Inclusive economic growth in Kenya: The spatial dynamics of poverty

Tyson, Judith; Diwakar, Vidya; Adetutu, Morakinyo O.; Bishop, James Research Report Inclusive economic growth in Kenya: The spatial dynamics of poverty ODI Report Provided in Cooperation with: ODI Global, London Suggested Citation: Tyson, Judith; Diwakar, Vidya; Adetutu, Morakinyo O.; Bishop, James (2020) : Inclusive economic growth in Kenya: The spatial dynamics of poverty, ODI Report, Overseas Development Institute (ODI), London This Version is available at: https://hdl.handle.net/10419/233929 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc-nd/4.0/ Inclusive economic growth in Kenya The spatial dynamics of poverty Prepared for the FSD, AfRO and OSIEA ‘Modelling an inclusive economy for Kenya’ programme Judith Tyson, Vidya Diwakar, Morakinyo Adetutu and James Bishop July 2020 Report Readers are encouraged to reproduce material for their own publications, as long as they are not being sold commercially. ODI requests due acknowledgement and a copy of the publication. For online use, we ask readers to link to the original resource on the ODI website. The views presented in this paper are those of the author(s) and do not necessarily represent the views of ODI or our partners. This work is licensed under CC BY-NC-ND 4.0. Cover photo: Fruit market in Lamu, Kenya, 2019. © Beks/Unsplash. 3 Acknowledgements The Overseas Development Institute (ODI) gratefully acknowledges the support of FSD Kenya in authoring and conducting fieldwork for this report. We thank FSD Kenya and Dirk Willem te Velde, as well as the many interviewees in Kwale and Busia who generously gave their time for this paper. The views presented are those of the authors alone and do not necessarily represent the views of ODI, FSD Kenya or any of the interviewees or peer reviewers. 4 Contents Acknowledgements 3 List of tables and figures 5 Acronyms 6 1 Introduction 7 2 Why has Kenyan growth not been more inclusive? 8 2.1 Introduction 8 2.2 The SIDA poverty framework inKenya 10 2.3 Conclusion 16 3 What are the implications for policy andinterventions? 22 3.1 The importance of SIDA ‘context’ 22 3.2 The need for context-led policy and interventions 23 3.3 Addressing extreme poverty 25 3.4 Conclusion 25 4 Conclusion 27 References 29 Annex 1 SIDA index methodology 34 Annex 2 The spatial productivity analysis 36 A2.1 Estimations using county-level panel data from 2005 to 2015 36 Annex 3 Fieldwork case studies 38 A3.1 Kwale 38 A3.2 Busia 44 5 List of tables and figures Tables Table 1 Growth hubs – county index scores 19 Table 2 Poverty logjams – county index scores 20 Table 3 The ‘mid-case’ counties – county index scores 21 Table 4 How the SIDA framework affects ‘pathways out of poverty’ 24 Table A1 SIDA index rankings and weightings 35 Table A2 Correlations of county productivity, regression results 37 Figures Figure 1 SIDA’s multidimensional poverty framework 9 Figure 2 Spatial map of Kenyan poverty by county (2015–2016) 9 Figure 3 Kenyan poverty and extreme poverty by county (2015–2016) 10 Figure 4 Economic opportunity by county 11 Figure 5 Kenyan environmental vulnerability bycounty 13 Figure 6 Kenyan political and institutional strength by county 13 Figure 7 Kenyan conflict and security by county 16 Figure 8 SIDA framework index and poverty levels by county 17 Figure 9 SIDA framework economic and social index and poverty levels by county 18 Figure 10 SIDA framework environmental, political and security index and poverty levels by county 18 Figure 11 Growth hubs – stylised facts 19 Figure 12 Poverty logjams – stylised facts 20 Figure 13 The ‘mid-case’ counties – stylised facts 21 Figure A1 Change in productivity estimates 37 6 Acronyms BMU beach management unit CDP County Development Plan CIDP County Integrated Development Plan CPAN Chronic Poverty Advisory Network FSD Financial Sector Deepening GCP gross county product GDP gross domestic product HISP Health Insurance Subsidy for the Poor ICT information and communications technology ILO International Labour Organization KIHBS Kenya Integrated Household Budget Survey KNBS Kenya National Bureau of Statistics LREB Lake Region Economic Bloc MSME micro, small and medium enterprise NAP National Adaptation Plan NGO non-governmental organisation NHIF National Hospital Insurance Fund NSSF National Social Security Fund ODI Overseas Development Institute OECD Organisation for Economic Co-operation and Development OLS ordinary least squares OSIEA Open Society Initiative for East Africa SACCO savings and credit cooperative SET Supporting Economic Transformation SEZ Special Economic Zone SIDA Swedish International Development Cooperation Agency SME small and medium-sized enterprise UNDP United Nations Development Programme UNICEF United Nations Children’s Fund WEF World Economic Forum 7 1 Introduction Kenya poses a conundrum. Its top-line economic growth has been impressive and it has been one of the fastest-growing economies in sub-Saharan Africa over the past decade. However, poverty has not fallen as fast as it should have in light of this strong economic growth. Extreme and urban poverty have remained stubbornly high. This paper investigates the discrepancy. A key issue is that the composition of economic growth has been unbalanced. Strong growth in financial services and information and communications technology (ICT) has created jobs, but just for the educated middle classes. The mass migration of unskilled workers from rural to urban areas has only been absorbed into low-wage, informal labour. Manufacturing has grown but not sufficiently and agriculture, on which most of the poor still depend, is largely a subsistence activity. It is necessary to dig deeper than the composition of economic growth, however. We do this using the Swedish International Development Cooperation Agency (SIDA) framework, which examines poverty dynamics in terms of four dimensions: (i) economic and social, (ii) political and institutional, (iii)environmental and (iv) security and conflict. As we discuss in Chapter 2, these factors vary significantly between the different regions in Kenya and those variations can either accelerate prosperity or entrench poverty. Thus, our analysis gives rise to a critical question: what can be done to make economic growth more inclusive? In Chapter 3, our analysis suggests it is critical to assess the context of economic development in relation to policy and interventions to make sure it is ‘good enough’ and, where it is not, to rely more on social protection and other direct interventions to alleviate poverty. Our research also shows that where the context is not ‘good enough’ – particularly in addressing extreme poverty – cash transfer programmes can lead to short-term welfare improvements. Longer term, however, they may trap the poor in typical low-income, informal work, so they need to be combined with proactive programmes to help the poor develop higher-income livelihoods. Lastly, there is a need for a forward-looking view of the context. In Chapter 4, we discuss the potentially overwhelming threats of population growth, climate change, environment and noninclusive politics and corruption. Tackling these must also be a central part of making economic growth more inclusive. 8 2 Why has Kenyan growth not been more inclusive? 1 The World Bank confirmed middle-income graduation for Kenya in 2015. The KNBS confirmed this in late 2014 after a review of methodology for assessing the economy’s telecommunications and property sectors (KNBS, 2016). 2 The Kenya National Bureau of Statistics (KNBS) defines households and individuals with a total monthly adult equivalent consumption expenditure per person of less than KSH 3,252 in rural and peri-urban areas and less than KSH 5,995 in core urban areas as living in ‘overall poverty’. Households and individuals with a total monthly adult equivalent consumption expenditure per person of less than KSH 1,954 in rural and peri-urban areas and less than KSH 2,551 in core urban areas are deemed to live in ‘hardcore or extreme poverty’. 2.1 Introduction Kenya has been one of the fastest-growing economies in sub-Saharan Africa in the decade to 2020, with gross domestic product (GDP) growth averaging 5.5% since 2011 and graduation to middle-income status in 2014 (KNBS, 2016).1 The buoyant economy has been underpinned by political devolution following the adoption of a new constitution in 2010 and an economic strategy that set the country’s ‘big four’ priority areas of manufacturing, universal healthcare, affordable housing and food security. Infrastructure and the business environment continue to improve, supporting private investment, but political patronage and corruption remain pervasive (IMF, 2018; Wankuru et al., 2019). Kenya’s strong GDP growth has been accompanied by impressive increases in per capita income of 4.6% annually and a decline in the poverty rate from 46.6% to 36.1% between 2005–2006 and 2015–2016. Extreme poverty fell from 19.6% to 8.6% in the same period, according to the Kenya Integrated Household Budget Survey (KIHBS) 2015–2016 (KNBS, 2016).2 However, the overwhelming majority of Kenyans, an estimated 80%, remain poor or near poor, so are vulnerable to falling back into poverty. The benefits of economic growth have been inequitable, with the richest 25% of the population consuming 60% of the increase in GDP (UNDP, 2017; Diwakar and Shepherd, 2018; Wankuru et al., 2019). What’s more, the absolute number of poor people has remained broadly unchanged, as Kenya’s population has more than doubled since1990 and many of the growing number of young people have failed to find decent employment (Diwakar and Shepherd, 2018; Wankuru et al., 2019). So why, given this strong macroeconomic growth, has poverty alleviation not been faster and more widespread? 2.1.1 The SIDA framework This paper examines that question in terms of the four dimensions set out in SIDA’s multidimensional poverty framework: (i)economic and social context, (ii)environmental risk, (iii) politics and institutions and (iv) conflict and security (Figure1) (SIDA, 2017). 