report instances of abuse is an essential it is expected to bring. A recent study by the element in bringing accountability to the actions World Bank suggests that implementation of the of officials and to reduce harassment, physical African Continental Free Trade Area would provide violence, and corruption at the border. a substantial boost to trade in Africa and generate significant economic benefits.38 For Ethiopia these Better infrastructure and coordinated trade could amount to an income gain of around 9 policy reforms that reduce trade costs for small- percent, and for Kenya over 11 percent, among the scale traders will encourage the use of formal highest in Africa, reflecting high current barriers to 38 World Bank (2020) The African Continental Free Trade Area: Economic and Distributional Effects, World Bank, Washington 38 HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM: OVERVIEW trade. These gains come from the removal of tariffs Next to investments in domestic and regional and reduction of non-tariff barriers on intra-African connectivity in the Horn to connect to markets, trade. But the biggest impact comes from trade local investments are needed to enhance facilitation measures, accounting for around three- connectivity within metropolitan regions. Addis quarters of the overall gains for both countries. The Ababa and Nairobi already are some of the major large boost in household consumption expected metropolitan regions of Africa, encompassing to result from greater trade openness could lift millions of people. Other economic centers are as many 8.2 million people in Ethiopia and 4.4 also growing in importance. The emphasis on million in Kenya out of moderate poverty. Although connectivity provision for border cities and port data limitations prevented detailed modelling cities should not preclude improving connectivity for Eritrea and Somalia, similar, if not larger around the biggest cities’ metropolitan regions. impacts, could be expected given their higher This additional focus is justified because the trade restrictiveness. Hence, this reinforces the virtuous effects of air transport connectivity importance of improvements in trade facilitation and rail corridors between major cities tend and the removal of tariff and non-tariff barriers to spill over to proximate cities. When multiple among the Horn of Africa countries and that this cities form a dense regional network, a larger would generate additional benefits beyond those number of producers and consumers can be directly related to greater trade within the region, connected. The infrastructure built by public by better connecting the countries to partners authorities in these urban regions opens markets elsewhere in Africa under the broader objective and opportunities in proximate towns through of African integration. the capitals’ gateway effects. 39 SECTION 5 | KEY ACTIONS FOR JOBS AND ECONOMIC TRANSFORMATION IN THE HOA SECTION 5 Key actions for jobs and economic transformation in the HoA Raising living standards in the Horn through the key actions based on the analysis presented more and better jobs will require a menu of in this overview, structured around the 2009 actions and reforms. Some of these actions will, WDR framework of “Institutions, Infrastructure, due to their very nature, be more domestically and Interventions”. Institutions refer to policies oriented or less amenable to a regional that should be spatially blind and universal in approach (for instance, improving education coverage (for instance, the provision of basic and strengthening skills) while others will require public services). Infrastructure refers to polices deliberate and sustained regional coordination and investments that are spatially connective (such as coordinated investments and reforms to (roads, railways, etc.). Finally, interventions are facilitate cross-border trade). This section lays out programs that are explicitly spatially targeted. 