Cover image: Eskay Lim / EyeEm / EyeEm / Getty images Series Editor- in-Chief Kunal Sen UNU-WIDER and University of Manchester About the Series Cambridge Elements in Development Economics is led by UNU-WIDER in partnership with Cambridge University Press. The series publishes authoritative studies on important topics in the field covering both micro and macro aspects of development economics. This Element outlines the origins and evolution of an international award-winning development intervention, index- based livestock insurance (IBLI), which scaled from a small pilot project in Kenya to a design that underpins drought risk management products and policies across Africa. General insights are provided on (1) the economics of poverty, risk management, and drylands development; (2) the evolving use of modern remote sensing and data science tools in development; (3) the science of scaling; and (4) the value and challenges of integrating research with operational implementation to tackle development and humanitarian challenges in some of the world’s poorest regions. This title is also available as Open Access on Cambridge Core. This title is also available as Open Access on Cambridge Core at www.cambridge.org/core E scap in g P o verty Trap s an d U n lo ckin g P ro sp erity in th e Face o f C lim ate R isk JE n SE n e t a l. ISSN 2755-1601 (online) ISSN 2755-1598 (print) Nathaniel D. Jensen, Francesco P. Fava, Andrew G. Mude, and Christopher B. Barrett et al. Escaping Poverty Traps and Unlocking Prosperity in the Face of Climate Risk Development Economics available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of use, https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core Elements in Development Economics Series Editor-in-Chief Kunal Sen UNU-WIDER and University of Manchester ESCAPING POVERTY TRAPS AND UNLOCKING PROSPERITY IN THE FACE OF CLIMATE RISK Lessons from Index-Based Livestock Insurance Nathaniel D. Jensen University of Edinburgh Francesco P. Fava University of Milan Andrew G. Mude African Development Bank Christopher B. Barrett Cornell University Brenda Wandera-Gache African Management Institute Anton Vrieling University of Twente Masresha Taye Institute of Development Studies Kazushi Takahashi National Graduate Institute for Policy Studies Felix Lung World Bank Munenobu Ikegami Hosei University Polly Ericksen University of Vermont Philemon Chelanga Agency for Inclusive Innovations Development Sommarat Chantarat Puey Ungphakorn Institute for Economic Research, Bank of Thailand Michael Carter University of California – Davis Hassan Bashir Agency for Inclusive Innovations Development Rupsha Banerjee International Livestock Research Institute use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core Shaftesbury Road, Cambridge CB2 8EA, United Kingdom One Liberty Plaza, 20th Floor, New York, NY 10006, USA 477 Williamstown Road, Port Melbourne, VIC 3207, Australia 314–321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre, New Delhi – 110025, India 103 Penang Road, #05–06/07, Visioncrest Commercial, Singapore 238467 Cambridge University Press is part of Cambridge University Press & Assessment, a department of the University of Cambridge. We share the University’s mission to contribute to society through the pursuit of education, learning and research at the highest international levels of excellence. www.cambridge.org Information on this title: www.cambridge.org/9781009558242 DOI: 10.1017/9781009558280 © UNU-WIDER 2024 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, with the exception of the Creative Commons version the link for which is provided below, no reproduction of any part may take place without the written permission of Cambridge University Press & Assessment. An online version of this work is published at doi.org/10.1017/9781009558280 under a Creative Commons Open Access license CC-BY-NC-SA 3.0 IGO which permits re-use, distribution and reproduction in any medium for non-commercial purposes providing appropriate credit to the original work is given, any changes made are indicated, and the new work is published under the same licence terms. When the licensor is an intergovernmental organisation, disputes will be resolved by mediation and arbitration where possible. To view a copy of this license, visit https://creativecommons.org/licenses/ by-nc-sa/3.0/igo When citing this work, please include a reference to the DOI 10.1017/9781009558280 First published 2024 A catalogue record for this publication is available from the British Library. ISBN 978-1-009-55824-2 Hardback ISBN 978-1-009-55825-9 Paperback ISSN 2755-1601 (online) ISSN 2755-1598 (print) Cambridge University Press & Assessment has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites referred to in this publication and does not guarantee that any content on such websites is, or will remain, accurate or appropriate. use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of http://www.cambridge.org http://www.cambridge.org/9781009558242 http://dx.doi.org/10.1017/9781009558280 http://dx.doi.org/10.1017/9781009558280 https://creativecommons.org/licenses/by-nc-sa/3.0/igo https://creativecommons.org/licenses/by-nc-sa/3.0/igo http://dx.doi.org/10.1017/9781009558280 https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core Escaping Poverty Traps and Unlocking Prosperity in the Face of Climate Risk Lessons from Index-Based Livestock Insurance Elements in Development Economics DOI: 10.1017/9781009558280 First published online: June 2024 Nathaniel D. Jensen, Francesco P. Fava, Andrew G. Mude, and Christopher B. Barrett et al. Author for correspondence: Nathaniel D. Jensen, njensen@ed.ac.uk Abstract: This Element outlines the origins and evolution of an international award-winning development intervention, index-based livestock insurance (IBLI), which scaled from a small pilot project in Kenya to a design that underpins drought risk management products and policies across Africa. General insights are provided on (1) the economics of poverty, risk management, and drylands development; (2) the evolving use ofmodern remote sensing and data science tools in development; (3) the science of scaling; and (4) the value and challenges of integrating research with operational implementation to tackle development and humanitarian challenges in some of the world’s poorest regions. This title is also available as Open Access on Cambridge Core. Keywords: drought, index insurance, pastoralism, Africa, research for development © UNU-WIDER 2024 ISBNs: 9781009558242 (HB), 9781009558259 (PB), 9781009558280 (OC) ISSNs: 2755-1601 (online), 2755-1598 (print) use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of mailto:njensen@ed.ac.uk https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core Contents 1 Index-Based Livestock Insurance to Address Risk-Based Poverty Traps 1 2 East African Pastoralism: Change and Variability 6 3 Index-Based Insurance for Pastoralist Regions 11 4 Institutional and Implementation History of IBLI 17 5 Accurate and Effective Contract Design 25 6 Creating and Serving the IBLI Market 37 7 Evidence of IBLI Impact, Quality, and Uptake of IBLI 42 8 Enabling Sustainable Scaling 51 Epilogue: A New Round of IBLI Scaling Begins 61 References 63 use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core 1 Index-Based Livestock Insurance to Address Risk-Based Poverty Traps Revisiting the Challenge of Persistent Poverty When people’s living standards fall below a minimum absolute or relative threshold that societies deem necessary to safeguard the dignity of human persons, they are typically labeled as “poor.”Most cultures have sought to explain and reduce poverty, investing in the intrinsically normative topic with consider- able moral authority (Iliffe 1987; Lipton & Ravallion 1995). Generations of scholars have sought to explain patterns of poverty and to identify interventions that might help reduce its tragic hold on humankind (Ravallion 2016). A key empirical regularity throughout modern history is that poverty status varies more between places than within them, generating a large social science literature that documents and tries to explain spatially con- centrated poverty (Lipton & Ravallion 1995; Jalan & Ravallion 2002; Bloom et al. 2003; Ravallion 2016). Dating back at least to Adam Smith (1776), economists have typically seen poverty as the natural consequence of insufficient accumulation of productive capital, and/or insufficiently advanced technologies to generate a stream of income from that capital sufficient to sustain adequate consumption of essential goods and services. Most poverty analysis starts from that conceptualization, pointing to spatial patterns of low capital accumulation and anemic rates of adoption of modern technologies – both often arising due to market failures, especially in finance – to explain widespread, deep poverty. Others take a more radical view of poverty, which they see as the natural result of surplus extraction from the weak by the powerful (Watts 1983; Iliffe 1987). The poorest places on Earth are defined not only by the prevalence and depth of the poverty residents experience but also by the persistence of that poverty (Barrett & Swallow 2006). Poverty analysis has advanced considerably as longitudinal data on the same households and individuals have become more widespread (Carter & Barrett 2006; Barrett et al. 2016). The evolving poverty dynamics literature consistently finds that an identifiable subpopulation dispro- portionally suffers sustained deprivation that others never experience. Normative concerns about persistent poverty have long motivated research on “poverty traps,”which are defined as an absorbing state of persistent poverty. A large body of literature on poverty traps has focused on why low levels of capital accumulation and failure to adopt advanced technologies might be self- reinforcing equilibria (Azariadis & Stachurski 2005; Barrett & Swallow 2006; Bowles et al. 2006; Kraay &McKenzie 2014; Barrett et al. 2016). Poverty traps have typically been modeled as low-level equilibria that arise from coordination 1Escaping Poverty Traps and Unlocking Prosperity use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core (including market) failures with a focus on deterministic systems (Dasgupta 1997; Mookherjee & Ray 2002; Azariadis & Stachurski 2005; Ghatak 2015). Many poverty trap narratives and models have a Sisyphean character to them, in which people placed in impossible situations are doomed because desirable outcomes are simply unattainable. Rags-to-riches stories excite the popular imagination in part because they offer hope of escape from poverty traps, even if one’s odds of success are slim. The more recent literature on poverty traps dispenses with old, deterministic assumptions and focuses instead on the central role that risk plays in persistent poverty (Barrett et al. 2019). A deep, and influential literature documents the poor’s considerable exposure to risk and the limited market- or technology- based tools they have available to mitigate risk (Stiglitz 1974; Fafchamps 2003; Dercon 