15 to maintain key welfare benefits (Diwakar and Shepherd, 2018; Wankuru et al., 2019). Unfortunately, the poor are generally less likely to be engaged in employee associations if they are not formally employed. Even where they are members of such associations, they appear less able than the non-poor to translate this engagement into tangible benefits, such as the receipt of National Social Security Fund (NSSF) or National Hospital Insurance Fund (NHIF) benefits from their employer. Gender is a pervasive dimension of poverty Kenya ranked 76th out of 144 countries in the Global Gender Gap Report 2017 due to gender inequalities in school achievement, health outcomes, political representation and labourforce participation (WEF, 2017). Even though girls’ access to education has improved over time, gender equality in other areas, such as political participation, remains weak (Ponge, 2013). This gender inequality affects the well-being of women and their wider households, as there is a strong positive correlation between gender inequality and poverty rates. Reducing gender inequality will require short-term efforts, such as affirmative and legal action, as well as embedded longer-term processes to change norms. Satisfactory grievance redress is related to community strength, not poverty Grievance redress mechanisms play an important role in a functioning political system, allowing people’s voices to be heard and providing a forum for injustices to be addressed. These may be formal or informal institutions and communities that band together over a common cause. There are multiple social movements, community organisations and community-led resolution processes, such as hearings by elders or community chiefs, as well as formal courts (Diwakar and Shepherd, 2018). Furthermore, the strength of grievance redress is not strongly linked to poverty, with counties such as those in the north-east having the most satisfactory outcomes, as reported by households. This suggests 17 Based on data in the index. See Annex 1 for more details. that community coherence rather than wealth may drive positive grievance resolution.17 When it comes to formal grievance processes, poor households and women are the least likely to have political representation or to seek or receive satisfactory resolution. Such disadvantages are driven both by a lack of income, limiting access to legal representation in the formal courts, and ethnic and regional inequalities (Diwakar and Shepherd, 2018). This is particularly unfortunate, as most of the grievances expressed by households in the KIHBS related to resources (such as personal property, divorce, land, natural resources), suggesting a vicious circle, whereby resource-poor households had no means of settling resource-related disputes and remained structurally embedded in resource poverty as a result. 2.2.4 Conflict and security Kenya has a high level of conflict and insecurity and this has a strong spatial dimension, with severe problems in the north-eastern counties and urban areas, including Nairobi, Kisumu and Mombasa (Figure 7). Conflict and insecurity push the poor deeper intopoverty The poor are more exposed to such problems and more vulnerable to their effects, because even micro-level household shocks, such as theft, can depress household well-being and driveimpoverishment. For example, the most common response to shocks is to draw down savings or make a distressed sale of assets. However, this can be problematic for the poor, who lack savings, leading to dysfunctional coping mechanisms, such as withdrawing children from school, resorting to child labour or reducing nutrition, leading to food insecurity and poor health. Conflict and insecurity are most pervasive in the north-eastern counties Such problems are more pervasive in remote areas, such as the north-east, and also in relation to terrorist organisations, such as alShabab 16 insurgents. This has been ineffectively controlled due to inadequate policing and state security arrangements (Pkalya et al., 2003). Such conflict has fuelled internal displacement in the north-east. Counties with high levels of displaced people are among the poorest in Kenya and suffer from other multidimensional factors, such as low school enrolment rates, malnutrition and inadequate water and sanitation (Diwakar and Shepherd, 2018). Such problems are compounded by the environmental context (with climate hazards, such as drought, exacerbating farmer–pastoralist conflicts) and the political context (sparking election-related violence, as in 2007, and a rise insexual and gender-based violence) (Thomas etal., 2013). 18 This section has also been informed by case-study fieldwork in Busia and Kwale (see Annex 3 for more detail). 19 See Annex 1 for a full description of the index methodology, data sources and sub-components. 2.3 Conclusion The four SIDA framework dimensions have given us an understanding of the various dimensions ofpoverty.18 Equally important is that they interact with each other to become ‘more than the sum of their parts’, meaning that poverty dynamics are not only multidimensional, but interrelated and act to either compound poverty traps or to provide permanent pathways out of poverty. These multidimensional and interrelated poverty dynamics can be seen at county level in Kenya. Figure 8 shows the percentage of the Kenyan population living in poverty in 2015–2016, with a score on the SIDA framework index for eachcounty.19 The index gives a score for each of the four dimensions, with a high and positive score indicating a supportive context for permanent pathways out of poverty and a low or negative score indicating a context that compounds poverty traps. So, for example, a county with strong economic and social opportunities, an environment with low levels of climate risk or environmental degradation, sound politics and institutions, and low levels of conflict and insecurity would have the highest score. Conversely, a county with poor economic and social opportunities, an environment with high levels of climate risk and environmental degradation, weak politics and institutions, and high levels of conflict and insecurity would have the lowest score. As can be seen, there is a close relationship between high scores on the SIDA framework index and low levels of poverty: counties with a higher positive score (supportive context) have lower poverty levels, while those with low and negative scores (compounding context) have higher poverty levels (Figure 8). It is also notable that the economic and social context has the strongest inverse relationship to poverty (Figure 9), whereas the other three SIDA framework dimensions have a more variable relationship to poverty (Figure 10). Figure 7 Kenyan conflict and security by county Note: (i) Rankings from 1 to 5 indicate quintiles, whereby 5is the upper quintile with the highest levels of security andlowest levels of conflict and 1 is the bottom quintile with the lowest levels of security and highest levels of conflict. (ii) Individual components of the conflict and stability index include conflict fatalities, the crime index, food insecurity and households experiencing shocks recorded by the KIHBS 2015–2016. See Annex 1 for further details of index methodology. Source: Authors’ analysis, based on KNBS (2016) 0–1 1–2 2–3 3–4 4–5 17 Growth hubs Moreover, those counties with the highest scores and least poverty tend to have supportive contexts on all four SIDA dimensions: positive economic and social contexts, stronger political and institutional environments and fewer problems with the environment, security and conflict (though some have notably poor scores on the environmental front). These interact to compound their positive effects on poverty and make them ‘more than the sum of their parts’. These advantages work together to create ‘growth hubs’ with strong economic potential that is realised, resulting in relatively low levels of poverty and low or negligible extreme poverty. They include urban centres,20 such as Nairobi, Mombasa and Kisumu, and rural counties where commercial and high-value agriculture with processing and exports has developed strongly, such as Nakuru (horticulture and vegetables) and Kericho (tea) (Figure 11 and Table 1). 20 Urban poverty saw a small absolute fall from 32.1% to 29.4%, from 2005–2006 to 2015–2016, due to mass migration. Poverty logjams At the other end of the spectrum are counties where the four dimensions compound each other to create poverty logjams that make a permanent escape from poverty difficult and often create high levels of extreme poverty. This is most common in counties in remote and marginal areas, including the north and east of Kenya (Figure 12 and Table 2). These counties typically suffer from weaknesses in all four dimensions. They are often (increasingly) arid, sparsely populated and highly reliant on subsistence agriculture, such as livestock. There tends to be a lack of services and infrastructure to support economic development. Communities are frequently marginalised, not only geographically, but also ethnically and politically, reducing their effectiveness in influencing policy and public spending in their favour. Furthermore, levels of crime and conflict are high, with ever more disputes over land usage and clashes with insurgents along the poorly guarded border withSomalia. Figure 8 SIDA framework index and poverty levels by county Note: See Annex 1 for further detail of index methodology and data. Source: Authors’ calculations based on KNBS data –1.5 –1.2 –0.9 –0.6 –0.3 0.0 0.3 0.6 0.9 1.2 1.5 –1.5 –1.2 –0.9 –0.6 –0.3 0.0 0.3 0.6 0.9 1.2 1.5 Poverty level SIDA framework index Proportion of population (%) Index score Garissa Turkana West Pokot Busia Marsabit Wajir Mandera Samburu Tana River Kisii Kitui Laikipia Kilifi Isiolo Kwale Vihiga Elgeyo Marakwet Migori Bomet Kakamega Makueni Bungoma Baringo Uasin Gishu Homa Bay Kajiado Trans Nzoia Nandi Nyandarua Nyamira Siaya Kisumu Lamu Kericho Embu Nakuru Taita Taveta Tharaka Nithi Narok Nairobi Meru Machakos Mombasa Muranga Kiambu Kirinyaga Nyeri 18 Figure 9 SIDA framework economic and social index and poverty levels by county Source: Authors’ calculations based on KNBS data 0 1 2 Rankings 3 4 5 6 EconomicPoverty Garissa Turkana West Pokot Busia Marsabit Wajir Mandera Samburu Tana River Kisii Kitui Laikipia Kilifi Isiolo Kwale Vihiga Elgeyo Marakwet Migori Bomet Kakamega Makueni Bungoma Baringo Uasin Gishu Homa Bay Kajiado Trans Nzoia Nandi Nyandarua Nyamira Siaya Kisumu Lamu Kericho Embu Nakuru Taita Taveta Tharaka Nithi Narok Nairobi Meru Machakos Mombasa Muranga Kiambu Kirinyaga Nyeri Figure 10 SIDA framework environmental, political and security index and poverty levels by county Source: Authors’ calculations based on KNBS data 0 1 2 Rankings 3 4 5 6 Garissa Turkana West Pokot Busia Marsabit Wajir Mandera Samburu Tana River Kisii Kitui Laikipia Kilifi Isiolo Kwale Vihiga Elgeyo Marakwet Migori Bomet Kakamega Makueni Bungoma Baringo Uasin Gishu Homa Bay Kajiado Trans Nzoia Nandi Nyandarua Nyamira Siaya Kisumu Lamu Kericho Embu Nakuru Taita Taveta Tharaka Nithi Narok Nairobi Meru Machakos Mombasa Muranga Kiambu Kirinyaga Nyeri SecurityPoliticalEnvironment 19 Figure 11 Growth hubs – stylised facts • Growth hubs and corridors • Higher-productivity sectors: agricultural processing, commercial agriculture and formal manufacturing, incl. in special economic zones • Infrastructure, including higher electrification rates and trade corridors (roads, railways, airports and seaports, including those with cargo facilities) • Urban areas lack adequate transport, water and sewage • Lower poverty rates, but pockets of poverty remain, incl. in urban slums • Degraded urban environment • Agricultural fecundity • Natural water sources and/or irrigation • Reasonable security and absence of pervasive conflict • Targets for terrorism • Dominant communities • Central areas • Influence with national government • Poor lack access and influence in these networks • Corruption Economic and social Political and institutional Environmental Conflict and security County Poverty levels (national quintile) SIDA dimension Economic andsocial Environmental Political and institutional Security and conflict All dimensions Nyeri 1.0 3.9 3.0 4.2 2.8 3.5 Kirinyaga 1.0 3.7 3.0 3.6 3.5 3.4 Kiambu 1.0 3.3 3.0 4.0 