5.1 Institutions: Building human capital The countries in the Horn are characterized by national policy priority. Based on the latest sparsely populated and peripherally located household surveys, a substantial share of youth lagging areas. People in those areas are on average (15-24) in the HoA do not complete primary poorer and suffer from multiple overlapping education (with large variations across countries deprivations in public service delivery. The policy and regions within countries). Consistent with the approach to integrate these lagging areas with emphasis placed by the JET agenda on building the leading areas should focus on institutions.39 capabilities, it is difficult to see how broad-based Examples of institutions include property rights, improvements in productivity, earnings, and living land and labor regulations, macroeconomic standards can be achieved with these low levels stability, and the provision of essential social of education. Improving human capital will require services such as health, education, and water a strong effort in the Horn’s borderlands and and sanitation. Institutions should be provided lagging rural areas, where human development and applied regardless of place. indicators and service delivery are poor. Improving girls’ education will also be particularly important The spatially blind provision of social services, given its effect on bringing down fertility rates and especially education, is a foundational and laying the foundations for a growth-boosting priority for countries in the Horn and a major demographic dividend. 39 See World Development Report 2009. 40 HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM: OVERVIEW 5.2 Infrastructure: Trade facilitation and harmonization and selective investments in physical infrastructure There are substantial projected gains from of infrastructure, investments in cross-border deeper integration and more trade in the Horn, in connectivity need to be selective and rigorously terms of higher incomes, increased food security analyzed within a cost-benefit framework. and greater stability. This will require policies and investments that are spatially connective: The Horn’s border areas in particular would Improving regional transport connectivity, benefit greatly from improved coordination and including through addressing missing links in harmonization of trade policies and improved domestic transport networks, and improving trade border infrastructure. The Horn’s borderlands are and border facilitation. While physical investments lagging areas, the result of a long and complex in cross-border corridors and infrastructure are history of fragility and marginalization, and are projected to have positive effects on household characterized by low economic density and high welfare, especially in the border areas, the economic distance to the domestic economic effects will be far greater and more widespread centers. Often, the nearest economic center for if the physical investments are combined with borderland populations is located at the other improved trade and border facilitation. Reducing side of the border, but thick borders and opaque impediments to trade and thinning borders can trade policies limit productive interactions with be at least as important to connect and spur the these centers. Facilitating cross-border trade and growth of markets as infrastructure spending that mobility is the least-cost way to increase market decreases travel time and cost, especially in the access for the populations in the border areas, HoA’s border areas. Investments in connective with expected positive impacts on welfare and infrastructure (both between and within countries), economic activity. In the few places in the HoA trade policy reforms and trade facilitation will need where borders are operational and relatively fluid, to be assessed from a regional perspective and and where some key infrastructure has been coordinated with neighboring countries in order provided, trade has flourished, and vibrant cross- to reap the biggest returns. Given the high cost border corridors of density have emerged. 5.3 Interventions: Metropolitan regions and selected secondary cities The ongoing rapid urbanization in the Horn can However, infrastructure and service provision in be an engine for structural transformation and smaller cities are lagging compared to capitals, better jobs, but requires targeted investments warranting substantial investments to prepare in secondary cities and around metropolitan for their projected growth. This is particularly areas. The Horn’s urban population has grown important for secondary cities in the border areas, from 20 million in 2000 to 50 million in 2020 which play a crucial role in service delivery to, and and is expected to increase to 90 million by 2035. market access for, mobile populations in sparsely This rapid urbanization will increase demand populated lowlands. for food grown in rural areas of the region, potentially benefiting rural producers. Most of In addition to targeted investments in the increase in urban population is projected to secondary cities, local investments are happen in secondary cities and towns, which needed to enhance connectivity within the usually have stronger links with their surrounding big metropolitan regions in the Horn. Addis rural hinterland and are more poverty-reducing. Ababa and Nairobi already are some of the major 41 SECTION 5 | KEY ACTIONS FOR JOBS AND ECONOMIC TRANSFORMATION IN THE HOA metropolitan regions of Africa, encompassing metropolitan regions allows nearby places millions of people. Improving connectivity and and people to benefit from agglomeration tackling congestion around the capital cities’ economies, unleashing network spillover effects. 