2004) The newer framing of risk-based poverty traps follows from the observation that another defining feature of places with high rates of deep, persistent poverty is disproportionate exposure to uninsured, catastrophic risk, often from multiple sources such as weather, markets, disease, and conflict. For example, across a range of societies at different stages of development, uninsured health shocks are consistently the single greatest cause of descent into persistent poverty (Krishna 2010), consistent with the literature that highlights how infectious disease risk exposure can trap individuals, or even entire communities, in long-term poverty (Bonds et al. 2010; Ngonghala et al. 2014; Ngonghala et al. 2017). This newer literature elevates the value of effective risk management to a status comparable to that of capital accumula- tion and improved technology adoption as central to enabling sustained improvements in living conditions (Barrett et al. 2019). In this view, deep, persistent poverty is not solely the consequence of bad initial conditions but rather of the combination of poor circumstances and excessive exposure to adverse shocks. Some poverty traps feature multiple equilibria wherein any individual1 may either escape poverty or collapse to an absorbing state of persistent poverty, depending only on their initial wealth and the sequence of shocks that they experience. Such a system generates what Ikegami et al. (2019) call “unneces- sary deprivation,” which occurs when individuals who have the capacity and means to be nonpoor are rendered poor by risk and shocks. Providing such individuals with better risk management tools should in principle reduce unnecessary deprivation and create substantial social and economic gains. Even for individuals who can in principle eventually escape poverty, risks and 1 Individual here can mean a single person, but also a single or even more complex family unit. 2 Development Economics use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core shocks lower their expected long-term well-being, slow their advance to improved living conditions, and generate costly transitory poverty. Better risk management tools can offer substantial social and economic gains for such people as well. Persistent Poverty in East Africa’s Arid and Semiarid Lands Index-based livestock insurance (IBLI) was conceived, launched, studied, and adapted within the context of the longstanding struggle to understand and reduce persistent poverty in a specific place: the arid and semiarid lands (ASALs) of East Africa.2 In many ways, this is an archetypal region, character- ized by widespread, deep, persistent poverty among populations routinely buffeted by a range of potentially catastrophic shocks. ASALs are the largest globally by area, covering roughly one-third of the Earth’s land surface, and host over one billion people, who commonly are pastoralists whose livelihoods predominantly rely on livestock production, often involving extensive grazing on communal lands, whereby seasonal movement in search of forage and water is important (de Leeuw et al. 2019). In relatively more humid ASAL areas, agropastoralists combine livestock with rainfed crop production (Nidumolu et al. 2022). Livestock are pastoralists’ main store of wealth; a productive asset that generates a plurality of community income and consumption goods, offers social status, and underpins many cultural rituals. Livestock are pastoralists’ main nonhuman productive asset and the production technologies involved in extensive grazing are few. In many ways, this makes pastoralist populations ideal for the study of stochastic poverty dynamics and the search to explain and unlock risk-based poverty traps. The decade-long, multidisciplinary Pastoral Risk Management (PARIMA) project set out to study such populations in the ASALs of northern Kenya and neighboring southern Ethiopia.3 The project identified the strong influence of drought risk on more salient food security and human health risks, which households perceive and attempt to manage (Smith et al. 2000; Barrett et al. 2001; Little et al. 2001; McPeak & Barrett 2001; Doss et al. 2008). A series of papers found that drought shocks led to considerable, avoidable human suffer- ing and that existing policy responses –mainly relief food shipped from distant countries – were slow to arrive and ineffective in mitigating the most serious 2 IBLI provides insurance against unusually low remote sensing (satellite) measures of forage availability that are strongly correlated with livestock productivity and mortality. Section 3 explains index insurance in greater depth. Sections 4 and beyond explain the particulars of IBLI in detail. 3 McPeak et al. (2011) summarize many findings of that research project. 3Escaping Poverty Traps and Unlocking Prosperity use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core human consequences that emerged from droughts (Mude et al. 2009; Nikulkov et al. 2016). Among the important findings from the PARIMA project, multiple data sets clearly established the existence of poverty traps in these communities (McPeak & Barrett 2001; Lybbert et al. 2004; Barrett et al. 2006; Santos & Barrett 2011). Multiple data sets consistently identified a threshold of 6–12 Tropical Livestock Units (TLUs),4 above which pastoralists could viably maintain large herds through transhumant or rotational grazing, and below which herd size collapsed to a low-level equilibrium of roughly one cow as it became infeasible to sustain the mobility required to sustain a larger herd (Lybbert et al. 2004; Barrett et al. 2006; Santos & Barrett 2011; Barrett & Santos 2014; Toth 2015). Moreover, the work established that uninsured catastrophic drought risk exposure is the primary cause of those poverty traps (Santos & Barrett 2019) and increases in the frequency of catastrophic drought due to climate change threaten to close off the high-level equilibrium options that remain, leading to system collapse (Barrett & Santos 2014). The drought risk-based poverty traps framing of the persistent poverty suffered by so many of the region’s pastoralists also helped explain why standard interventions often failed in the long-term. Post-drought restock- ing, for example, rarely restored herd sizes to the point where households regained the ability to migrate seasonally, and the frequency of drought meant that herds could rarely grow to a viable size before the next drought struck (Toth 2015; Santos & Barrett 2019). Meanwhile, emergency food aid and other transfers commonly failed to equip poor households to build assets, nor did they prevent collapse into destitution for formerly nonpoor pastoralists who had lost much of their herd due to a catastrophic drought, swelling the involuntarily sedentarized subpopulations in ASAL towns that increasingly overwhelmed underfunded social protection programs (Ikegami et al. 2019). New tools were clearly needed to help pastoralists manage catastrophic drought risk. IBLI was initially developed as a microinsurance scheme for pastoralists in an ASAL system characterized by multiple equilibrium poverty traps. However, its effectiveness as a drought risk management tool drew broader interest as a scalable risk management instrument applicable to individuals and households at the micro level, among governments at the macro scale, as well as a range of meso-scale organizations in between. 4 TLUs allow aggregation across livestock species based on body mass and nutrient intake requirements. For East Africa, ILRI deems one adult cow weighing 250 kg equivalent to 1.0 TLU, a camel equivalent to 1.4 TLUs, and sheep and goats each equivalent to 0.1 TLU. 4 Development Economics use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core Financial Innovation to Unlock Risk-Based Poverty Traps Conventional forms of social assistance, whether emergency food aid or cash transfers, are reactive as they respond only to the needs of those who have already collapsed into unnecessary deprivation. They do not address the under- lying structural causes that generate that collapse, nor necessarily help individ- uals maintain resilience to withstand and recover from shocks, or even advance economically in their wake. We set out to identify alternative interventions that might work better than those conventional mechanisms. Index-based risk transfer products were seen as a potential instrument for unlocking poverty traps, both by preventing descents into poverty and by inducing productivity-increasing investment and lending to facilitate such investment (Chantarat et al. 2007; Barnett et al. 2008; Chantarat et al. 2011; Chantarat et al. 2013; Chantarat et al. 2017). We designed an IBLI product with the intention to reduce negative impacts from drought risk and thereby to facilitate escapes from the poverty traps among the region’s residents. Similar objectives motivated parallel efforts elsewhere, as a range of agricultural index insurance products were designed in various settings to try to reduce risks associated with extreme weather events (as explained in greater detail in Section 3). A similarly named IBLI product emerged at roughly the same time in Mongolia, albeit with a different design and aimed at extreme weather events rather than droughts (Mahul & Skees 2007; Bertram-Huemmer & Kraehnert 2018). For a range of reasons explained in the coming sections, the East African IBLI product has generated greater – or at least better docu- mented – impacts and diffused more broadly than most other agricultural index insurance products, which have largely remained pilots or small-scale projects (Carter et al. 2017). Although social gains from financial risk management tools that disrupt poverty traps can be high (see Section 3), financial innovation needs to satisfy three key requirements. First, it must be high quality, so it reliably delivers payments when needed. Second, it must deliver assistance speedily during or near the onset of a shock to prevent individuals from losing or depleting their assets (e.g., through distress sales or abandonment with migration). Third, it must be trusted such that individuals will shift their behavior in advance of indemnity payments. These triple requirements of quality, speed, and trust informed our approach to developing IBLI. We hypothesized that these goals could be more easily attained with a pre-financed commercial contract than through a politically mediated transfer process that would always be subject to the vagaries of public sector budgets and politicians’ short-run interests. These challenges of quality, 5Escaping Poverty Traps and Unlocking Prosperity use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core speed, and trust required both excellent product design, aided by the emergence of new remote sensing techniques and technologies (see Section 5), as well as strong partnerships between researchers and operational agencies, both com- mercial and public, to continuously adapt the product and its outreach (see Sections 4, 6, and 8). As an experiment, IBLI had two distinctive characteristics. Firstly, launching IBLI required collaboration with commercial reinsurers, underwriters, retail agents, and a wide range of social and environmental scientists, as well as international donors, national and local governments, communities, and non- profit partners. The resulting partnerships brought together organizations and individuals with markedly different motivations to develop, adapt, and diffuse IBLI. This posedmajor management challenges but also broadened insights and ultimately buy-in to IBLI as the original design proved successful (Banerjee et al. 2019; Johnson et al. 2019). Secondly, IBLI needed rigorous impact evaluation. Did it really obviate the adverse, especially the catastrophic, impacts of drought? Did IBLI induce behavioral responses by pastoralist households and communities emboldened to risk scarce investible resources into economic advancement? Did it reduce descents into poverty, facilitate escapes from poverty traps, and generally boost welfare? And was it cost-effective in doing so, especially as compared to popular alternative investments, such as cash transfer programs? What pro- grammatic and design lessons could be learned to inform the scaling of risk management tools more broadly, beyond just the original IBLI product and the specific place where it originated? These are among the many questions that this Element will address. Before doing that, it is essential to understand the social and environmental setting of IBLI’s place-specific origins in tackling the challenge of risk-based poverty traps. 