Policy Research Working Paper 11283 Climate and Social Sustainability in Fragility, Conflict, and Violence Contexts Jose Antonio Cuesta Leiva Connor Huff Social Policy Global Department A verified reproducibility package for this paper is January 2026 available at http://reproducibility.worldbank.org, click here for direct access. Policy Research Working Paper 11283 Abstract Climate change is widely recognized as a driver of violent classifications, recent social sustainability frameworks, and conflict, but its broader social effects remain less under- controls for population and geography. The results reveal stood. Ignoring these dimensions risks a vicious cycle where strong correlations—not causation—between climate climate policies might undermine socially just adaptation. events and contexts of fragility, conflict, and violence. Cli- Evidence is still limited on how climate shocks influence mate impacts are most pronounced in both national and political participation, trust, or migration. This paper helps subnational conflict settings. The study also finds robust fill that gap by examining links between climate change, links between fragility, conflict, and violence and low levels conflict, and social sustainability, with a focus on inclusion, of social sustainability, reflecting its role as both a driver and resilience, cohesion, and legitimacy. Using secondary data consequence of conflict. Some dimensions—such as violent from 2019–24, the study applies simple correlation-based events and insecurity—appear weaker in areas most affected methods to test three hypotheses on the nature, severity, by climate shocks. Two of the hypotheses are supported, and composition of these associations. The analysis com- and one remains inconclusive. bines multiple climate impact measures, new conflict This paper is a product of the Social Policy Global Department. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at jcuesta@worldbank.org. A verified reproducibility package for this paper is available at http://reproducibility.worldbank. org, click here for direct access. RESEA CY LI R CH PO TRANSPARENT ANALYSIS S W R R E O KI P NG PA The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Climate and Social Sustainability in Fragility, Conflict, and Violence Contexts Jose Antonio Cuesta Leiva1 and Connor Huff2 (1) World Bank, Global Department of Social Development, jcuesta@worldbank.org (2) University of California Los Angeles, Department of Political Science, connorhuff@ucla.edu Authorized for distribution by Robin Mearns, Director, Gender, World Bank Group Key words: Climate change, conflict, fragility, violence, social sustainability JEL codes: Q01, Q56 Acknowledgements: This study contributes to the background research for the World Bank’s flagship “Peace and Social Dividends of Climate Action”. This work was partially funded by the State and Peacebuilding Fund Grant Number TF0C7472. The authors would like to thank Patrick Barron, Oli Brown, Jana El-Horr, Meltem Ikindji, Lindsey Jones, Amir Khouzam, Lucia Madrigal, Stanislava Mladenova, Miguel Purroy, Janna D. Tenzing for insightful feedback and recommendations. Remaining errors are entirely the responsibility of the authors. 1. Introduction Achieving the world's development goals requires joint progress in economic, environmental, and social sustainability—essentially, fostering economic growth and jobs, addressing climate change, and building peaceful societies. While the connections between economic and environmental sustainability were established decades ago (IUCN, UNEP & WWF, 1980), the link between climate change and social sustainability remains underexplored (Agliardi, Casari & Xepapadeas, 2019). This gap is partly due to the insufficient theoretical and empirical foundations of social sustainability (Barron et al., 2025). Ambiguous definitions of social sustainability and a lack of consensus on effective actions to promote it hinder global progress. In addition, the evidence of the connections of conflict with climate change is not free of empirical challenges. As a result of the lack of progress in understanding the dynamics connecting those forms of sustainability, major challenges persist, including stagnant economic growth, poverty, and inequality (Alvaredo et al., 2018, Milanovic, 2024), climate change (Stern, 2007), and rising conflict (World Bank, 2021). To address these challenges, this paper aims to fill key evidence gaps in the relationship between social and environmental sustainability, focusing specifically on unpacking the interconnections between climate change, fragility, and conflict. To clarify the concept of social sustainability, we adopt the recent definition provided by Barron et al. (2023) in their comprehensive conceptual and empirical framework: socially sustainable societies are those in which all community members feel included in the development process and believe that both they and future generations will benefit from it. These communities collaborate to overcome challenges, deliver public goods, and allocate scarce resources in ways that are perceived as legitimate and fair, ensuring long-term well-being. This definition highlights