5.4 Additional considerations The analysis conducted for the HoA REM points to three additional policy considerations and priorities. First, the analysis has shown that there is a strong help focus development partners’ support and case for a concerted and coordinated effort to maximize its overall impact. improve living standards and reduce fragility in the Horn, including by applying regional, cross- Second, job generation in the formal private border strategies. While spread across different sector in the Horn, while essential, will only countries, the border areas of the HoA face similar deliver a fraction of the job opportunities needed challenges: poor service delivery and human to accommodate the booming labor force. While development outcomes, lagging infrastructure, jobs policy must take account of the differences high economic distance from domestic density in economic structure between the countries and international density (thick borders), of the region, workers in the Horn would likely vulnerability to climate change, and fragility. The benefit most from policies for broad-based, cross- returns to investments and policy reforms in the sectoral productivity growth (notably including in border areas will be amplified by coordinated, agriculture), centered around maintaining macro- subregional efforts. For instance, cross-border fiscal stability, mobilizing more public revenues models of education and health service delivery and deploying these towards greater and more could be considered, taking advantage of target efficient spending on public services (notably populations having greater scale when people education) and infrastructure, as well as improving living on either side of a border are included the investment climate and access to finance. (e.g., to realize scale economies in the provision of home-language education), and could be Third, there is an urgent need for more, better particularly cost-effective where populations are quality and more timely economic data for the isolated domestically but closer to service delivery region (including using more recent technology facilities across a border. Similarly, investments in and methods, such as remote sensing and big better infrastructure and connectivity in a border data such as mobile phone data) to support area of one country will only have limited effects analysis and sound policies. These tools are if the border remains thick or infrastructure on likely to be of particular use in improving our the other side remains poor. Coordinated disaster understanding of the issues at the main border preparedness and response effects could help crossings and along the key cross-border value protect the poor and vulnerable, especially in chains, as well as those arising in secondary cities the face of climate change. Across all these – critical dimensions for driving development in dimensions, a regional perspective would also the borderlands of the HoA. 42 HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM: OVERVIEW Annex Map: The Horn of Africa and its main cities Karima Haiya Abha SAUDI ARABIA Atbara Re Najran d S U D A N Jizan Sadah Se JIZAN ERITREA a Keren R E P. O F MITSIWA HARBOR Khartoum Kassala Massawa SALIF Sana’a YEMEN Asmara RAS ISA MARINE TERMINAL Al Mukalla ALAl Hudaydah AHMADI Dhamar 'Ataq ASH SHIHR OIL TERMINAL Wad Madani AL MUKALLA Adigrat Al Bayda Gedaref Ibb RUDUM TERMINAL BALHAF Sennar Mekele AL MUKHA Taizz Kosti ASSAB Assab Lahij den of A Aden ADEN Gonder G u l f DJ IBOU TI Tadjoura BOOSAASO Bahir Dar Djibouti DORALEH Ali Sabieh Boosaaso Dese Dikhil BERBERA Melut Berbera Asosa E T H I O P I A Dire Dawa Malakal Burco Qardho Nek'emte Addis Harar Babile Hargeysa Ababa Nazret Laascaanood Gore Fiq Garoowe Jima Ginir Awasa Goba SOMALIA SOUTH Imi Gode SUDAN Arba Minch Dhuusa Mareeb Negele Ferfer Juba Kapoeta Filtu Beledweyne Dolo Bay Rhamu Luuq Moyale Mandera Lodwar Baydhabo Gulu El Wak Jawhar Lira Marsabit MUQDISHO UGANDA K E N Y A