3.8 3.5 Muranga 1.0 3.3 4.0 3.4 3.0 3.4 Mombasa 1.0 3.9 1.0 3.0 3.3 2.8 Machakos 1.0 3.6 1.0 4.0 3.0 2.9 Meru 1.0 3.4 4.0 2.6 3.3 3.3 Nairobi 1.0 3.2 3.0 2.6 4.3 3.3 Narok 1.0 3.2 5.0 2.2 2.5 3.2 Tharaka Nithi 1.0 3.0 3.0 2.4 3.5 3.0 Taita Taveta 2.0 3.8 5.0 3.8 2.0 3.6 Nakuru 2.0 3.7 4.0 2.8 3.8 3.6 Embu 2.0 3.4 3.0 3.4 3.5 3.3 Kericho 2.0 3.3 2.0 3.2 3.5 3.0 Lamu 2.0 3.4 4.0 1.6 2.3 2.8 Kisumu 2.0 3.6 1.0 2.0 2.8 2.3 Average 1.4 3.5 3.1 3.1 3.2 3.2 Table 1 Growth hubs – county index scores 20 Mid-case counties Lastly, there are the ‘mid-case’ counties, whose economic and social context is typically modest, often because of their peripheral location and economic concentration on agriculture. However, other dimensions are typically ‘good enough’ (though not good). For example, many of these counties have infrastructure that is not excellent, but ‘good enough’, with recent and ongoing improvements. Similarly, their political and institutional strength is not top-class, but may include county governments with credible economic development plans and an ability to execute. They typically have positive community relations and reasonable security, with issues mainly involving petty crime. One notable exception is that they often have environmental problems either relating to climate change or poor management of the local environment (Figure 13 and Table 3). In the next section, we will draw on this analysis to consider the implications for inclusive economic growth, including policies and programmes designed to increase inclusion. Figure 12 Poverty logjams – stylised facts • Reliance on informal, low-productivity occupations including subsistence agriculture and informal services • Undiversified income and limited or no remittances • Poor health, education and financial-service quality and access • Lack of infrastructure: electricity, roads, mobile/internet access • Lack of market access • Limited ‘ecosystem’ of MSMEs and economic linkages • Arid locations • Climate vulnerability • Environmental degradation • Poor land management • Conflict and crime • Food insecurity • Exposure to shocks • Marginalised communities • Remote areas • Weak institutions including local government, rule of law and property rights • Gender inequality • Inadequate grievance resolution • Corruption • Lack of contracted employment • Lack of land and property ownership Economic and social Political and institutional Environmental Conflict and security County Poverty levels (national quintile) SIDA dimension Economic andsocial Environmental Political and institutional Security and conflict All dimensions Mandera 5.0 2.4 4.0 2.6 2.3 2.8 Wajir 5.0 1.9 5.0 2.8 3.5 3.3 Marsabit 5.0 2.0 5.0 3.2 2.0 3.1 Busia 5.0 2.2 1.0 2.8 3.0 2.3 West Pokot 5.0 2.0 2.0 3.0 2.0 2.3 Turkana 5.0 1.7 3.0 3.0 2.5 2.5 Garissa 5.0 1.6 3.0 2.4 4.0 2.7 Average 5.0 2.0 3.3 2.9 2.5 2.7 Table 2 Poverty logjams – county index scores 21 Figure 13 The ‘mid-case’ counties – stylised facts • Physical proximity to growth hubs and trade corridors with potential to build linkages to the local economy • Improving or adequate infrastructure including links from county to growth hubs and trade corridors • Other factors either improving or adequate (health, education, financial access, infrastructure) • Opportunities for income growth and diversification • Proactive management to tackle environmental problems • Reasonable security and absence of conflict • Local government supportive of pro-poor economic growth • Credible county economic plan that is consistent with pro-poor outcomes and planned investment • Supportive national policy and development partners Economic and social Political and institutional Environmental Conflict and security County Poverty levels (national quintile) SIDA dimension Economic andsocial Environmental Political and institutional Security and conflict All dimensions Siaya 2.0 3.0 1.0 3.0 2.5 2.4 Nyamira 2.0 3.0 2.0 2.4 2.3 2.4 Nyandarua 3.0 3.6 4.0 3.4 3.8 3.7 Nandi 3.0 3.2 4.0 3.4 4.0 3.7 Trans Nzoia 3.0 2.9 4.0 3.6 4.3 3.7 Kajiado 3.0 2.9 5.0 3.4 3.5 3.7 Homa Bay 2.0 2.3 1.0 2.2 2.8 2.1 Uasin Gishu 3.0 3.4 2.0 2.6 2.8 2.7 Baringo 3.0 3.2 3.0 2.8 2.3 2.8 Bungoma 3.0 3.0 1.0 3.4 2.5 2.5 Makueni 3.0 2.7 2.0 3.2 2.8 2.7 Kakamega 3.0 2.7 1.0 2.8 3.0 2.4 Bomet 4.0 2.9 2.0 3.8 4.8 3.4 Migori 3.0 2.2 1.0 2.0 3.0 2.1 Elgeyo Marakwet 4.0 2.9 3.0 3.2 3.8 3.2 Vihiga 4.0 3.1 1.0 3.6 3.5 2.8 Kwale 4.0 3.1 2.0 3.0 3.0 2.8 Isiolo 4.0 2.4 4.0 3.4 3.0 3.2 Kilifi 4.0 2.9 2.0 3.0 2.8 2.7 Laikipia 4.0 2.8 4.0 2.4 2.5 2.9 Kitul 4.0 2.2 3.0 3.6 2.8 2.9 Kisii 4.0 2.2 2.0 2.8 3.3 2.6 Tana River 5.0 2.6 5.0 3.0 2.5 3.3 Samburu 5.0 2.4 5.0 2.8 2.5 3.2 Average 3.4 2.8 2.7 3.0 3.1 2.9 Table 3 The ‘mid-case’ counties – county index scores 22 3 What are the implications for policy andinterventions? 21 See case-study countries in Annex 3 for a more detailed discussion of the evidence. 22 See econometric analysis in Annex 2 that substantiates this finding. The SIDA framework dimensions – economic and social, environmental, political and institutional, and conflict and security – have an overarching effect on poverty reduction. This is because they are significant determinants of whether a household can successfully improve its livelihood and whether this improvement is permanent or temporary. In this section, we explore this issue further and consider what can be done to make economic growth more inclusive. 3.1 The importance of SIDA ‘context’ The four SIDA framework dimensions are key determinants of poverty levels and dynamics due to their powerful and pervasive effects on livelihoods at the household level (Table 4).21 In the economic and social context, infrastructure – including power, transport networks, telecommunications and finance – is fundamental to economic activity and livelihoods. Indeed, such is the importance of infrastructure that differences are a key cause of inequality between counties.22 Infrastructure determines households’ ability to undertake ‘value-added’ activities. For example, without power, there can be no agricultural processing or manufacturing; without roads, it is difficult to access new markets; without finance there can be no investment to boost farm or factory productivity. Human capital is also an essential part of the economic and social context. Weak vocational skills prevent the poor from taking waged employment. Know-how is needed for success in improving agricultural production and marketing. Well-connected communities (those with ‘social capital’, another form of human capital) are better able to help each other, lift aspirations and connect with national and international organisations. In the environmental context, climate is a key risk to development. Counties that are highly exposed to climate risk – in particular, the north-eastern and coastal counties of Kenya – are already experiencing disruption to agricultural and livestock production, damaged or destroyed infrastructure and increased conflict following disasters such as droughts, flooding and cyclones. These events are powerful shocks to poor households from which it is difficult to recover. They can be a major reason why households are trapped or revert to living below the poverty line (Diwakar and Shepherd, 2018). Degradation of the local environment is also key. In urban slums, poor sanitation and hygiene undermine health. In rural areas, degradation caused by overly intensive agriculture, overfishing and overgrazing reverses any gains in agricultural productivity (World Bank, 2010; 2016; Wankuru et al., 2019). Some counties are starting to tackle these problems through better agriculture and 23 fishery management, often in partnership with expert bodies, such as universities and international donors, but more needs to be done – andquickly. The political and institutional context is also an overarching dimension. It determines whether governments deliver public infrastructure and services and provide critical coordination with the private sector (Lin et al., 2012).23,24,25 In Kenya, the success of devolution and the quality of county government are key to prospects for poverty alleviation. Growth hubs typically have strong county governments and significant connections to national government that give them an advantage in areas such as accessing national budgets and affecting nationalpolicy. Mid-case counties are also benefitting from successful devolution, including good county development plans and only moderate levels ofcorruption. In contrast, poverty logjam counties may have county governments that are less able, with pervasive corruption – although the presence of strong communities can provide informal governance, which is positive. Lastly, security and conflict are prevalent factors, especially in extreme poverty. In the north-eastern counties, there are severe security problems due to conflict between ethnic groups and from insurgents. This creates fundamental problems, such as food insecurity and violence. In other counties, in contrast, security problems are more likely to be less severe, stemming from issues such as petty theft, prostitution and drug-taking. 23 Annex 3 details the fieldwork case studies in Busia and Kwale that substantiated this analysis. 24 For example, price intelligence; market requirements; supplier and buyer networks; business networks with growth hubs; disintermediation of ‘middlemen’. 25 For example, through mentors, role models and supporting migration. 26 Diversifying income by developing income streams from non-farm businesses in otherwise agriculture-dependent households can be of great benefit. Income can also be diversified across the total household ‘portfolio’, for instance through different occupations in a household. Migration can also be important, as it enables family members to send remittances to relatives, supporting consumption and providing funds for investment in livelihoods. 27 Assets can include financial assets, such as savings and insurance, traditional assets, such as land and livestock, and business assets. 