2 East African Pastoralism: Change and Variability The 300million or so Africans who inhabit ASALs face serious challenges. The compounding effects of natural and environmental factors – such as unpredict- able weather and spatially variable soil quality – policy and politics, and infrastructure make pastoralism in East Africa a risky endeavor. Droughts, the most common severe shock that hits ASALs, are often correlated with other shocks (e.g., conflict, disease, macroeconomic) and commonly cause cata- strophic loss of wealth and income for many people within affected communi- ties, frequently leading to humanitarian disasters. IBLI was designed to insure against drought, a “covariate shock” that affects large areas (distinct from 6 Development Economics use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core “idiosyncratic shocks” that strike just one or a few families at a time), and specifically for pastoralists in an area straddling the Ethiopia-Kenya border in East Africa. Climate is a key determinant of rangeland productivity, as vegetation growth follows rainfall amount, frequency, and duration (Coppock 1994; Coppock et al. 2017). Forage and water availability drive variability in ASAL livestock production. Pastoralism has evolved over centuries to manage the spatial and temporal variability of water and pasture. A key defining feature of East Africa’s ASALs is low and highly variable rainfall, with a bimodal seasonal pattern in most cases. These areas typically receive as little as 200 to 300 mm of rainfall annually, and rarely more than 600 mm (Williams & Funk 2011). Unpredictable rainfall patterns, combined with calcareous soils of low carbon and mineral content (Homewood 2008), result in low crop yield potential and render crop agriculture unreliable. Livelihoods therefore depend heavily on extensive grazing of cattle, camels, goats, and sheep. Livestock enable sporadic crop cultivation – mainly of maize – as the animals import essential soil nutrients and water by grazing elsewhere and then concentrating manure and urine within overnight enclosures that people can subsequently farm. During periods of good rains and availability of inputs, pastoralists often diversify into crop cultivation as a temporary relief and a means of supporting livestock, at least on stover (Catley et al. 2013). Even so, crop yields remain low and crop failures are commonplace. Because they are central to pastoralist livelihoods, livestock is equally central to pastoralists’ individual and community identities. Livestock ownership is not just a store of wealth but is equally a centerpiece of sociocultural activities and a leading source of social status. Livestock and their products are embedded in a variety of rituals and ceremonies, beginning with a person’s birth, and continuing through their circumcision, marriage, childbirth, and passing. Complex usufruct rules and agreements traditionally allowed pastoralists the flexibility they needed to ensure access to precious dry season reserves. However, this same flexibility also makes pastoralists vulnerable to land loss and exclusion from customary ranges (Homewood 2008). In recent decades, spatial expansion of towns and cultivated farmlands, as well as the gazetting of protected areas, have increased land fragmentation and increased exclusive uses for purposes other than grazing, reducing pastoralists’ ability to access crucial grazing and water reserves (Galvin et al. 2002; Munyao & Barrett 2007). Heavy grazing from restricted mobility can also degrade rangelands (Galvin et al. 2002) and threaten their sustainability. In addition, woody shrubs are expanding across rangelands because of both management practices and increases in carbon and nitrogen emissions (Galvin et al. 2002). Proliferation of woody 7Escaping Poverty Traps and Unlocking Prosperity use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core cover (or “bush encroachment”) has been compounded by governments’ (including Ethiopia’s) misunderstanding of the role of fire in mesic savanna ecosystems, resulting in ill-advised, strict fire bans that enable woody species to expand, degrading rangeland productivity and biodiversity (Johansson et al. 2021). The introduction of the fast-growing, non-native woody species Prosopis juliflora in ASAL environmental rehabilitation programs has likewise caused considerable damage in many rangelands, generating conflict between conservationists and pastoralists and lawsuits for damages caused by the Prosopis (Maundu et al. 2009). On top of increasingly restricted land and water access, droughts seem to have grown more frequent and severe in recent decades. Rainfall variability increases with aridity and climate change in this region (Overpeck & Udall 2020). The bimodal pattern in most of the Horn of Africa brings “short rains” from October to December and “long rains” from March to May. The “short” rains exhibit more interannual variability and are especially affected by El Niño Southern Oscillation (ENSO) events (Mutai & Ward 2000), with El Niño years bringing more precipitation and La Niña bringing less. Unfortunately, La Niña events are growing more frequent with global warming (Cai et al. 2015). Indian Ocean temperature anomalies can also influence precipitation in the absence of an ENSO event (Zhao & Cook 2021; Doi et al. 2022). Analysis of decadal rainfall trends in East Africa showed significant declines in long rains precipitation and increased unpredictability in the region between 1960 and 2009 (Williams & Funk 2011). Liebmann et al. (2014) found that the short rainy season has become wetter while the long rains are drier, but the significant increase in the short rains is compromised by strong year-to-year fluctuations. Ayugi et al. (2022) projected more frequent, longer, and stronger intensity droughts in this ASAL region in the future. These patterns – and the associated potential for system collapse (Barrett & Santos 2014) – underpin the need for regular revisiting of IBLI product design and pricing (see Section 5). In severe or prolonged droughts, livestock mortality rates increase sharply. Livestock population dynamics are determined by short-term losses during drought and longer-term trends in resource conditions, thus it can take several years for a herd to recover after a major drought and longer if several rainy seasons fail (as has been the case recently) and herd mobility is constrained. Significant droughts struck the region in 2011, 2014, 2016–2017, 2019, and 2021–2022, and the popular perception is droughts are becoming more severe in their impacts (Funk et al. 2015; Ayugi, Eresanya et al. 2022). Pastoral communities have long been marginalized by colonial and postco- lonial central governments. Pastoral systems are socioculturally alien to the foreign and highland populations that have long dominated Ethiopia, Kenya, 8 Development Economics use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core and other countries in the region. Few colonists or postcolonial leaders wanted to live in the harsher ASAL regions. Therefore, the infrastructure and institu- tions created to serve leaders’ (and their core constituencies’) own needs were concentrated outside the ASALs. Governments have often supported, explicitly or implicitly, the privatization of communal pastoralists grazing areas, gazetting protected areas or mining concessions, and even enclosures of rangelands previously held in common property with overlapping access rights among groups. Meanwhile, the central state has been notably absent in offering police protection, which contributes to a widespread sense of lawlessness in these ASALs (Catley & Iyasu 2010; Wild et al. 2019; Lind et al. 2020).5 Even when trying to help pastoralists, insufficient understanding of the rationale for and logic of pastoralism has often led to misguided development interventions, especially with respect to market development, rangeland rehabilitation, and early warning. Perhaps the most tangible material manifestation of pastoralists’ marginal- ization is their relative lack of infrastructure. They have fewer schools, fewer health facilities, limited electricity or telecommunications connectivity, insuffi- cient water, and sanitation facilities, and fewer maintained or all-season roads (McPeak et al. 2011). Indeed, the last stretch of the pan-African highway – which stretches from Egypt to South Africa – to get hard surface paving (e.g., asphalt or concrete) was in northern Kenya. The lack of roads, electricity, and so on, makes manufacturing and services difficult and hampers private invest- ments in the livestock sector, such as in slaughterhouses, canneries, dairy processing plants, and other value addition services. As ASAL populations live far from the major cities, this marginalization has been easy to ignore. This is changing in Kenya and Ethiopia, albeit slowly. Moreover, change is not always driven by communities’ best interests, as with improvements made in northern Kenya connected to (largely foreign-financed) hydrocarbons exploration and trade infrastructure (e.g., the Lamu Port, South Sudan, Ethiopia Transport) LAPSSET corridor through Isiolo and oil discovery in Turkana). Beginning in the late 2000s, however, mobile telephone service began in parts of southern Ethiopia and northern Kenya. Inexpensive phones and services offered unprecedented connectivity to distant markets and financial services such as mobile banking and digital payments. Communication is now much easier, and households can send and receive money, reducing two major impediments faced by these populations in prior years (McPeak et al. 2011). 