four critical dimensions of social sustainability: social cohesion, inclusion, resilience, and process legitimacy, each explored below. In exploring these dimensions’ interactions with conflict, it is important to consider first the existing research on climate change and conflict, recently characterized as an “exploding literature” (Burke et al., 2024a: 30). This literature has shown that climate change can influence the risk of violent conflict, including organized armed conflict, interpersonal violence, and even suicides (Burke et al., 2009, Baysan et al., 2019). However, analyses disagree on the extent of these influences. According to Mach et al. (2019: 193), studies "have repeatedly triggered dissenting perspectives... and failed to clarify areas of agreement and reasons for disagreement." Significant uncertainties remain about when and how climate change has caused conflict to date, as well as how future conflicts might impact climate risks. Additionally, the limitations of current methodological approaches raise concerns about the evidence base of effective policies reducing climate risks and enhancing human security both in the short and long term. Building on this background, this paper presents stylized facts on the connections between climate change and social sustainability, identifying commonalities and distinctions in climate change impacts across social sustainability components, and analyzing variations within fragile, conflict-affected, and violent (FCV) contexts. It uses data from the newly constructed Social Sustainability Global Database (Cuesta, Madrigal & Pecorari, 2024), refines definitions of fragility and conflict at subnational levels, and explores both conditioned and unconditioned 2 correlations among three key elements—climate change impacts (floods and droughts), social sustainability, and FCV—over the most recent period of 2019-2024. In doing so, the paper focuses on three sets of research questions, which we turn into research hypotheses that can be tested: Q1: How do the impacts of climate change vary between fragile, conflict-affected, and violent settings and more stable contexts? Which types of FCV situations experience the most severe climate impacts? Do these impacts differ by the type of climate-related risk? H1: The effects of climate change are more severe in contexts with higher levels of fragility, conflict and violence. The intensity and nature of these impacts vary depending on the type of context and the specific climate hazard. Q2: How strongly are poor social sustainability outcomes linked to fragile, conflict-affected, and violent settings? Which types of FCV contexts tend to experience the lowest levels of social sustainability? H2: Social sustainability tends to be lower in settings with more severe fragility, conflict and violence. Q3: Are areas with lower social sustainability more likely to experience the negative impacts of climate change? H3: Low social sustainability is not only common in more fragile and conflict-affected settings but also in areas facing more severe climate impacts. The analysis begins with a review of existing research, which has primarily focused on how climate change influences conflict. It then details the data and methods, followed by empirical findings and an assessment of their robustness. 2. Literature review: Unpacking the links across climate change, FCV and social sustainability: What do we know? Climate change undoubtedly poses a major threat to humanity and the planet (Ko et al., 2024). Between 2010 and 2020, people caught in climatic hazards were 15 times more likely to die than those in regions with very low vulnerability (IPCC, 2022). Evidence links it to environmental degradation (Gyau-Boakye, 2001), worsening economic inequality (Paavola, 2017), rising poverty (Das Gupta, 2014), the spread of infectious diseases (Van de Vuurest & Escobar, 2023), increasing hunger (Jaramillo et al., 2023), and declining health outcomes, including lower life expectancy (Roy, 2024) and higher mortality (Pottier et al., 2021). It is destroying crops, fueling wildfires, killing livestock, and forcing people to abandon traditional ways of life (Minale et al., 2024). Climate-induced displacement is accelerating, with large areas becoming uninhabitable (Stapleton et al., 2017) and tens of millions expected to be uprooted in the coming decades (Clement et al., 2021). Conflict on its own also poses another major threat to humanity, with widespread consequences. It hampers economic development (Gates et al., 2012), weakens public health (Ghobarah, Huth 3 & Russet, 2003), limits access to resources (Raleigh, 2011), and harms psychological well-being (Mollica et al., 1999). Conflict disrupts economies (Collier, 1999), deepens poverty and destabilizes livelihoods (Brück, d’Errico & Pietrelli, 2019). State repression further suppresses political engagement (Petrova & Rosvold, 2024), shapes dissent (Hatz, 2019), and erodes trust in government (Desposato et al., 2021). This threat has grown, not diminished: an average of 50 countries experienced armed conflict each year between 2012 and 2021—up from 34 in the previous decade (Burke et al., 