3.2 The need for context-led policy and interventions There is a significant body of literature on ‘pathways out of poverty’, showing that households can permanently escape from poverty when three things are achieved in combination: an increase in absolute income, a diversification of income and the accumulation of assets (Shepherd and Diwakar, 2019). Increasing household income is one way out of poverty, but diversification26 and asset accumulation27 protect households from shocks and increase their resilience – ensuring their escape is permanent, not temporary. Furthermore, these factors encourage spending on education, healthcare and investment in business assets, all of which further solidify the escape from poverty (Shepherd and Diwakar, 2019). One thing that is evident from this analysis is that the literature on how households escape from poverty needs to be put in the broader context, as this is a key determinant of individual and household ability to escape poverty. In other words, the escape from poverty is subject not just to individual characteristics, but the circumstances in which people find themselves. This has two important implications. First, to increase the inclusiveness of economic growth, there is a need to tackle the overarching problems in the dimensions of the SIDA framework, such as infrastructure, human capital, politics and the environment. This is not to say that households should not receive assistance in parallel with tackling these problems. However, such interventions should 24 only be attempted where the context is (not necessarily perfect, but) ‘good enough’. A key insight from Chapter 2 is that, in Kenya, ‘good enough’ contexts are to be found in ‘growth hubs’ or ‘mid-case’ counties. So, for example, before attempting an intervention, the political environment will need to be examined (the credibility of national and local government and their policies, for instance) to see if political bodies are tackling problems such as infrastructure, healthcare, education, conflict, security and corruption and whether they have credible economic plans that can be built upon and coordinated with household-level programmes. This is particularly relevant in the case of policies and interventions that aim to foster private-sector development and job creation. These are very common among national governments, development agencies and nongovernmental organisations (NGOs). However, they are unlikely to succeed in providing Table 4 How the SIDA framework affects ‘pathways out of poverty’ SIDA framework dimension Effects Economic and social Infrastructure Power Mechanisation, irrigation, processing, manufacturing Transport Physical access to markets ICT ‘Soft’ market intelligence25 Access to finance (via mobile banking) Finance Investment in productive assets (agricultural inputs, plant, equipment, private power connections) Working capital for businesses (trade and value-chain financing) Household resilience and stability Human capital Health, education Basic human well-being and skills Vocational skills Ability to apply higher-value techniques in agriculture Ability to gain waged employment Communities, networks (‘socialcapital’) ‘Soft’ market intelligence Social aspirations and values26 Links to ‘expert’ research organisations, community organisations, carbon credit schemes Environmental Climate risk Aridity, erratic rainfall, flooding, cyclones Disruption to agriculture Destruction of public and private infrastructure Increasing conflict Local degradation Soil, water, coastal habitats, rubbish, sanitation Agricultural production Human well-being, including health and food security Political and institutional Government and public services County governments, corruption, devolution Quality of economic planning, infrastructural development, education and health, community engagement Funding and execution of public programmes of education, health and social protection Security and conflict Crime rates and conflict Policing, border security Loss of assets Food security Mortality 31 KNBS (2018a) Kenya FinAccess survey, 2018. Nairobi: KNBS (http://statistics.knbs.or.ke/nada/index. php/catalog/96). KNBS (2018b) Foreign investment survey report 2018. Joint report by the Kenya National Bureau of Statistics, the Central Bank of Kenya and the Kenya Investment Authority. Nairobi: KNBS (www.knbs.or.ke/?wpdmpro=foreign-investment-survey-report-2018). KNBS (2019) Gross county product 2019. Nairobi: KNBS (https://africacheck.org/wp-content/ uploads/2019/02/Gross-county-Product-County-GDP.pdf). Kodiaga, T. (2013) Water security and climate resilience project (WSCRP): environmental and social management framework (ESMF) (P117635). Final report. Nairobi: Government of Kenya, Ministry of Water and Irrigation (http://documents.worldbank.org/curated/en/582711468048849575/ Final-report-WSCRP-environmental-and-social-management-framework-ESMF). Krishnan, A., te Velde, D.W. and Were, A. (2019) Integrating Kenya’s small firms into leather, textiles and garments value chains: creating jobs under Kenya’s big four agenda. London: ODI (https:// set.odi.org/wp-content/uploads/2019/05/Summary-paper_MSMEs-in-Kenya-value-chains_May- FINAL-3.pdf). Kuguru, M. (2018) ‘NHIF targets to list 362,000 elderly persons for Universal Health Coverage’. TheStar, 20 August (www.the-star.co.ke/news/2018/08/20/nhif-targets-to-list-362000-elderly- persons-for-universal-health_c1805848). Kwale County Government (2019) Kwale county annual development plan FY2017/2018. Kwale, Kenya: Kwale County Government (www.youthagenda.org/mainwebsite/wp-content/ uploads/2019/12/Kwale-County-Annual-Development-Plan-2018.pdf). Lin, J.Y. (2012) New structural economics: a framework for rethinking development and policy. Washington DC: World Bank (http://siteresources.worldbank.org/DEC/Resources/84797- 1104785060319/598886-1104951889260/NSE-Book.pdf). LREB – Lake Region Economic Bloc (2018) The lake region economic blueprint: a better life. Nairobi: Council of Governors (https://cog.go.ke/phocadownload/reports/Lake_Basin_web.pdf). McMillan, M., Page, J., Booth, D. and te Velde, D.W. (2017) Supporting economic transformation: an approach paper. London: ODI (www.odi.org/publications/10748-supporting-economic- transformation-approach-paper). Michalopoulos, S. and Papaioannou, E. (2012) (revised 2013) National institutions and subnational development in Africa. NBER Working Paper 18275. Cambridge MA: National Bureau of Economic Research (www.nber.org/papers/w18275). Mulefu, F.O., Mutua, F.N. and Boitt, M. (2016) ‘Malaria risk and vulnerability assessment GIS approach. Case Study of Busia County, Kenya’. IOSR Journal of Environmental Science, Toxicology and Food Technology 10(4): 104–112 (www.iosrjournals.org/iosr-jestft/papers/vol10-issue4/ Version-1/N100401104112.pdf). Mulinge W., Gicheru, P., Murithi, F., et al. (2016) ‘Economics of land degradation and improvement in Kenya’ in E. Nkonya, A. Mirzabaev and J. von Braun (eds) Economics of land degradation and improvement: a global assessment for sustainable development. Cham, Switzerland: Springer. Nyarko, Y. (2016) ‘Commodity exchanges in Africa: transforming African agriculture and finance’. Policy Brief. London: London School of Business, International Growth Centre (www.yawnyarko.com/ assets/ publications/05.06_CommodityExchangesinAfrica-TransformingAfricanAgricultureandFinance.pdf). OECD – Organisation for Economic Co-operation and Development (2017) Promoting pro-poor growth policy guidance for donors. Paris: OECD (www.oecd-ilibrary.org/docserver/ 9789264024786-7-en.pdf?expires=1579714852&id=id&accname=guest&checksum= 466D23DFF3C73542A2C23F62CA04A773). Omoro, F. (2001) Household welfare monitoring and evaluation survey. Embu District: Kippra Dp No.125. Embu, Kenya: Embu County Government. 32 Oxford Policy Management (2018) Evaluation of the Kenya hunger safety net programme phase 2: impact evaluation final report. Oxford, UK: Oxford Policy Management (www.opml.co.uk/files/ Publications/a0013-evaluation-kenya-hunger-safety-net-programme/impact-evaluation-final-report- ex-summary.pdf?noredirect=1). Pape, U.J. and Mejia-Mantilla, C. (2018) Kenya poverty and gender assessment 2015–2016: reflecting on a decade of progress and the road ahead. Washington DC: World Bank (http://documents.worldbank.org/curated/en/905491550155961925/Kenya-Poverty-and-Gender- Assessment-2015-2016-Reflecting-on-a-Decade-of-Progress-and-the-Road-Ahead). Pkalya, R., Adan, M. and Masinde, I. (2003) Conflict in northern Kenya: a focus on the internally displaced conflict victims in northern Kenya. Rugby, UK: Practical Action (https://practicalaction.org/ docs/region_east_africa/conflict_in_northern_kenya.pdf). Ponge, A. (2013) ‘Gender mainstreaming and women empowerment in political party processes in Kenya: implementing the new constitution in earnest’. International Journal of Humanities and Social Science 3(3): 62–71. REACH (2015) Country diagnostic report, Kenya. REACH Working Paper 3. Oxford, UK: University of Oxford, School of Geography and the Environment (https://profiles.uonbi.ac.ke/gouma/files/ country_diagnostic_report_kenya.pdf). Republic of Kenya (2016) Kenya national adaptation plan 2015–2030: enhanced climate resilience towards the attainment of Vision 2030 and beyond. Nairobi: Republic of Kenya (www4.unfccc.int/ sites/NAPC/Documents%20NAP/Kenya_NAP_Final.pdf). Shepherd, A. and Diwakar, V. (2019) Pathway to sustained poverty reduction: balancing growth from above and below. London: ODI (www.odi.org/sites/odi.org.uk/files/resource-documents/12772.pdf). Shepherd, A., Scott, L., Mariotti, C., et al. (2013) Chronic poverty report 2014–2015: the road to zero extreme poverty. London: ODI (www.odi.org/sites/odi.org.uk/files/odi-assets/publications-opinion- files/8834.pdf). Shepherd, A., Bird, K., Dacorta, L., et al. (2019) Thefourth chronic poverty report: growth. London: ODI (www.odi.org/publications/11404-fourth-chronic-poverty-report-growth). SIDA – Swedish International Development Cooperation Agency (2017) ‘Dimensions of poverty SIDA’s conceptual framework’. Stockholm: SIDA (www.sida.se/English/publications/149106/dimensions- of-poverty-sidas-conceptual-framework/). Thomas, K., Masinjila, M. and Bere, E. (2013) ‘Political transition and sexual and gender-based violence in South Africa, Kenya, and Zimbabwe: a comparative analysis’. Gender & Development21(3): 519–532. Tyson, J.E. (2018a) ‘One-stop border posts and informal livelihoods in Busia, Kenya’. Development in Practice 28(4): 562–573. Tyson, J.E. (2018b) Financing special economic zones: different models of financing and public policy support. Supporting Economic Transformation. London: ODI (https://set.odi.org/wp-content/ uploads/2018/09/SET_Financing-Models-for-SEZs_Final.pdf). Tyson, J. and Beck, T. (2018) Capital flows and financial sector development in low-income countries. Growth Research Programme Synthesis Report. London: ODI (https://degrp.odi.org/wp-content/ uploads/2018/10/DEGRP-Capital-flows-report-2018.pdf). UNDP – United Nations Development Programme (2017) Income inequality trends in sub-Saharan Africa: divergence, determinants and consequences. New York, NY: UNDP (www.africa.undp.org/content/rba/ en/home/library/reports/income-inequality-trends-in-sub-saharan-africa--divergence--dete.html). UNICEF – United Nations Children’s Fund (2017) Situation analysis of children and women in Kenya, 2017. Nairobi: UNICEF (www.unicef.org/kenya/media/136/file/SITAN-report-2017-pdf.pdf). Wankuru, P.C., Dennis, A.C.K., Angelique, U., et al. (2019) Kenya economic update: unbundling the slack in private sector investment: transforming agriculture sector productivity and linkages to poverty reduction. Kenya Economic Update No. 19. Washington DC: World Bank (http://documents.worldbank. org/curated/en/820861554470832579/Kenya-Economic-Update-Unbundling-the-Slack-in-Private- Sector-Investment-Transforming-Agriculture-Sector-Productivity-and-Linkages-to-Poverty-Reduction). 