5 As Wild et al. (2019) explain, pastoralists’ underrepresentation in national and global health statistics is another form of marginalization, especially because those statistics are used to direct public funds. 9Escaping Poverty Traps and Unlocking Prosperity use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core East Africa’s rangelands are also home to large and diverse wildlife popula- tions, attracting tourists from around the globe. Until the past few decades, wildlife coexisted with domesticated livestock, as both populations moved across the rangelands as seasons changed. In Kenya, two-thirds of the wildlife population are found in communal lands, groups, and private ranches (Western et al. 2009) rather than in nationally protected areas. The importance of working with pastoral communities to maintain wildlife populations is generally recog- nized (Reid et al. 2016; Western et al. 2020). However, community-based or other forms of inclusive tourism enterprises may not benefit all community members. Competition over land and other resources remains a key challenge, especially when protected areas exclude pastoral livestock, and as other devel- opment schemes and urbanization take up land and fence off mobility corridors (Munyao & Barrett 2007). Although there is considerable evidence that wildlife and livestock can be managed together, implementation of that model is not widespread, and too often wildlife conservationists – including large-scale private ranches that support ecotourism, conservancies, or similar services – and pastoralists engage in conflict over land tenure. Beyond contestation over land rights, herd movement has often induced inter-clan and inter-ethnic con- flict with sedentarized populations, environmental conservation agencies, or both (Bassi 2005). The overlap between wildlife and livestock populations also produces dis- ease interactions, although pastoralists have traditionally known when to move animals to avoid vector-borne diseases that increase with rains. Governments since the colonial era have been quick to impose quarantines on pastoral areas when infectious disease outbreaks occur, protecting the highland herds around the major cities at considerable cost to pastoralists and the traders who inter- mediate between the ASALs and the highlands (Barrett et al. 2003). Advocates of One Health approaches – which recognize that the health of people, animals, plants, and the environment are interdependent – argue for studying and man- aging healthy animals, people, and ecosystems in a more integrated fashion, especially emphasizing its benefits for pastoralists (Greter et al. 2014). The compounding effect of natural risks – poor soils, variable rainfall, frequent droughts, livestock, and human disease – and manmade ones arising from weak property rights in land and water (McCarthy et al. 2000), weak infrastructure, and political marginalization confront pastoralists with consider- able uninsured covariate risk from drought and disease among others. Governments and donors have historically mounted slow and insufficient responses to such disasters, mainly food aid shipments and limited post- drought restocking (Mude et al. 2009; Nikulkov et al. 2016; Santos & Barrett 2019). 10 Development Economics use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core The increasing frequency and severity of droughts, and the absence of adequate social protection response, have led to mass livestock mortality events that leave millions of pastoralists vulnerable to collapse into poverty traps (Lybbert et al. 2004; Barrett et al. 2006; Santos & Barrett 2019). IBLI was initially developed in this context, as a tool for pastoral populations to protect themselves from poverty traps that originate in drought risk. As we explain in subsequent sections, as IBLI spread to a wider range of countries, the changing context has necessitated adapting the product design (e.g., from predicted livestock mortality to forage scarcity) and delivery channels, as well as its scaling. 3 Index-Based Insurance for Pastoralist Regions Index insurance offers a prospective solution to poor rural communities’ expos- ure to the risk of extreme weather events (Barnett et al. 2008), inspiring a range of efforts to develop products well-suited to specific contexts (Carter et al. 2017; Jensen & Barrett 2017). This section explains the basic logic of index-based insurance in general and how this logic has been implemented for IBLI in East Africa’s ASALs specifically to address the covariate drought shocks pastoralists face, and to leverage markets to cost-effectively transfer the systemic drought risk characterizing the region. The potential of IBLI in these pastoralist regions extends far beyond simply exploiting a financial tool for solving a risk management failure. Conventional humanitarian aid and social protection programs, such as cash transfers and relief food distribution, commonly react to people falling into poverty. In targeting those that are already poor, such programs do not prevent people’s collapse into poverty nor dismantle the structural forces that generate chronic poverty in the first place. This section also explores the additional economic logic for index insurance for individuals and as a complement to existing social protection programming in these drought-prone regions characterized by poverty traps. Index Insurance and its Strengths and Weaknesses Insurance products can provide an adaptive, market-based solution to help manage risks. In advanced market economies, households and businesses typically seek – and sometimes are legally obliged to hold – insurance against catastrophic losses to prime income-earning assets such as life, health, or property (including automobile and home).6 Such insurance contracts are 6 Globally, life insurance accounts for roughly 45% of all premiums, while health insurance and property and casualty account for roughly one-quarter each (Binder et al. 2021). The insurance industry is built around insuring assets, not annual income flows from assets. Hence the need for 11Escaping Poverty Traps and Unlocking Prosperity use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core traditionally designed as indemnity insurance, in other words as contracts that reimburse policyholders in the event of a verifiable loss they incur. In the context of livestock insurance, examples include several of the plans available in the United States through the United States Department of Agriculture’s Risk Management Agency. In some settings, conventional indemnity insurance designs may not be com- mercially feasible because of incentive problems associated with moral hazard and adverse selection7 and high transaction costs to monitor policyholders’ behavior and verify their loss claims. This is especially true in places like East Africa’s ASALs, where most of the population lives in remote locations and where their limited wealth restricts the sums insured. As such, the fixed costs of information verification make it nearly impossible to profitably offer conven- tional contracts. Hazell (1992) offers several striking examples of conventional loss-adjusted contracts where the insurance provider cannot cost-effectively verify losses, with national insurance programs from the 1980s paying out two to five times the premiums collected, a financially unsustainable design. Index insurance products can fill the gap left by this market failure for conventional insurance contracts.8 Index insurance employs a cheap-to- measure “index” that correlates with individual losses, but that cannot be meaningfully influenced by any party to the contract. For example, a suitable index could be a river’s water level to approximate a household’s flood-related damages. Index insurance can thereby avoid moral hazard and adverse selection problems because loss verification is independent of the behavior and type of the insured. Index insurance can also significantly reduce transaction costs to generate risk profiles, set appropriate premium rates, and verify losses by using an index available at low cost in near-real time. In the case of IBLI, the index is based on remote sensing (satellite) measures of forage availability that are strongly correlated with livestock productivity and mortality (see Section 5). In the ASAL context, index insurance obviates the asymmetric information and costly loss verification problems that render conventional indemnity insurance infeasible, opening the door to offering commercial insurance to low-wealth households in remote locations. Despite the benefits of index insurance, it has several weaknesses. First, the use of an index that is only correlated with, but not identical to, individual government subsidization in order for crop, unemployment or other forms of term-specific income insurance to be viable. 7 Adverse selection occurs when clients purchase insurance that is offered at premium rates that are set using estimates of the client’s risk that are lower than they actually face. Moral hazard arises when having insurance induces behavioral change, in particular that the client engages in riskier behavior because they no longer bear the full cost of all potential adverse outcomes. 8 See Carter et al. (2017) for a review. 12 Development Economics use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core losses, also opens the door to “basis risk,” including both “false negatives” – a pastoralist who suffered a drought-related loss is not indemnified because the index failed to signal it – and “false positives” – a pastoralist who has not had losses is indemnified as if they had. False positives raise the premium cost of insurance (Elabed et al. 2013; Carter & Chiu 2018a). Jensen et al. (2016) evaluated the first IBLI index, which was used from 2010 until 2015, and found that it covered only 31 percent of households’ herd mortality risk, with the remainder lost to index imperfections that relate to differences between household-specific and area-average rates of livestock mortality. The index was revised in 2015 to allow payouts to take place earlier, before drought impacts have been fully realized. While Jensen et al. (2019) found that the new index correlated well with covariate livestock mortality observations, it has not been evaluated comprehensively for basis risk since then (see Section 5). Second, compared to conventional indemnity-based insurance, index insur- ance products are also relatively complex financial instruments from a policyholder perspective. Not only do they require an understanding of basic insurance mechanics, financial planning, and trust in the insurance pro- vider, but add the complexity of understanding and accepting an index which, in the case of IBLI, is observed from space, and is subject to the mentioned basis risk. Low financial literacy among pastoralists in the ASALs may limit index insurance demand (Patt et al. 2009), although evidence suggests that an accurate understanding of IBLI