Wajir Mogadishu Marka MARKA Mbale Kitale Modogashe Samatar Bu'aale BARAAWE Kampala Eldoret Kulamawe Habaswein Liboi Bilis Qooqaani Isiolo Jinja Kakamega Meru Kisumu Masaka Nakuru Nyeri Garissa KISMAAYO Kismaayo Embu Mwingi Bukoba Musoma Nairobi INDIAN Mwanza Lamu OCEAN Namanga LAMU Arusha Voi MALINDI Shinyanga Moshi KILIFI Babati Mombasa Nzega MOMBASA Singida 0 100 200 Miles Tabora TANGA Tanga T A N Z A N I A CHAKE CHAKE 0 100 200 Kilometers ZANZIBAR IBRD 46041 | Dodoma Zanzibar JUNE 2021 This map was produced by the Morogoro Dar DAR es Salaam ES SALAAM Cartography Unit of the World HIGHWAYS Bank Group. The boundaries, colors, denominations and any Mikumi other information shown on this MAJOR ROADS Iringa map do not imply, on the part of Kibiti the World Bank Group, any judgment on the legal status of any NATIONAL CAPITALS territory, or any endorsement or acceptance of such boundaries. INTERNATIONAL BOUNDARIES KILWA KIVINJE 43 REFERENCES References This overview draws on the following Horn of Africa Regional Economic Memorandum Background Papers (2021): 1. Economic Geography Analysis (Tom Bundervoet and Takaaki Masaki) 2. Jobs in the Horn of Africa (Tom Farole, Jan von der Goltz and Tove Sahr) 3. A Review of Cross-Border Trade in the Horn of Africa (Paul Brenton and Habtamu Edjigu) 4. Infrastructure and Structural Change (Matias Herrera Dappe and Mathilde Lebrand) 5. A Framework for Enhancing Intra-Regional Connectivity in the horn of Africa (Charles Kunaka and Ben Derudder) 6. Overview of Digital Development in the Horn of Africa (Tim Kelly and Eric Dunand) Barros, R.P., Ferreira, F., Molinas Vega, J. & IMF Annual Report on Exchange Arrangements Saavedra, J. (2009). Measuring Inequality and Exchange Restrictions (2019). of Opportunities in Latin America and the Li, X., Zhou, Y., Zhao, M. and Zhao, X. (2020). A Caribbean.” Washington DC: The World Bank. harmonized global nighttime light dataset Brenton, P. and M. Nyawo. (2019). Food Prices, 1992-2018. Nature Scientific Data 7(168). Access to Markets and Child Undernutrition Montenegro, C. and Patrinos, H. (2014). in Ethiopia. Policy Research Working Paper Comparable Estimates of Returns to 8823, World Bank, Washington, DC. Schooling Around the World. Policy Research Bloom, D., Canning. D & Sevilla, J. (2003). The Working Paper 7020. Washington DC: The Demographic Dividend. A New Perspective on World Bank. the Economic consequences of Population Psacharopoulos, George; Patrinos, Harry Change. RAND Corporation Anthony. (2018). Returns to Investment in Federal Democratic Republic of Ethiopia. (2018). Education: A Decennial Review of the Global Poverty and Economic Growth in Ethiopia Literature. Policy Research Working Paper No. (1995/96 – 2015/16). Ethiopia: Planning and 8402. Washington, DC: World Bank. Development Commission. Roberts, Peter; KC, Shyam; Rastogi, Cordula. Ghosh, T., Powell, R. L., Elvidge, C. D., Baugh, K. 2006. Rural Access Index: A Key Development E., Sutton, P. C., and Anderson, S. 2010. The Indicator. Transport paper series; no. TP-10. Open Geography Journal 3: 147-160.Sources: World Bank, Washington, DC. HYDE3.2. Ronald Lee and Andrew Mason (2006) “What Gribble, J. & Bremner, J. (2012). “Achieving a is the Demographic Dividend?” Finance and Demographic Dividend.” Population Bulletin Development (September) 16-17. 67 (2), The Population Reference Bureau. Schmidt, E., Dorosh, P., Jemal, M. and Smart, J. 2018. IDA. (2019). Special Theme: Jobs and Economic Ethiopia’s Spatial and Structural Transformation. Transformation. Washington DC: The ESSP Working Paper 119. International Food International Development Association. Policy Research Institute (IFPRI). World 44 HORN OF AFRICA REGIONAL ECONOMIC MEMORANDUM: OVERVIEW Urbanization Prospects (UN 2018). 2009: “Reshaping Economic Geography”. Washington DC: The World Bank Tesfaye, A. and N. Amaha. (2018). A Review on Cross-Border Livestock Trade Across World Bank Group; U.K. Department of Dry Land Borders of Ethiopia: The Trends International Development. 2019. Poverty and Implications’, Journal of Scientific and and Vulnerability in the Ethiopian Lowlands: Innovative Research 7(2): 36-42 Building a More Resilient Future. World Bank, Washington, DC: World Bank. United Nations. (2019). Department of Economic and Social Affairs, Population Division. World World Bank. 2019. Transport note for the high- Population Prospects 2019, Online Edition. level conference on promoting economic Rev. 1. integration in the Horn of Africa, February World Bank (2009), World Development Report 2019. Unpublished. 45 46