33 Wanyama, F.O. and McCord, A. (2017) The politics of scaling up social protection in Kenya. ESID Working Paper No. 87. Manchester, UK: University of Manchester, Effective States and Inclusive Development (www.effective-states.org). WEF – World Economic Forum (2017) The global gender gap report 2017. Geneva, Switzerland: WEF (www.weforum.org/reports/the-global-gender-gap-report-2017). WHO – World Health Organization (2009) WHO guide to identifying the economic consequences of disease and injury. Geneva, Switzerland: WHO (www.who.int/choice/economicburden/en/). World Bank (2010) Kenya: agricultural productivity and sustainable land management project. Washington DC: World Bank (http://documents.worldbank.org/curated/en/675841468273620112/ Kenya-Agricultural-Productivity-and-Sustainable-Land-Management-Project). World Bank (2012) ‘Special focus: creating jobs’ in Kenya economic update: Kenya at work – energizing the economy and creating jobs. Washington DC: World Bank (www.worldbank.org/content/dam/ Worldbank/document/Africa/Kenya/kenya-economic-update-december-2012-special-focus.pdf). World Bank (2013) Growing Africa: unlocking the potential of agribusiness. Washington DC: World Bank (https://siteresources.worldbank.org/INTAFRICA/Resources/africa-agribusiness-report-2013.pdf). World Bank (2016) Transforming livelihoods along Kenya’s coastal areas. Washington DC: World Bank (www.worldbank.org/en/news/feature/2016/07/11/transforming-livelihoods-along-kenyas-coastal-areas). 34 Annex 1 SIDA index methodology The index draws on a framework adapted from SIDA’s multidimensional poverty framework (see Figure 1 in the main document). To date, however, the framework has been largely used as a qualitative tool. In this study, we wish to extend the SIDA framework to create a quantitative analysis using an original methodology to provide a numerical index of its four dimensions and an overall numerical score across all four dimensions of its framework. This SIDA index was created by examining each of the four dimensions and identifying critical aspects and data that could indicate the quality of each dimension. The data for each dimension was then selected from the KIHBS 2015–2016 (KNBS, 2016), the Gross County Product 2019 (KNBS, 2019), the Kenya FinAccess Survey 2018 (KNBS, 2018a) and recent ODI SET and CPAN research, as well as other sources of research and data, including national statistics and the broad economic literature (detailed in the data prepared by the ODI and published by FSD Kenya in conjunction with this paper). The details of the dimensions are as follows. •Economic and social dimensions: agricultural productivity (as indicated by county agricultural output per capita); density of SMEs per capita; financial inclusion; primary education graduation rates; distance to the nearest all-weather road; the number of markets in the county; electrification rates as a percentage of the population with access to mains electricity; county public spending per capita on education; and county public spending per capita on health •Environmental dimensions: environmental risk at county level, as determined by the KNBS •Political and institutional dimensions: households reporting experience of bribery in the past two years; households reporting experience of grievances and disputes in the past two years; gender inequality in employment contracts; percentage of the population engaged in employee associations; percentage of the population owning their own homes •Conflict and insecurity: levels of fatality and crime; percentage of the population suffering from food insecurity; percentage of households reporting shocks in the past year The data for each of these factors was ranked into quintiles for each county, with a low ranking suggesting a negative indicator and a high ranking suggesting a positive indicator. (For example, a high score for agricultural productivity would be positive, so a high level on this data point would score 5; however, high food insecurity would be negative, and so a high level on this data point would score1.) These quintiles were then averaged across all of the data fields for each SIDA dimension to give the average quintile ranking for each dimension (TableA1). A composite index was then created by weighting the four dimensions to give an overall SIDA framework index. The weightings were: 60% for economic and social ranking, 10% for environment, 20% for political and institutional dimensions and 10% for conflict and security dimensions. ODI researchers can provide the database, with full details of the index computations used for the study. 35 SIDA index ranking Weighting Economic and social 30% Agricultural productivity High MSMEs per capita High Fin inclusion High Primary education High Distance to road Low No. of markets High Electricity mains High Public spending per capita – education High Public spending per capita – health High Environmental 5% Natural risk Low Political and institutional 10% Bribes Low Employment gender inequality Low Grievance Low Employment association High Own household High Conflict and insecurity 5% Fatalities Low Crime index Low Food insecurity Low Shock Low Table A1 SIDA index rankings and weightings 36 Annex 2 The spatial productivity analysis We model each county’s performance as a function of geographical and spatial variables (see TableA2). This multivariate analysis allows us to tease out the relationship between indicators, while keeping constant other factors that might affect productivity. A2.1 Estimations using county-level panel data from 2005 to 2015 We estimate the productivity performance of each county using a panel data stochastic production frontier model, assuming that each county produces its output using its populace and public capital stock Y i t = T L( K, L, t ) i t + v i t – u i t [1] where T L( K, L, t ) i t represents the translog approximation of the production function. In the absence of direct, consistent information on county-level output, we follow the literature (for example, Henderson et al., 2012; Michalopoulos and Papaioannou, 2012; Hodler and Raschky, 2014) on using satellite data on night light or luminosity density for 2005–2015 as a proxy for county-level income/ output, Y i t. Inspired by Burgess et al. (2015), we use total road expenditure (sum of development expenditure [new investment] and recurrent expenditure [maintenance]) as a proxy for county-level capital K i t, while population is treated as the measure of labour L i t. t is a time trend measuring technical progress. The composed error term v i t – u i t is such that v i t is the traditional idiosyncratic error representing noise, while u i t is a one-sided error term that represents the level of inefficiency (inother words, it measures the productivity slack of each county in terms of its actual output, relative to its potential on the frontier, as well as relative to other counties). Hence, our county efficiency performance is estimated as: E ƒ ƒ i t = E [ exp ( –u i t ) | i t] [2] i t = v i t – u i t [3] Note that there may be a degree of uncertainty in the performance scores, as the production function inputs and outputs are proxies rather than inputs and outputs in the traditional sense. Even so, the sensible relationships between the productivity performance and spatial/geographical variables suggest that the estimates fit the data and reality to a reasonable extent. We also link county-level productivity to poverty. To do this, we regress poverty rates from 2005 and 2015 on the estimated performance scores for a correlation exercise. The idea is to see the extent to which improved economic performance affected poverty. This regression should allow for a cursory evaluation of whether improved economic productivity performance is associated with poverty reduction. Second, the regression could be used as a sense check on the analyses conducted so far, as one would theoretically expect increased productivity to help reduce poverty. We then approximate estimates at each data point using the coefficient from the ordinary least squares (OLS) regression model. 37 A2.1.1 Results The regression results indicate that having a port terminal, larger area and proximity to the highway between Mombasa–Nairobi–Kampala are associated with reduced inefficiency, or increased productivity, during the sample period, where productivity is measured in terms of luminosity density for 2005–2015 as a proxy for county-level income/output. Moreover, a 10% increase in the distance to Nairobi is associated with a 2.2% decrease in productivity. Counties with ports or proximity to the highway or capital may be able to benefit more from improved services, infrastructure and transportation links as a result, so these relationships are intuitively sound. Area size is interpreted as elasticities at the sample mean, so the positive association between area size and increased productivity may not hold along the entire data distribution of counties, particularly for small-sized growth hubs, where the size-productivity relationship is reversed. Figure A1 maps the changes in efficiency scores over time to demonstrate spatially how regions have evolved relative to others. These scores represent the proximity of a county’s actual output to its production frontier, or its optimum level of output given the various inputs outlined in Table A2. The closer a county’s actual output lies to its production frontier and relative to other counties, the more efficient it is said to be. Simply put, as efficiency scores increase, a county becomes more efficient in translating its inputs into productive outputs. Separately, we link county-level productivity to poverty in an additional exercise for the years 2005 and 2015, where poverty estimates are available. Our estimates suggest that a 10% improvement in countylevel productivity is associated with an approximate 1% reduction in the poverty rate. This analysis is for illustrative purposes only, as productivity is just one aspect likely to affect poverty rates across countries, as this report exemplifies. 