contract terms has only a limited effect on demand (Takahashi et al. 2016; Jensen et al. 2018). Finally, although index insurance sharply reduces underwriters’ costs of claim verification, the sales and indemnity distribution costs of an active insurance distribution network nonetheless remain high in remote rural areas, driving up premium rates. Data from one IBLI underwriter in Kenya show that for every United States dollar (USD) collected in IBLI premium, it cost on average USD 1.26 in operations and USD 1.76 in payouts, that is, about USD 3 in total to administer the policy (Lung et al. 2021). As many of these are fixed costs, this underscores the importance ofmarket development efforts to get to a commercially viable scale (Section 6). Economic Logic for Micro-level Index Insurance in Pastoralist Areas Index insurance can help resolve conventional insurance market failures, espe- cially if care is taken with product design and quality control (Carter et al. 2017; Jensen & Barrett 2017). Products like IBLI that aim to ensure productive assets may be even more viable than index insurance products that aim to insure 13Escaping Poverty Traps and Unlocking Prosperity use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core annual income realizations, consistent with the observation that most insurance policies globally insure assets, not income streams. This underlines the logic for index insurance; it can resolve an important financial market failure faced by poor households in rural areas. Potentially the most important added value of index insurance in contexts like East Africa’s ASALs comes from the role it can play in the presence of a risk- based poverty trap. In the rest of this section, we consider the economic case for even imperfect index insurance as a social protection tool to alter poverty dynamics in pastoral regions. Uninsured catastrophic drought risk exposure is the core mechanism that drives pastoralists into poverty traps (Lybbert et al. 2004; Barrett et al. 2006; Santos & Barrett 2019). Against this backdrop, IBLI was introduced in the Marsabit district of northern Kenya in January 2010 and the Borana plateau of southern Ethiopia in August 2012. When a subset of the authors of this Element approached potential funders to support the design and piloting of IBLI, we hypothesized that IBLI would offer a higher benefit–cost ratio and would result in lower long-term social protection expenditures than the usual mix of food aid and regular cash transfers targeted at the already-poor. In simple terms, we argued that a USD 15 annual insurance subsidy for vulnerable households would prove cheaper than letting the vulnerable slip into chronic poverty where they would become eligible for a USD 15 per-month cash transfer. That intuition has been developed more formally in a sequence of papers (Carter & Ikegami 2009; Ikegami et al. 2019; Janzen et al. 2021) that establish that index insurance can indeed reduce the total cost of social protection through two key mechanisms. The first is a “vulnerability reduction effect.” Insurance can protect households’ assets against catastrophic losses and maintain their economic viability at a relatively low cost, reducing the risk that they become chronically poor and require ongoing social protection expenditures. The second is an “investment incentive effect.” Insurance enhances households’ incentive to prudentially invest more in productive assets, making it less likely that they will require social protection assistance in the future. Those studies find that small herd sizes – those below the 6–12 TLU threshold identified before – whether initially or following a shock, can trap chronic poverty households who would otherwise grow their herds and not be poor, generating an “unnecessarily deprived” subpopulation due to some com- bination of low initial livestock wealth, misfortune, or both. IBLI was designed to change these poverty dynamics. It provides a safety net to the nonpoor who suffer drought-related herd mortality shocks that might otherwise cast them into unnecessary deprivation in the longer term. At the same time, IBLI can induce more investment by initially poor households by reducing the risk that they lose 14 Development Economics use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core that investment in the next drought. This generates a “paradox of social protec- tion,” reflecting the dynamic trade-offs that arise in a world of risk and poverty traps (Ikegami et al. 2019). The social protection paradox arises when concen- trating exclusively on the most destitute and ignoring the vulnerable, near- or barely poor leads to worse outcomes eventually for the poorest. The reason is that one needs to invest in preventing shocks, like droughts, from casting people unnecessarily into destitution else the ultra-poor population grows and over- whelms limited humanitarian budgets and ultimately harms the poorest relative to what could have been achieved by balancing humanitarian assistance with effective risk management (Ikegami et al. 2019). Therefore, besides offering an important risk management instrument in settings characterized by poverty traps, IBLI can also provide a cost-effective means to address vulnerability and the structural forces that generate chronic poverty. Most of those gains come from the vulnerability reduction effect. If in addition, the insurance is subsidized using a loosely targeted program, long-term poverty falls further, primarily due to the investment incentive effect, which leads some previously poor households to escape poverty. These results based on empirical data and micro-level simulations have helped stimulate demand for IBLI also at the macro scale. Economic Logic for Macro-level Index Insurance IBLI was developed and rolled out as a micro-level insurance product, that is, one sold directly to individual pastoralists. But index insurance can also be used at the macro level where it is purchased by national or sub-national governments, or by nongovernmental development or humanitarian organ- izations (Fava et al. 2021; Lung et al. 2021). In settings characterized by poverty traps, the logic of the preceding section can make index insurance an attractive policy instrument for combatting catastrophic risk, ensuring social protection in the face of shocks like droughts, and inducing private investment. Two subcategories of IBLI macro-level programs have been implemented to date. One is IBLI as a sovereign risk insurance program where governments purchase the policy and receive payouts which they commit to deploy based on a pre-agreed response plan to mitigate the impact of the insured risk. In 2021 African Risk Capacity (ARC) Ltd., a specialized agency of the African Union, added a pastoral component to the portfolio of products it offers African governments prioritizing coverage for pastoral regions, with a similar design as IBLI. The second is a “modified macro product,” which is basically the micro-scale IBLI aggregated into bulk purchases of the policy by institutions on 15Escaping Poverty Traps and Unlocking Prosperity use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core behalf of individual households who directly receive any indemnity payout if the index triggers. This approach was piloted by the Government of Kenya via the Kenya Livestock Insurance Program (KLIP) and by the World Food Programme (WFP) in the Satellite Index Insurance for Pastoralists in Ethiopia (SIIPE) program. Sections 4 and 5 provide further details on these programs. Macro-level programs can help overcome the weaknesses of index insur- ance outlined earlier. First, the ARC-type macro product eliminates the idiosyncratic risk component of basis risk for the policy holder (e.g., a government) because individual household-specific risks cancel each other out. Modified macro products, by enrolling more households, can help support informal insurance networks that manage idiosyncratic risk and basis risk within communities (Takahashi et al. 2019). Second, with the government as the sole policyholder/purchaser, many challenges with respect to financial literacy and trust can be overcome at a lower cost. ARC also provides comprehensive capacity-building services to clients, which is done much more easily for such a centralized macro product than for a spatially dispersed micro program. Third, macro and modified macro products require less product distribution infrastructure than micro-level programs, being focused on a single policyholder. Costs for onward distri- bution to shock-affected individuals remain, however, and can be significant, for example, in the form of household targeting and registration needs (Fava et al. 2021). Macro-level insurance programs, including IBLI, can also prove worthwhile to governments and nongovernmental organizations (NGOs) from a financial management perspective (Barrett & Maxwell 2007). When stochastic events (like droughts in pastoralist regions) create stochastic budgetary liabilities for governments and NGOs, insurance can offer a more effective means to pre- arrange the needed response funding compared to other budgetary tools such as reallocations, international borrowing, and fundraising appeals (Clarke et al. 2017; Carter et al. 2021). The economic logic of index insurance as a response to insurance market failures in low-income agrarian settings has motivated a wide range of donor- and government-funded interventions throughout the world over the past ten to twenty years (Carter et al. 2017). The added benefit of asset insurance in settings characterized by poverty traps makes IBLI especially compelling, both as a micro-scale product targeted at individual purchasers and as a macro-level policy instrument for governments or NGOs. However, much depends on key details around product design, distribution, and an enabling policy framework, which we discuss in subsequent sections. 