2024a). Unsurprisingly, climate change and conflict reinforce each other, as conflicts weaken resilience to climate shocks, while resource scarcity fuels tensions. Many climate-vulnerable countries also endure prolonged armed conflict, with disasters and violence often hitting the same communities (Basheer & Elagib, 2024). Fragile states account for about 60% of climate-related disaster deaths (Peters & Budimir, 2016), as conflict diverts resources, damages infrastructure, and undermines disaster risk management (Walch, 2018). Extreme weather events inflict more lasting economic damage in fragile states, where they cause an average of 4 percent in cumulative GDP losses after three years, than in other countries, where the impact is closer to 1 percent (Jaramillo et al., 2023). Of the 39 countries in the World Bank’s FY24 List of Fragile and Conflict-affected Situations, 28 score among the lowest 50 countries on a global ranking of climate vulnerability (ND GAIN, n.d.). When quantifying and unpacking those mutually reinforcing links, evidence suggests that climate variability, hazards, and trends have influenced armed conflicts within countries over the last century. Summarizing experts’ estimates, Mach et al. (2019) report that climate change has contributed to 3%–20% of conflict risk to date. They also report that climate variability has significantly increased risk in about 5% of conflicts. Looking ahead, a 2°C rise above preindustrial levels is projected to raise conflict risk by 13%, climbing to 26% under a 4°C warming scenario (Mach et al., 2016). Burke et al. (2024a) project a similarly significant impact to that reported by Mach et al.’s studies, suggesting that extreme climate conditions will continue to elevate conflict risk. However, their impacts are smaller than those suggested by previous studies. Drawing on more recent studies with broader geographic coverage, finer temporal conflict data, and higher- resolution spatial data, they find that a one standard deviation increase in local temperature is linked to a 2.5% rise in intergroup conflict and a 1.9% rise in interpersonal violence—estimates that remain highly significant but are about three-quarters and a fifth smaller, respectively, than earlier findings. 1 With projected temperature increases of 2 to 4 standard deviations in coming decades—and absent effective adaptation—conflict risks could rise by 20%–40% for intergroup violence and 6%–12% for interpersonal violence. 1 Hsiang, Burke & Miguel (2013) and Burke, Hsiang & Miguel (2015) report that a normalized standard deviation increase (relative to the local distribution) in a climate variable (such as warmer temperature) leads to an 11% increase in intergroup conflict, and similarly to a 2.3% increase in interpersonal violence. Both of these effects are statistically significant at high levels of confidence. These results are consistent with Carleton et al.’s (2017) study using data on food security in Africa, which estimated that a one standard deviation increase in temperature will cause a 10.8% average increase in conflict incidents and 16.2% average increase in the rate of violent crime. 4 Research has also linked resource scarcity to violent conflict, with extreme temperatures and precipitation amplifying risks by reducing output (Homer-Dixon 1999). Climate change intensifies resource scarcity, heightening competition for land and water, particularly in fragile states (Mekonen & Berlie, 2021). In Ethiopia, increasing droughts have led to habitat loss, soil erosion, and biodiversity decline (Woodward, Perkins & Brownet, 2010). The 2015 drought left one in ten citizens food insecure as rainfall fell 50% below average (Kassegn, Ebrahim & Yildiz, 2021). Water-related conflicts have erupted among pastoralists in Kenya, Ethiopia, Somalia, Mali, and Mauritania (Mulugeta, 2011). Floods have a more immediate but complex conflict impact. While they rarely ignite new conflicts, they prolong existing ones by displacing populations and straining weak institutions (Ghimire & Ferreira, 2015). Flood-induced political unrest depends on governance, population pressures, and political exclusion (Ide, Kristensen & Henrikas, 2021). Others have found that while sudden temperature shifts show no clear link to conflict, prolonged cooling correlates with instability, suggesting societies adapt to moderate changes but struggle with sustained disruptions. This evidence aligns with findings on short-term climate shocks (Miguel, Satyanath & Sergenti, 2004) and long-term agricultural productivity (Iyigun, Mueller & Qian, 2024). The latter’s research also emphasizes that the strength and nature of the effect of climate change and conflict risks depend on economic, political, and institutional contexts. Climate shocks may have a weaker impact on conflict in areas where incomes are less climate- dependent or where stronger states can buffer economic losses—through safety nets, social protection, or military deterrence (Burke, Hsiang & Miguel, 2015). Still, the dynamics are far from straightforward. Their findings show that household wealth does little to mitigate the link between rising temperatures and civil conflict, whereas countries