0.816099–1.000000 0.683882–0.816099 0.503938–0.683882 0.237943–0.503938 0.928394–1.000000 0.785036–0.928394 0.447903–0.785036 0.157524–0.447903 2005 2015 Correlation Coefficient Mines –0.206 Port terminal –3.518* Area size –0.435* Border with Tanzania or Uganda 0.352 Distance to Nairobi 0.228* Five richest counties –0.385 Highway between Mombasa-Nairobi-Kampala –1.191* Note: The asterisk denotes statistical significance at 1%. Table A2 Correlations of county productivity, regression results Figure A1 Change in productivity estimates 38 Annex 3 Fieldwork case studies As part of the preparations for the main working paper, researchers undertook fieldwork in the case-study counties of Kwale and Busia to deepen the evidence and conceptualisation of the spatial dynamics of poverty and the SIDA framework. These counties were selected as being representative of a mid-case county and a county lying on the cusp of being a poverty logjam, respectively. The goal of the fieldwork was to assess in greater depth the nature of the four SIDA dimensions and how they interact on the ground. The fieldwork was completed using qualitative interviews combined with a detailed review of the data underpinning the SIDA index for these counties. Interviews were conducted with a wide variety of informants, including county government officials, local business groups, farmer, fisher and pastoralist focus groups and development agencies operating in the counties. These included the World Bank, the University of Nairobi and various NGOs. A3.1 Kwale Kwale has a poverty rate of 47.4%, well above the national average of 40.5%, and lies in the second and bottom quintile nationally. The population also suffers from widespread poor nutrition and infectious diseases, including malaria and diarrhoea. Child deprivation and stunting are high (REACH, 2015). Extreme poverty is better than the national average, although equality is marginally worse. A3.1.1 Economic and social context The Kwale County Development Plan (CDP) for 2017–2018 reports that productivity in agriculture, livestock and fisheries remains low for a number of reasons (Kwale County Government, 2019). Agriculture is reliant on rain-fed water sources. Farmlands and coastline fisheries have been affected by the degradation of soil and water sources. Poor agronomic practices, resulting from limited skills and knowledge of crop and livestock husbandry, and small farm size suppress productivity. There are poor market linkages and a lack of marketing skills and information. Farming cooperatives exist, but they are not well organised, and many are inactive. This causes farmers and fishermen to rely on middlemen, reducing the prices they receive for their produce. The CDP reports that there is little capital investment in agriculture, resulting in limited or no mechanisation and inadequate storage and transport facilities. Furthermore, investment to rectify this has been constrained because overall financial access is low. Formal access is limited, as only 22.5% of Kenyan land has title deeds, making it difficult to provide collateral for borrowing (Kwale County Government, 2019). Our findings from fieldwork, including focus groups and interviews with key informants (from areas such as the county government and the private sector) were broadly consistent with these sources. Agricultural production is concentrated in the interior of the county, with fertile areas nearer the coast and arid agricultural lands in the hinterland. In the fertile coastal areas, the soil provides opportunity for cash crops such as mangos, oranges, cashews and coconut. These are orchard-based, so help conserve soil and prevent erosion in addition to potentially yielding good income. Currently, there is little in the way of marketing and value 39 addition. However, two fruit-processing plants are being built and the government is building a large fruit and vegetable market on the coast road to sell into the markets in Mombasa and Tanzania. In the hinterland, there is a concentration of small, low-productivity farms that focus on the production of subsistence crops. Some of the barriers to improving productivity are typical of rural areas in Kenya, such as a lack of access to finance and reliance on brokers to intermediate goods to market, leading to lower farm-gate prices. Few examples of agricultural processing or manufacturing were reported by focus-group interviewees. Households in the hinterland areas are not necessarily much poorer than those living in the coastal areas. Unlike the coast, however, there is little economic opportunity in the hinterland, where all the forest has been denuded for charcoal, the soils are sandy and not good for crops and there is no water. There have been attempts to address this by building dams, but these have proved sinks for capital, which gets siphoned off through corruption (only one of four has actually been built) and have added very little value locally. The one functioning dam has been used to start an irrigation project, but there have been operational problems, including the cost of running irrigation equipment and the silting up of dams. Interviewees complained that they cannot compete with the superior produce coming in from the Taita hills. People have also been forced to grow particular crops recommended by the government (onions, tomatoes, kale, etc.). Left to their own devices, they would prefer to grow maize, because it can be used as a subsistence crop. However, such crops do not justify the cost of irrigation, leading to a vicious circle of agricultural development. Overall, the hinterland offers little economic potential other than through labour markets and businesses that exploit the substantial opportunities in the nearby coastal belt and large towns. There is a large road being built through this area, which will link Tanzania and upcountry Kenya with Kwale, bypassing the Mombasa bottleneck. Livestock farming has been viable in these arid areas, but has suffered from the breakdown of government extension services and agricultural policy. The hinterland has also been degraded by uncontrolled charcoal production and sale as a ‘cash crop’, compromising the already marginal productivity of the land and potentially leading to more severe degradation in the longer term. One exception to this was a project we visited in a remote fishing community in the south of the county. In this project, villagers had worked with the county government to establish seaweed farming and processing on the coastal beaches. The project was largely the domain of the village’s women, each with an individual patch of a natural tidal area for farming and processing. The farming involved installing wire netting in tidal areas around mangrove swamps, where seaweed was planted and harvested. Farmers had also planted mangrove saplings as part of a project to reinvigorate the mangrove swamps. The village had established processing facilities to dry the seaweed, process it into soap or sell it in bulk. Sales were made to a single intermediary who visited the village periodically. The villagers reported reasonable increases in income from this activity. However, they also had no market information and were ‘price takers’ from the single intermediary. They were unable to optimise the gains of their processing and production due to a lack of market information on pricing, a network to sell their products and transportation. Another exception was a coconut factory that processed raw coconuts to produce coconut oil for cosmetics and food. As part of our fieldwork, we visited the factory and interviewed key informants at the firm. Its production relied largely on basic machinery and was highly labour intensive. This had limited production in terms of volume and quality. Interviewees also reported constraints in terms of finance, sales and marketing. Financially, the firm was reliant on funds from farmer cooperatives and businessmen in the district and business growth was constrained by a lack of access to factoring and general trade finance. It had already established exports within the East African community and had growing exports to Europe and the United States. However, it had not developed to its full potential in terms of marketing its coconut oil, by publicising the health benefits, or producing coconut milk. 40 The firm was planning to invest $150,000 in improved plant, which would increase its production capacity and quality (and meet export standards), allow it to process both oil and coconut milk, and improve productivity by reducing processes and consolidating them into a single site. It is interesting to consider the knock-on effects of such investment. Kwale has significant areas of coconut plantation, as well as informal coconut trees. The firm currently processes 400,000 to 600,000 coconuts a month. However, it had reached the upper limit of its ability to process coconut and, in fact, had purchased no coconuts for several months because of stockpiling after a good harvest. The expansion of the facilities should lead to a significant increase in the volume of coconuts that can be processed daily, leading to significant growth in the backward linkages to coconut farmers. These include relatively small farmers, as the firm also has transport facilities that enable it to collect directly from the farm gate. However, the effects of the development are more negative in terms of direct employment. Most notably, the introduction of new facilities will reduce the number of employees significantly, from 160 to an estimated 20–30, although the remaining jobs are likely to be higher paid, as they will be highly skilled employees. Other new jobs could be created, too, for example, in sales and marketing and export facilitation. Efforts have been made by the county government and local farmers to extended production away from subsistence farming towards market gardening and livestock. Focus-group interviewees indicated mixed results. There have been improvements, led by the county government, including improved control of disease and pests, both for both crops and livestock, and greater access to power. However, the results of irrigation projects have been mixed, as rainfall has been erratic and because maintenance and the availability of equipment, such as electronic pumps, have been erratic. There has also been a lack of connectivity between the goods produced and market demand. For example, interviewees reported that tomato production had not been matched by market demand. On the coast, the main agricultural activity is fisheries. The fish landing sites are organised by Beach Management Units (BMUs) – community organisations established by way of 2007 legislation for beneficiaries of the beach. Focus groups were conducted with various BMUs along the coastline. Some fisheries remained based on traditional subsistence fishing activity and full focus groups in this category reported a deteriorating situation. Their traditional activities had been based on the use of traditional, home-made boats, which were only suitable for fishing within the interior of the reef. Activities had also been focused on subsistence, with limited sales of fish products. Interviewees in this category reported decreasing catches due to declines in fish stocks within the reef and a rise in aqua-pests such as sea urchins. The World Bank has linked these problems to environmental deterioration, including the pollution of marine environments by soil washed down from up country, and to overfishing. Interviewees reported receiving some assistance from the county, including more modern boats and equipment, but they did not have the technical expertise to use them, so they had either been abandoned or been rented to Tanzanian fishermen, who had greater technical expertise. Interviewees reported attempting to increase cash sales of fish and squid, but this was constrained by a lack of storage facilities. When asked what the barriers were to gaining access to storage facilities, they cited the cost of connection to the mains as prohibitive, even though mains electricity was close by, because of the tourist industry. They also said they had suffered from theft and vandalism of equipment and reported minimal sales to the tourist industry, despite their proximity to major international hotels. There was also a need for a coordinated development strategy for fisheries. The country government was leading this, but lacked resources. Some issues were beyond its control. The marine economy is being exploited by offshore fishing by Europeans, Chinese, Japanese and Tanzanians with little or no value creation for the local economy. The only strategic response has been to ban Tanzanian fishermen from landing in Kenya or fishing inside the reef, but this has not tackled the problems in deeper coastal areas and has led to overfishing and the depletion of fish stocks. 