16 Development Economics use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core 4 Institutional and Implementation History of IBLI From its conception to its modern-day scale, IBLI has progressed from a commercial pilot in one Kenyan county with a single retail sales channel, to a market with several products sold across three countries through multiple sales channels with support from a variety of actors and institutional arrange- ments. This progression has also moved IBLI beyond the direct influence of the original research and implementation partners and has ushered in a range of changes that illustrate the opportunities and risks that come with scaling a successful pilot. In this section, the main milestones of the IBLI journey are illustrated together with the conceptual pillars supporting IBLI’s operational implementation model. We highlight the systematic (and unusual) integration of a demand-responsive scientific research arm with evidence-backed and partner- led market development, which has been critical to its success and provides critical background for Sections 5–8. Piloting The IBLI program originated in 2007 as a research collaboration between the International Livestock Research Institute (ILRI), Cornell University, and the University of California – Davis with the objective of studying whether insur- ance couldmitigate the negative consequences of droughts for pastoralists in the region. After several years of research, product design, and stakeholder engage- ment, an index insurance policy was developed for Marsabit County, Kenya. Developing a new insurance market in a remote county with little exposure to insurance required considerable investments and innovative institutional arrangements. To be successful, IBLI needed cost-effective, efficient, and trustworthy channels for providing extension services, collecting insurance premiums, and disbursing payouts to insured pastoralists (Matsaert et al. 2011). The resulting marketing arrangement included a single local underwriter (UAP Insurance) supported by a global reinsurer (SwissRe) and Equity Insurance Agency (EIA) the insurance agency subsidiary of Equity Bank, one of Kenya’s fastest growing Banks at the time. To help the implementing partners recoup their initial investments in developing a new product whose timeline to commercial viability was not guaranteed, ILRI signed an agreement with EIA and UAP that gave them exclusive rights to sell the IBLI product for three years. IBLI was first launched in 2010 by EIA and UAP as a purely commercial microinsurance product sold through a network of insurance agents directly to individuals. Clients could purchase insurance coverage for camels, cattle, sheep, and goats. Coverage rates for each animal type were originally set to broadly reflect their market value and there was no minimum or maximum 17Escaping Poverty Traps and Unlocking Prosperity use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core coverage rate set, irrespective of herd size or composition. Insurance policies provided coverage for twelve months and payouts were made either through bank accounts, mobile money accounts, or in person by cash or check. After several sales windows, evidence that IBLI coverage was having posi- tive impacts on buyers generated interest, additional investments by donors and insurance firms, and pressure to expand IBLI beyond Marsabit County. This expansion initially proved challenging because the product used a livestock mortality index (see Section 5), which had been parameterized for Marsabit using a unique dataset of historic livestock losses that was only available in a few select areas in the region (Chantarat et al. 2013). The demand for geographic expansion, combined with fruitful collaborations with remote sens- ing experts, spurred the development of a new index that tracked relative local forage conditions, rather than predicted livestock losses, and which could be parameterized using existing global datasets, effectively allowing IBLI policies to be developed for any region. The downside to that design innovation was the index was effectively decoupled from prospective purchasers’ direct losses, raising new questions about product quality. While stakeholders were asking for geographic expansion, several factors, including the considerable costs of marketing, sales, and distribution, along with the monopoly granted to the exclusive insurance provider, resulted in several missed sales seasons by EIA in Marsabit. That experience underscored that implementation processes were as important as product quality to ensure pastor- alists had new, effective drought risk management options. Those missed sales seasons precipitated an adjustment in institutional arrangements. In 2012, the exclusivity agreement with the EIA and UAP Insurance was canceled, paving the way for new commercial partners and product innovation. One such innovation was the development by Takaful Insurance of Africa (TIA) of an Islamic Sharia- compliant version of IBLI to meet the needs of the region’s sizable Muslim population. Commercial partners also began to experiment with partnering with NGOs and local government agencies to reduce supply chain costs and increase demand (Mburu et al. 2015). Throughout this period, ILRI worked with donors to support public–private partnerships allocated public funds to subsidize product development and extension, and with technology firms to develop more cost- effective channels for customer education and last-mile product delivery. Micro-scale (and Growing Pains) Between 2012 and 2016, the IBLI market grew to include three new insurance firms and several new reinsurance arrangements, all while scaling outward to a total of seven arid and semiarid counties in Kenya (Johnson et al. 2019) and 18 Development Economics use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core into the Borana Zone of Ethiopia. While developing policies for the new regions required parameterization of insurance policy features through a collaborative process between pastoralists, insurance firms, and researchers, the process was relatively straightforward. At the same time, it became clear that developing low-cost and effective extension and delivery channels was more challenging. IBLI products were completely new to the pastoralists, and insurance agents had the heavy burden of not only explain- ing the concepts of commercial insurance but also the subtleties of the index product. Also, most local insurance firms had no experience selling index insurance, or even agricultural insurance, nor had they ever worked directly with pastoral populations in remote rangelands. Unsurprisingly, IBLI’s first five years were plagued with supply-side issues, including missed sales seasons, poorly trained agents, and uninformed clients, as insurance companies worked to develop these new markets and related infrastructure. As IBLI scaled in Kenya there was considerable churn in the insurance market. The original insurance company stopped selling IBLI, two new insurance companies entered the market, and then one subsequently exited but has since reentered. These changes created gaps in product availability, inconsistent framing of the product, and changes to insurance agents and information channels. Such inconsistencies undercut the desired image of stability, transparency, and security for insurance policies and the firms behind them. Pastoralists’ demand for IBLI was low and variable in this volatile period. Section 6 further discusses the difficult period from 2010 to 2015 in Kenya. The IBLI product itself also evolved during this time. In 2012, IBLI policies transitioned from insuring against average livestock losses to insuring against local relative forage scarcity. In 2015, the policies in Kenya went through another large shift, from a product that made payments after a drought (asset replacement policy) to one that made payments during the drought (asset protection policy) (see Section 5). In principle, this shift increased the value IBLI offered clients (Jensen et al. 2019) and has driven much of the subsequent discourse on anticipatory climate risk financing. But the added value of receiv- ing indemnity payments when animals are stressed by drought but can still survive if provided supplemental feed, veterinary services, and/or water depends on the availability of those goods and services for purchase using IBLI payouts. The limited markets for livestock services in East Africa’s ASALs may call that value addition into question. Aided by meso-scale purchases (discussed in the next section), the commercial sector expanded in the original markets. In Kenya alone, IBLI policies were commercially available for over 220,000 km2 of rangelands by 2020. But pastoral- ists’ rate of individual purchases of the commercial IBLI product remained modest 19Escaping Poverty Traps and Unlocking Prosperity use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core (Figure 1). Between 2010 and 2020, pastoralists purchased fewer than 50,000 policies, the vast majority covering two TLUs or less, insuring a cumulative value of over USD 10 million, and with renewal rates in subsequent years consistently less than 50 percent (Lung et al. 2021). At no point was more than five percent of the human population or two percent of livestock in regions with active IBLI availability insured through private IBLI purchases.9 It is not clear to what extent modest individual purchase levels reflect weak demand for drought insurance generally, issues related to this specific product (e.g., product quality, price, timing of premium payments), or a continuation of the supply-side obstacles faced during the first five years. Figure 1 IBLI coverage rates of the human (long dash) and livestock (short dash) populations in active IBLI regions, both read against the left-hand axis. The right-hand axis and solid line indicate the cumulative value of the total sum insured. Notes: Figures are for active regions in Kenya and the Borana Zone of Ethiopia only. In 2010, the active region included an area of 63,000 km 2 and a total population of less than 0.3 million individuals. By 2020, the active region included 406,000 km2 and about 6.5 million people. The estimates of the ratio of population covered assumes 5.5 members per household. The livestock estimates do not include camels. The figure does not include insurance coverage through KLIP or SIIPE. 9 By comparison, livestock owners in eastern Kenya spent about USD 10 per animal to vaccinate roughly 16 percent of adult cattle against East Coast Fever, a disease that is responsible for an estimated one million cattle deaths per year (McLeod & Kristjanson 1999; Marsh et al. 2016). 20 Development Economics use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core Meso- and Macro-scale Growth From 2010 to 2015, IBLI had only been a microinsurance product sold directly to individual pastoralists by local insurers. In 2015, the Government of Kenya added IBLI to its social protection programming by launching the KLIP, motivated in part by the logic described in Section 3. The KLIP program purchases insurance coverage for five TLU on behalf of beneficiary households, who, ideally, are targeted vulnerable households that fall just above the wealth threshold for eligibility for the Hunger Safety Net Programme (HSNP), a cash transfer program targeting the poorest households in the dryland counties of Kenya. This approach of purchasing (or heavily subsidizing) insurance for a targeted group is commonly referred to as meso-scale insurance, to distinguish it from individual purchases of insurance (micro-scale) or institutional or government purchases of insurance for themselves (macro-scale). Figure 2 illustrates the timeline of IBLI’s evolution in these three scales. KLIP initially purchased insurance on behalf of 5,000 households in two counties in northern Kenya. By 2017 it had grown to pay USD 2.4 million in premiums annually to provide coverage to 18,000 households each year for an annual total sum insured of USD 12.6 million (Fava et al. 2021). That is nearly equal to the total cumulative insured value through individual micro-scale IBLI Scale ‘10 ‘11 ‘12 ‘13 ‘14 ‘15 ‘16 ‘17 ‘18 ‘19 ‘20 ‘21 ‘22 ‘23 Micro IBLI launched in Marsabit Kenya in 2010 and grew to include several insurance firms and to provide coverage in eight ASAL counties of Kenya. The micro product continues to be sold across northern Kenya. In 2012, IBLI launched across the Borana Zone of Ethiopia and continues to be sold there by Oromia Insurance Company (OIC). Meso The Government of Kenya supports KLIP, which purchases IBLI on behalf of targeted pastoralists in eight ASAL counties. WFP implements SIIPE in Ethiopia, which provides conditional insurance transfers and has grown to cover 5,000 households. CST launches IBLI in Dassenech Woreda, Ethiopia. ICRC pilots IBLI in Meyumuluke Woreda, Ethiopia. WFP pilots IBLI in Zambia. World Bank’s DRIVE project subsidizes IBL Iacross ASALsin Kenya, Ethiopia, Somalia, and Djibouti. Macro ARC offers a rangeland customization to its sovereign product, which is based on the IBLI logic, and has been purchased by Burkina Faso, Chad, Mauritania, Niger, Senegal, Somalia, and Sudan. Figure 2 Timeline of IBLI scaling. 