in Africa with higher GDP per capita exhibit notably lower conflict sensitivity to temperature shocks. This suggests that robust state capacity might play a more decisive role than individual-level economic incentives in reducing climate-driven conflict risks. Buhaug (2010) and Buhaug, Hegre & Strand (2010) argue that climate variability alone does not predict armed conflict, highlighting the important role that political and socio-economic factors play in shaping these linkages. Theisen, Gleditsch & Buhaug (2013) find mixed evidence, suggesting climate change’s impact is mediated by governance and institutional strength. Mobjörk, Krampe & Tarif (2020) argue that deteriorating livelihoods fuel grievances, migration can heighten tensions over scarce resources, armed groups exploit resource scarcity for tactical gain, but only when elites manipulate vulnerabilities, conflicts escalate. Ultimately, some view climate change as a key conflict driver, while others emphasize governance and socio-political conditions. Still, some studies stop short of identifying any dominant driver. Burke et al. (2024a) note that the expanding research base points to a wide range of relevant factors beyond climate— such as demographics, migration, infrastructure, politics, institutions, and even psychological and physiological responses. The mechanisms linking climate extremes to conflict appear highly context-specific, making it difficult to single out any one factor as the primary driver. As this remains a frontier area of research, no clear consensus has yet emerged. Climate-induced economic shocks also drive conflict, as rainfall and temperature fluctuations disrupt agriculture and deepen poverty. Miguel, Satyanath & Sergenti (2004) linked rainfall- induced economic shocks to armed conflicts in 41 African nations, while Burke et al. (2009) 5 projected a 54% rise in Africa’s civil war likelihood by 2030 due to rising temperatures. Historically, cooling periods reduced agricultural yields, triggering famine and war (Zhang et al., 2007). In China, 70%-80% of war peaks, dynastic transitions, and social unrest coincided with colder periods, largely due to declining agricultural productivity (Zhang et al., 2020). Similarly, in Eastern Europe, temperature fluctuations significantly influenced conflict, reflecting the region’s agricultural dependence (Lee et al., 2019). Climate change also fuels conflict through physiological and psychological stress, increasing discomfort, aggression, and violence (Koubi, 2019). Both extreme heat and cold elevate hostility, fostering violence (Anderson, 2001). This heightened aggression can trigger large-scale displacement, sparking conflicts in receiving areas (Warnecke, Tänzler & Vollmer, 2010). Climate change further exacerbates intragroup violence by intensifying resource scarcity— freshwater, arable land, forests, and fisheries—especially in overpopulated regions (Froese & Schilling, 2019). Extreme temperatures deplete essential survival resources, deepening tensions. Evidence on the social impacts of climate shocks remains inconclusive. Some studies link conflicts to lower trust but higher political participation (Mackay, Mavisakalyan & Tarverdi, 2024), while others argue that social vulnerabilities heighten conflict risks (Koubi, 2019) and trap nations in instability cycles (Buhaug & von Uexkull, 2021). Climate disasters exacerbate socio-economic hardships, fueling political violence (Theisen et al., 2013), though such conflicts often remain localized (Njiru, 2012). Crises test state legitimacy (Lupu & Peisakhin, 2017), but strong institutions and trust can prevent escalation (Petrova, 2022). As a result, the social link between climate change and military conflict remains contested (Salehyan, 2008). Climate’s influence on conflict remains weaker and more uncertain than non-environmental factors (Burke et al., 2009; Adger, 2014; Mach et al., 2019). This uncertainty stems from complex climate–conflict linkages, which depend on socioeconomic trajectories, global recessions, state capacity, ideological shifts, intergroup inequality (e.g., ethnic divisions), history of violence, and global governance. There is strong consensus that low socioeconomic development is a major predictor of intrastate conflict (Hegre & Sambanis, 2006), though debate remains on whether it fuels conflict through grievances or opportunities. Another source of uncertainty is how climate responses can create unintended consequences, affecting conflict risks (Adger, 2014). For instance, food export bans after climate-induced crop failures may destabilize other regions, while adaptation policies that benefit some groups over others could heighten tensions (Mach et al., 2019). Optimists argue that migration, technology, and institutional shifts mitigate risks (Nunn & Qian, 2011), while pessimists contend that accumulated climate shocks deplete food stocks and fuel instability (Bai & Kung, 2011). In reality, adaptation and intensification often coexist, making the net impact an empirical question. Far less research has examined how conflict shapes climate change. Evidence suggests, however, that conflict may exacerbate vulnerability to