47 such as extension officers. However, the current budgetary emphasis on ‘bricks-and-mortar’ investments often resulted in a trade-off away from behavioural change and capacity-building activities. Agricultural productivity is affected by soft skills, including human capital, but also the strength of the county’s infrastructure. Promisingly, the World Bank is leading a major irrigation project in Busia County, the Lower Nzoia Irrigation Project, which is accompanied by support to increase agricultural production and establish improved market access and value-chain engagement. The county government has also developed 18 small-scale irrigation projects. Some focus groups noted that there were rivers near to study communities, providing additional opportunities. However, despite these opportunities, focus-group participants felt that little was being done to capitalise on existing resources to improve farming yields. Members of one community-based organisation noted that the irrigation project in their area had collapsed and pipes had been dumped there, with no sign of progress due to the misallocation of funds. Also, while farmers perceive irrigation to be helpful in improving yields for horticultural crops, there were also palpable fears about irrigation infrastructure: •water dams might be used for the disposal of dead animals and thus create hygiene concerns •people would be chased away from their land with little compensation •rich men had bought land in the area early, knowing that it would be used for the dam and would thus lead to high compensation •electrification remained too expensive for many poor households (‘If a household has electricity, there must be a school nearby’ noted one community respondent, reflecting the high cost of electricity), leading some farmers to siphon electricity from the mains. Sometimes these problems arose from an emphasis by leadership on going to scale without first understanding how production could be optimised. As a result, irrigation engineers would often create a water-storage unit and put all resources into the creation of the physical infrastructure, without the subsequent energy or budget for production or maintenance. According to several key informants, this focus on ‘bricks-and-mortar’ infrastructure stemmed from political aspirations. As one interviewee noted, the ‘social audit of political leadership is focused on what is tangible and what can be seen’. Beyond enhancing production, Busia’s engagement in storage, agricultural value chains and trade is also low. Food-storage facilities are predominantly traditional, although there are some modern silos owned by the National Cereals and Produce Board in Malaba. Livestock is widely kept, but husbandry and slaughter facilities are inadequate (Kodiaga, 2013; Busia County Government, 2019). The county has sugar processing at two factories, though other types of processing are limited. For example, although there are cotton mills, they are non-functional. Many focus-group respondents remarked that even though they had raw resources, there was limited value addition. In fisheries, there are two ports on the shores of Lake Victoria, at Sio Port and Port Victoria, but both are in poor condition. A planned fish-cooling plant near Lake Victoria has not yet been commissioned and a cassava-processing factory is not yet complete (Busia County Government, 2019). Other factors are also constraining economic development along agricultural value chains in Busia. Apart from an international trucking route, transport infrastructure is poor; only 10% of county roads have been tarmacked. In the focus groups, however, respondents were quick to note improvements in the road network over the past five years, for example, in reducing the time required to get to nearby towns. The tarmac roads were also brought closer to some communities during this period, enabling the transport of goods to market in a shorter time, even during rainy periods. In one community, a makeshift bridge was constructed to allow bikes to access the sweet-potato farms and which also facilitated links to the market. The effects of these improvements were not uniformly felt; other villages, for example, noted that while there were promising signs that road networks and other infrastructure might improve their well-being, no tangible benefits had been observed as yet. 48 In the absence of further improvements in hard infrastructure, many poor households in Busia are improving their welfare through group-based efforts. In focus groups, table banking was a critical driver of upward mobility for the poor. These were believed to be an improvement on merry-go- rounds chamas,35 by consistently increasing the amount of money available in the group pot. Other self-help groups also provided access to finance and training elements to help empower women. One female respondent noted, ‘I was a housewife before, with no knowledge of kitchen gardening. Now I use wastewater for my garden, so can produce and save’. In the fisheries sector, women joining self-help groups and table banking stated with pride that they were ‘taking over the role of husbands’ in providing household income. These soft infrastructure components have also been supported by hard infrastructure in financing and digitalisation. For example, most of the county is covered by fibre-optic networks and financial access is reasonable (Busia County Government, 2019). Women often invested funds from table banking in their micro-businesses, such as cereals trade or retail, to buy seeds or fertiliser for agricultural activities, to repair or build houses, or to provide school fees for children. Sometimes husbands were unaware of the amount of savings, allowing women relative autonomy on spending decisions. Some would also progress from table banking to a SACCO to access capital, though there were fears of loans being too large and, thus, too risky compared with informal lending arrangements. These fears extended at times into a near-phobia of taking formal loans, which one participant attributed to earlier perceptions of micro-finance institutions and other lenders as exploitative. Table banking was, to some extent, perceived as a substitute for the dearth of social-assistance schemes, which focus-group respondents felt were rare and transient. One key informant echoed this sentiment, saying that social assistance was not taken seriously by the county government. The number of recipients was too small and there was a perception of excessive focus on financial assistance in place of other, complementary assistance, for example the provision of agricultural farm inputs to improve livelihoods, or empowering recipients to use finances through financial literacy training. While these were sometimes evidenced on paper, some informants felt that there was a lack of implementation, often linked to a shortage of staff. In this context, social protection officers felt the best way to nurture self-sufficiency was to encourage beneficiaries to form groups to start table banking, using cash transfers received from the government. Against this backdrop, some things have taken a turn for the better in Busia recently. As part of the development of the East African trade corridors, a new international trucking route has been built that passes through both Busia and Malaba en route from Kampala, Uganda, to Nairobi. The towns are now the sites of new ‘one-stop’ border posts, set up by the national government in partnership with Trade Mark East Africa and opened in 2018. These one-stop posts have drastically reduced the amount of time spent crossing the border. A cross-border traders’ association noted that it might take as little as 15 minutes for a bus to be cleared. The one-stop post has also helped reduce the porous routes along the border, providing a cheaper, safer and often less-taxed alternative. Such a major road should also facilitate access to markets – such as the major urban centres in Kampala and Nairobi – by improving transport. However, the economic implications of these developments have been mixed. There is evidence that the new border post has increased opportunities for small traders, including agricultural goods, by increasing market activity in the border area and through increased engagement with wholesalers, who then sell on goods in major urban areas (Tyson, 2018a). Cross-border traders interviewed for this study also revealed that the one-stop infrastructure had improved foreignexchange transparency, reduced cheating on exchange rates and allowed for smaller denominations to be traded, benefitting poorer traders. However, the advantages were few compared with the problems cited. The county government complained that it had lost revenues that were previously collected at the border post (Busia County 35 An informal cooperative society used to pool and invest savings, often just for women, common in East Africa, particularly Kenya 49 Government, 2019). There had also been a decline in jobs and incomes that were dependent on the border posts, such as porterage and small restaurants and hotels (Tyson, 2018a; Busia County Government, 2019). Focus-group discussions revealed additional challenges related to: •Inefficiencies: When tax systems failed, typically at least once or twice a week, there would be long waiting periods. Traders were not allowed to ride their bikes on part of the road, making it cumbersome to walk their bikes with supplies. This is being addressed by a proposal to include designated bike lanes. •Price concerns: Traders felt there was double taxation, having to pay for goods to enter Kenya at the border, then again when selling them at the market. Accordingly, by the time it came to selling, the price was necessarily high and not competitive. These price differentials with markets in Uganda meant that, in many instances, Kenyan traders were opening businesses such as pubs in Uganda rather than contributing to their local economy. On the other hand, discussants felt that many foreign investors along the Malaba border would exploit Busia’s resources to become rich. For example, they would get quarry at cheap prices in Busia and then build houses for rent at high prices to local residents. •Low levels of human capital: Many traders were illiterate and unable to quantify their businesses, preventing them from borrowing money from formal institutions to expand their businesses. More generally, traders and non-traders felt that the county budget for health and other human capital services was inadequate, with needs based on the census, which often disregarded migrant Ugandans seeking healthcare in Busia. •Limited benefits for locals: Though the revenue-collection process had been simplified, there was no local verification, so youth unemployment was not addressed. More generally, with money now flowing from Kampala, Kigali, Congo or South Sudan directly to Mombasa or Nairobi, local Busia residents often suffered. Even to access local fish from Busia, Ugandans would go straight to Nairobi, leaving Busia fisherfolk with little negotiating power on the local market