21Escaping Poverty Traps and Unlocking Prosperity use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core purchases of IBLI between 2010 and 2020. In part due to financial challenges associated with COVID-19, the KLIP program lapsed in 2022 and did not resume as we wrote this in April 2023. In 2018, theWorld Food Programme (WFP) and the Regional Government of the Somali Region, Ethiopia, jointly launched a meso program like KLIP called Satellite Index Insurance for Pastoralists in Ethiopia (SIIPE) (Frölich et al. 2019). As of 2021, SIIPE provided conditional, fully subsidized insurance for a limited amount of coverage, underwritten by a coalition of local private insurance firms and reinsured through international markets, to 28,300 pastoral households in the Somali region (WFP 2021). WFP has also piloted a similar scheme in Zambia (WFP 2022). The SIIPE program offers an example of an alternative approach to IBLI provision. Unlike comparable collaborations in northern Kenya and southern Ethiopia, the IBLI-ILRI team had only a short-term engagement with the WFP- SIIPE team, focused on product design and capacity development. WFP then led the implementation activities including the extension and sales activities, which were previously either left to insurance firms or ILRI intervention, by adding responsibilities to its existing field staff. Three key learning points are worth highlighting. First, once the tools and processes are developed and in place users like WFP may lead such efforts with minimal backstopping from technical partners. Second, there can be large advantages to using existing field staff for the last-mile distribution processes for insurance. The marginal costs of additional extension and sales activities for field staff that are already operating in the communities are small compared to the costs of onboarding and training new insurance agents, and they may already have relationships with community members that can support their sales activities. Third, implementation divorced from technical monitoring and evaluation runs some risks as regards product quality assurance. This latter issue has become increasingly salient as IBLI scales and in the absence of effective regulation requiring a credible signal of quality (see Section 8). Other local and international organizations have also started using IBLI in their resilience-building operations and social protection programs by subsidiz- ing insurance premiums for drought-vulnerable pastoralists. In 2020, a joint entity of three institutions (the Catholic Agency for Overseas Development [CAFOD], the Scottish Catholic International Aid Fund [SCIAF], and Trόcaire) known as CST worked in Ethiopia with ILRI and a local insurer – Oromia Insurance Company (OIC) – to develop an IBLI policy for the Dassenech woreda (administrative district) of South Omo. The collaborating partners provide a 70 percent premium subsidy to pastoralists in the region. The International Committee of the Red Cross (ICRC) has partnered with OIC and 22 Development Economics use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core ILRI to use IBLI to support long-term displaced pastoral populations in the East Hararghe Zone of Ethiopia. In January 2021, ICRC started offering an 80 per- cent premium subsidy to residents of the Meyumuluke woreda as part of its livelihood and resilience-building programming. Meeting these new project- level objectives and targeted interventions requires adaptation in how insurance products are designed, roles are allocated among stakeholders, and products are sold. The public use of private insurance mechanisms garnered considerable interest from several governments, especially those with large pastoralist popu- lations. In 2019, delegates from the Intergovernmental Authority on Development (IGAD) region met in Addis Ababa at the “High-Level Ministerial Policy Roundtable and TechnicalWorkshop” to discuss the potential for regional collaboration and coordination between countries as they devel- oped their own IBLI programs. This resulted in several donors commissioning regional feasibility studies in the IGAD region (Lung et al. 2021) and separately in the Sahel (Thebaud 2016; Fava et al. 2018). IBLI is also a major element of the World Bank-funded multicountry De-risking, Inclusion and Value Enhancement of Pastoral Economies in the Horn of Africa (DRIVE) project that was launched in 2022 (World Bank Group [WBG] 2022). The various discussions also highlighted the importance of the underlying regulatory envir- onment and complementary risk management tools to address distinct risk layers (see Section 8). Figure 3 shows IBLI’s expansion from the initial pilot in Kenya through to 2022. While KLIP, SIIPE, and DRIVE are all examples of meso-scale programs, the IBLI contract design has also been employed by Africa Risk Capacity Limited (ARC Ltd) in a rangeland customization of its sovereign (macro) insurance product that it offers across the Sahel and East Africa.10 The largest share of IBLI coverage has thus come through coordinated, bulk purchases under KLIP, SIIPE, and similar macro- or meso-scale programs. The geographical and vertical expansion of the IBLI agenda, from a micro- oriented pilot in Marsabit with a single insurance firm to supporting several insurance firms and collaborating partners operating at micro-, meso- and macro- scales across multiple countries, required continuous adaptation of IBLI to accommodate unique characteristics and objectives of varying stakeholders (pastoralists, state, and non-state actors), institutions (finance, governance, etc.), 10 ARC Ltd is a financial affiliate of the ARC group, a specialized agency of the African Union, established in 2012 to help African governments improve their capacities to better plan and effectively respond to extreme weather events. ARC Ltd, was founded in 2014 to provide index- based insurance focused on climate related disasters to provide ARC Group with a concrete instrument to advance its mission. 23Escaping Poverty Traps and Unlocking Prosperity use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core and infrastructure while striving to maintain quality standards. Throughout the development of the IBLI agenda over almost fifteen years, four key component areas have been central to the IBLI modus operandi since the program’s inception. Those four components are 1. Accurate and effective contract design: Continuous efforts for increasing precision and value of the policies for pastoralists while supporting sustain- able scale. See Section 5. 2. Creating and serving the IBLI market: Developing low-cost, effective methods for client and stakeholder awareness, educating for requisite cap- acities, and product service delivery. See Section 6. 3. Evidence of IBLI impact, quality, and uptake: Rigorously evaluate IBLI’s impacts on households and its broader societal value and disseminate the resulting evidence. See Section 7. 4. Policy and institutional infrastructure: Supporting design of an enabling policy environment to facilitate appropriate public–private partnership (PPP) infrastructure for the delivery of a sustainable program. See Section 8. Figure 3 The diffusion of IBLI and IBLI-like products in Africa. 24 Development Economics use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core While the weight of attention among these components has shifted over the years – largely in response to specific bottlenecks, or opportunities, encountered at various points along the IBLI journey – the symbiotic integration of these four elements has always been critical to uncover the optimal value from IBLI and guiding its trajectory to market and scale. The next several sections discuss each of these components in turn. 5 Accurate and Effective Contract Design IBLI’s contract design evolved over time, ultimately pioneering early trigger mechanisms to provide monetary support that can prevent livestock from dying and evolved hand-in-hand with the evolution and scaling of the product and program. This section reviews the key milestones in IBLI’s design, taking into consideration both the technical development of the remotely sensed drought indicator used for the IBLI index and the insurance design framework. The Original IBLI Design IBLI’s initial objective was to insure pastoralists against drought-related livestock losses, which were identified as the main risk to their welfare and livelihoods. The initial IBLI product was based on an index that estimated area- averaged livestock mortality rates at the end of an insurance season and provided payouts when the estimated average losses were greater than a pre- specified threshold, with the intention that the payouts could be used to “replace” lost livestock (Chantarat et al. 2013). This “asset replacement” contract was developed and validated using a statistical relationship between longitudinal observations of household-level herd mortality and a deviation of the satellite-derived Normalized Difference Vegetation Index (NDVI) from the long-term mean when tracked from the start of the rainy season until the end of the following dry season. A coarse-resolution (~8 km) satellite NDVI product – based on data from the Advanced Very High Resolution Radiometer (AVHRR) onboard the US National Oceanic and Atmospheric Administration (NOAA) satellites – was selected as the most suitable predictor of drought-induced mortality because its long-term NDVI time series capture seasonal and interannual variations in rangeland vegetation health and abundance that are associated with spatiotem- poral weather variability. While weather parameters, such as rainfall, are also a basis of multiple index insurance initiatives (Leblois & Quirion 2013), weather station coverage in Africa is generally sparse, and existing station data are often not easily accessible, making it hard to assess the accuracy of satellite-derived rainfall products. Moreover, summarizing rainfall amounts for 25Escaping Poverty Traps and Unlocking Prosperity use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core a season does