price. •Heightened vulnerability: Focus-group discussants noted some instances of transactional sex out of fear of goods being seized, though this has reduced thanks to better policing at the border. Having a good rapport with local officers was still critical, however, so that local women did not have to pay each time they crossed the border. Frequent stories emerged, too, of officers collecting bribes or seizing the goods of small traders. Other examples of corruption included groups illegitimately registered as vulnerable groups (people with disabilities, women, youth) to collect small grants; there was limited awareness among rightful claimants of how to access these grants. All these points suggest that cross-border trading has been affected by structural and political constraints common to other aspects of the agricultural and fisheries production cycles. The next subsection explores other contextual dimensions, outside of but linked to the socioeconomic context, which have affected the ability of poor households to escape poverty through agriculture and fisheries. A3.2.2 Environmental context Busia has suffered from environmental degradation, placing it in the highest risk quintile nationally. Deforestation is widespread, with only 2.2% of the county having tree cover in what was once a natural rainforest. This is the result of cutting down trees for firewood, illegal logging and a lack of enforcement of environmental regulations. Combined with high levels of rainfall during rainy seasons, this has led to soil erosion and the pollution of water sources. Other problems include illegal sand harvesting and the poaching of endangered animal species (Busia County Government, 2019). Across livelihoods, weather unpredictability has been a leading cause of downward mobility for households in or near the poverty line, according to focus-group discussions. Even five years ago, the rains were perceived to be more reliable than today. The occurrence of drought alongside inconsistent rainfall has led boreholes to dry up and meant that farmers are perennially worried about water 50 shortages. One focus group attributed this to human activity, such as people cutting trees down for charcoal or turning wetland into farmland, which has also led to reduced rainfall. Moreover, the lakes that do exist are often polluted. Focus-group discussants in one village said they lost several fish stocks last year from a water-borne disease, emanating from dead livestock being thrown into bodies of water. Another focus group noted that earlier this year, they had almost no vegetables due to unreliable rain and worm infestation. The fall army worm has plagued farms in the region for the past two years, with disease affecting rice and other crops. One farmer lamented that as a result of these challenges, ‘at the end of the day, you might not have enough food to feed your family’. A3.2.3 Political and institutional context While permeating numerous sectors, including in the degree of progress made on human capital, infrastructure and trade, the political environment in Busia, in and of itself, shows some weaknesses, according to household survey data. Busia lies in the lowest quintile in terms of corruption and bribery. There have been recent charges relating to corruption, including of senior county government officials (although this has not yet been tested in court). Focus-group participants reinforced the importance of social and political connections across sectors: ‘If I did not vote for you, my child will not get that bursary or assistance. So we do not depend on it’. According to focus-group discussions, devolution has had largely negative outcomes for Busia farmers. One group claimed it had created a duplication of county roles, fuelling nepotism and limiting resources for village development projects. ‘We expect to be near resources in devolution. However they are scattered. There might be budget, but there is little implementation’. According to some respondents, agriculture extension officers used to be more visible under the national government. With devolution, however, there was perceived to be limited support of transportation from the county government, limiting the engagement of extension officers. Another farmer noted that before devolution, she received seeds from the Ministry of Agriculture’s extension officer. Today, despite the One Acre Fund, seeds are only offered as loans and thus require out-of- pocket payments. Another concern related to a lack of prioritisation, with the county government focused on buying tractors. Farmers felt this reflected a lack of contextual understanding, as tractors did not address the root problems associated with limited yields. In another instance, the government bought agricultural machines, but all were grounded and did not reach farming communities. A key informant noted that in the case of rice farmers in Bunyala, problems of corrupt management, marketing and the lack of a vibrant cooperative were compounded by productivity problems stemming from machinery breakdowns and limited farmer efforts to improve rice yields. Another key informant confirmed this, noting that management savings in cooperatives were often not transparent, with members not getting timely payments, creating frustrations that led to the failure of some cooperatives. Moreover, county officials noted that in some cases, there were instances of elite or rich people wanting to join farmer groups and squeeze out smaller farmers. This was linked to selfinterested leadership that prevented trickle-down benefits to farmers. The County Integrated Development Plan (CIDP) focuses on a number of strategies to develop the economy between 2018 and 2022, including the establishment of industrial and special economic zones (Busia County Government, 2019). However, such zones are usually only successful when they achieve clustering effects and the plan does not clearly identify how this will be achieved (Tyson, 2018b). Similarly, there are plans to develop markets and trade but they are predominantly regionally focused and there is a lack of strategic identification of markets and products. More generally, the CIDP is high level and there appears to be a lack of strategic integration of its high-level goals and the individual initiatives which have been budgeted. The CIDP also focuses on the modernisation of agriculture, including fisheries and livestock. Plans include increasing the effectiveness of cluster production units and stepping up value-added activities. Specific activities include providing professional training for farmers, improving the quality and 51 quantity of agricultural inputs and the public provision of tractors and processing equipment. Plans to support engagement in value chains include the construction of factories and storage facilities and the provision of transport vehicles and equipment (Busia County Government, 2019). There are multiple active agricultural programmes with similar goals in Busia. These include development agencies, such as the World Bank and the United Nations, as well as multiple NGOs, self-help groups and national development agencies. As part of the CIDP, the county government is planning to deepen and extend these programmes. County plans also inform and are informed by efforts at the national and regional levels. On the latter, the Lake Region Economic Bloc has articulated key challenges and opportunities in the area in its blueprint report (LREB, 2018). Challenges include declining yields, population pressures and the decreasing size of farm holdings, poor agricultural practices, a lack of title deeds, low uptake of research, a host of climate risks, crop diseases and inadequate market infrastructure (LREB, 2018). The blueprint identifies opportunities for agriculture in Busia including: •farming of food crops (maize, cassava, finger millet, beans, sorghum, rice, sweet potatoes, cowpeas, groundnuts, etc.), horticulture (pineapples, tomatoes, kale, cabbage, etc.), and cash crops (cotton, tobacco, sugarcane, oil palm and pepper) •fishing in Budalang’I and Funyula •promotion of oil crops, root tubers, indigenous vegetables and tissue culture (LREB, 2018). Key LREB informants, similar to those at the national and county levels, noted additional constraints in relation to human resources and low finances. The regional bloc depends entirely on money from counties, often perceived as too low and irregular, with delays in arrival, limiting the bloc’s ability to plan and execute development initiatives. There was also acute awareness of having to rely on outdated data, so plans are under way to create a regional data centre. Devolution has also come with financial constraints, creating a cadre of intermediary officials that focus-group discussants perceived as sometimes obstructing the flow of money and limiting the ability of poor households to meet basic needs. For example, over the past five years, there has been a jiga (chigoe flea) infestation in schools. While the Ministry of Health offered training and provided medication to treat the initial outbreak, no medication is available to treat future outbreaks. This was attributed to corruption diverting the flow of resources. Other challenges noted in focus groups centred on the quality of education, even where access had been improved. For example, though the county government had established learning institutions, in many instances, books were lacking or housing was inadequate for children coming from other parts of the county. A3.2.4 Conflict and security Levels of reported crime are high and Busia is in the second-lowest quintile in this regard, according to an analysis of survey data. Busia County, however, has seen improved security over the past decade, albeit starting from a low base, according to focus-group discussants. This was attributed to the efforts of the county’s former member of parliament, who fostered good communication on community policing, a good relationship between the county governor and the villages, so that they were informed early of any signs of unrest and were on alert, and chief elders maintaining vigilance. Even so, stories of theft from neighbours were common in focus-group discussions. People would steal cows, business assets, farming inputs and fishing supplies. ‘This pulls us behind even if we move up a little’, lamented one respondent. Other forms of insecurity related to cross-border issues, particularly in the fisheries sector. Focus groups highlighted the harassment they experienced from Ugandans when Busia fisherfolk ventured into their waters in search of fish. ‘We are told to eat omena raw’, said one respondent. Others spoke of the deaths of fishermen resulting from excessive harassment by their Ugandan counterparts, creating heightened livelihood insecurity. Evidence. Ideas. Change. ODI 203 Blackfriars Road London SE1 8NJ +44 (0)20 7922 0300 [email protected] odi.org odi.org/facebook odi.org/twitter ODI is an independent, global think tank, working for a sustainable and peaceful world in which every person thrives. We harness the power of evidence and ideas through research and partnership to confront challenges, develop solutions, and create change.