not account for within-season rainfall distribution that is import- ant for vegetation development. NDVI can thus be a more direct indication of forage availability. Further, the NDVI data were freely available in near real time with nearly thirty years of continuous historical observations (i.e., AVHRR time series date back to 1981). The asset replacement contract was designed to cover the average risk for a covariate region – an insurance unit – over two consecutive insurance seasons, with an insurance season defined as a combination of the rainy season and the following dry season. Clients could purchase coverage for camels, cattle, sheep, or goats, which were then aggre- gated into TLU, and insurance payouts were made by multiplying the cost of replacing one TLU with the predicted area-averaged losses (Chantarat et al. 2013). While the original mortality index functioned successfully, three main draw- backs emerged as interest in IBLI continued to grow. First, the quality of the statistical relationship between NDVI and livestock mortality was heavily dependent on the quality and availability of the longitudinal household-level livestock mortality data used to calibrate the model, leading to potentially high basis risk, especially if these mortality data were sparse or inaccurate. Second, the lack of robust ground data for designing the mortality contract was a major limiting factor for the geographic expansion of the coverage.11 Third, clients and other stakeholders indicated a strong preference for a product that paid out prior to livestock loss with the goal of financing coping mechanisms to safe- guard livestock and avert massive wealth loss. The first two of these drawbacks – the need for long panels of livestock loss data and the sensitivity of product quality to errors in those household data – were addressed in 2012 when a new contract was developed that used seasonal NDVI anomalies directly as an index of forage scarcity.12 This new contract did not rely on livestock mortality data, but rather on the well-established relation- ship between NDVI and the green biomass production of rangelands (Fava & Vrieling 2021). Abstracting from livestock mortality to forage scarcity was possible because, for most extensive pastoral systems, forage availability is 11 The original IBLI product was designed for Marsabit District using rich, monthly household survey data collected by a government program (described in Mude et al. 2009). Those predic- tions were validated out-of-sample using two years of quarterly household survey data from the same district collected by the PARIMA project (described in McPeak et al. 2011). A household survey data series including high frequency, longitudinal, livestock mortality is rare. We are not aware of any similar data series from pastoral areas. 12 Degradation of NOAA-17 AVHRR data in 2011 prompted a transition of IBLI data source to the Moderate Resolution Imaging Spectroradiometer (MODIS) sensor on the Terra platform. MODIS data were available only since 2000, but Vrieling et al. (2014) demonstrated the possibility to extend the AVHRR-based index with the MODIS-based index, enabling reliable continuity of the original index time series. 26 Development Economics use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core a fundamental determinant of livestock survival, as alternative feed resources are largely unavailable or unaffordable. The global availability of NDVI data and the independence from ground datasets allowed expanding IBLI geo- graphic scope and prompted the launch of the product in several new areas. The third drawback was addressed in 2015 with the development of the “asset protection” contract which modified the previous contract by anticipating the payout timing. The idea of this contract was that the insurance payouts could help pastoralists protect their livestock before they died. Whereas livestock losses principally occur in or directly following the dry season, those losses are the result of inadequate forage growth during the wet season due to below- normal rainfall. Wet-season NDVI could thus be used as an indicator of wet- season forage accumulation and therefore coming (dry-season) forage scarcity. The new early indicator of forage accumulation, coupled with the introduction of electronic payments, preceded the devastating impacts of the drought, thus providing an earlier basis for payout, and perhaps for wealth preservation (Vrieling et al. 2016; Fava & Vrieling 2021). Design of the Asset Protection Contract The asset protection contract covers the risk of a significant deficit of forage growth during the rainy season, which leads to insufficient forage to feed the livestock during the subsequent dry months. The accumulated wet season forage production is gradually depleted during the dry season through decom- position and by livestock and wildlife grazing. When droughts strike, less forage accumulates during the wet season, leading to forage deficits that cannot support livestock nutrition for the duration of the dry season. The result is that livestock die from starvation and/or become more vulnerable to fatigue, dis- eases, predators, and other risk events, such as heavy rains or floods at the beginning of the next wet season, unless households spend additional resources on inputs (e.g., forage, water, relocation, veterinary services). An asset protec- tion contract can, in principle, enable households to use indemnity payments to purchase those inputs and prevent their animals from perishing. However, the effectiveness of the indemnities for herd protection depends on whether existing active markets in those inputs can respond to a surge in demand due to indem- nity payments. The NDVI processing chain to calculate the IBLI forage scarcity index for the asset protection contract includes three main steps (Vrieling et al. 2016). First, the NDVI data are spatially aggregated by taking the area-average NDVI per insurance unit for each (ten-day) NDVI composite. The insurance units are defined by a combination of operational criteria (i.e., administrative boundaries 27Escaping Poverty Traps and Unlocking Prosperity use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core or natural boundaries like rivers) and local knowledge of seasonal herd mobility patterns, ethnic boundaries to traditional grazing ranges, and local agroecology. The approach followed for IBLI starts from the lowest level of mapped admin- istrative units and works with local communities and institutions to combine or adapt these to define meaningful and clearly delineated insurance units, mask- ing out non-rangeland areas such as impervious surfaces, and large bodies of water (Chelanga et al. 2017). Second, NDVI time series are temporally aggregated during the season to derive a seasonal index. Defining the start and end period of aggregation can be guided by expert knowledge of rainfall/vegetation seasonality or by analysis of the satellite-derived temporal NDVI profiles (Vrieling et al. 2016). Finally, the aggregated seasonal NDVI is normalized to obtain an index that indicates how the seasonal forage compares to season- and unit-specific forage conditions during the past fifteen to twenty years (i.e., the full length of the data time series). There are several approaches for normalizing the aggregated season NDVI values, for example, z-scoring (subtract mean and divide by standard deviation), linear scaling between the minimum and maximum historic values (i.e., the vegetation condition index [VCI]), or percentile calculation. When a pre-defined index threshold value is reached,13 payouts are made proportion- ally to the severity of the forage deficit. The indemnity is calculated as a fraction of the total sum insured, corresponding to the estimated cost of keeping one TLU alive. Asset protection contracts are currently provided by several private firms and programs in Eastern and Southern Africa (see Fava et al. 2020; Lung et al. 2021; Section 4 for an overview). While the backbone of the design is the same for all the products, differences in the purpose among the various drought risk finan- cing schemes (i.e., microinsurance, modified macro social protection, sover- eign-level insurance) have led to several customizations of the parameters and to adaptations of the design and risk layering approaches. Innovating IBLI Product Design Progress in Earth observation technologies and applications creates new oppor- tunities to support drought index-insurance (Benami & Carter 2021; Fava & Vrieling 2021; Vroege et al. 2021), while anticipatory risk financing is increas- ingly promoted as a key part of climate adaptation strategies in low-income economies (Weingärtner & Wilkinson 2019). While the most significant 13 For IBLI, typically the trigger (i.e., the threshold index value below which the index triggers payouts) has been set as the 20th percentile of the index empirical distribution function. The exit (i.e., the threshold index value below which the maximum payout is triggered) has been set with different approaches over time. 28 Development Economics use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781009558280 Downloaded from https://www.cambridge.org/core. IP address: 86.183.149.231, on 06 Sep 2024 at 07:21:45, subject to the Cambridge Core terms of https://www.cambridge.org/core/terms https://doi.org/10.1017/9781009558280 https://www.cambridge.org/core transition in the IBLI design was from predicted herd mortality to forage scarcity contracts and from asset replacement to asset protection contracts, the index design has evolved continuously in response to research findings and stakeholder feedback (Fava & Vrieling 2021). This section summarizes the key challenges, lessons learned, and future opportunities, both in terms of the technologies supporting index design and in terms of the broader insurance product design framework. Advances in the Biophysical Index IBLI’s forage scarcity index is a measure of relative seasonal vegetation activity, and as such provides an indication of reduced forage development in specific seasons due to drought (Fava & Vrieling 2021). Nonetheless, many alternative drought indices exist (West et al. 2019) derived from precipitation, soil moisture, or evapotranspiration data products. In recent years, the accuracy of such products has improved, in part, due to sensor improvements (Vroege et al. 2021).While the link between these products and forage availability may be less direct, they could potentially benefit IBLI in several ways. First, given that green vegetation abundance is not solely a function of water availability and can be influenced also by non-palatable green vegetation, products that accurately describe different aspects of the water cycle may be used to calibrate the forage index to single out drought-induced reductions of forage (Enenkel et al. 2019). Second, because of the time lag between drought stress and its effects on vegetation (Udelhoven et al. 2009), these products may allow for earlier identification of drought, possibly facilitating earlier payments. In