climate-related risks. Armed conflicts weaken institutional responses to disasters and reduce resilience to climate shocks (Barnett, 2006). In Nepal, prolonged conflict eroded disaster risk reduction frameworks (Jones, Oven & Wisner, 2016). Conflict-affected states experience higher disaster mortality rates—40% higher during 1960–2010 (Marktanner, Mienie & Noiset, 2014) and 34% higher during 1989–2018 (Caso, Hilhorst & Mena, 2023). In the Philippines, violence in Mindanao compounded the impact of Typhoon Haiyan, straining governance and humanitarian resources (Field, 2018). Conflict also 6 worsens climate-induced migration, escalating competition for scarce resources and fueling violence (Hoffmann, 2022). Resource scarcity further reduces the willingness of warring parties to negotiate (Mera, 2018). In conclusion, climate change amplifies resource scarcity, economic shocks, and migration, all of which heighten conflict risks. Scarcity of essential resources, like water and arable land, often escalates tensions, particularly in fragile states. Poor land and resource management, driven by environmental degradation, population growth, and climate change, fuels conflict and impedes resolution. However, climate change alone is an unlikely trigger conflict; sociopolitical factors such as governance, institutional strength, and social cohesion are key mediators. While some studies suggest climate shocks erode trust and increase political violence, the broader social impacts are debated. Notably, conflict can worsen vulnerability to climate change by weakening state responses and reducing resilience. Although these findings link climate to conflict and reveal key mechanisms, they fall short of testing the hypotheses needed to unpack these relationships. Despite the recent growth in literature, it remains focused on case studies and best practice examples, many of which are too localized to scale. The same cases are often repeated, even though each FCV context is unique. Research tends to emphasize English-speaking countries, neglecting successful conflict resolution, peaceful adaptation, and instances where conflict has not emerged despite serious climate impacts (van Schaik et al., 2019). Additionally, most studies focus on conflict, with fewer exploring fragility, social sustainability, interpersonal violence, and a broader range of climate risks. Moving forward, more evidence is needed on the role of social mediators in the climate-conflict dynamic. The interplay between social sustainability, climate change and conflict is mutually reinforcing, highlighting the need for comprehensive policies addressing both climate adaptation and conflict prevention in socially sustained ways, that is, in ways that reduce social vulnerabilities. This intersection raises urgent global justice and security concerns, calling for collective and coordinated action. Climate-induced conflicts disproportionately impact developing countries, despite their lower per capita emissions. Assistance to vulnerable nations in building resilience and adapting to climate impacts; investment in peacekeeping to mitigate climate-driven conflicts; and fostering international collaboration to protect populations from instability are among such types of collective international action needed. 3. Methods, data sources and indicators The discussion above highlights the need for stronger evidence to guide policy responses that can simultaneously reduce conflict and climate risks. While the links between climate change and conflict remain mixed, the complexity of these interactions calls for robust, cross-national studies that establish clearer patterns. Further research is needed to explore the role of social mediators—such as trust, participation, and resilience—in shaping how communities adapt to both climate change and conflict. Building on the lessons from our review, the paper advances the evidence base on the links between climate change, conflict, and social sustainability by (a) purposefully incorporating fragility into conflict and violence analysis, while addressing social 7 sustainability aspects beyond trust; (b) shifting focus from quantifying complex correlations to providing practical insights into their nature, examining them across conflict types and climatic risks; and (c) moving away from comparing multiple drivers to developing testable hypotheses on interlinkages, enhancing the basic understanding of these relationships. The study breaks down these relationships into clear research questions and testable hypotheses, using straightforward analysis to build on and expand existing knowledge about how climate affects conflict risks. Before proceeding, however, it is important to acknowledge the limitations of current approaches. Analytical challenges contribute to uncertainties about the role of climate in conflict, including its mechanisms, the conditions under which it emerges, and the implications of future climate change (Mach et al., 2019). Many studies focus on specific regions using time-series analysis but often overlook socio-economic and political factors or fail to differentiate between conflict types (Sakaguchi, Varughese & Auld, 2017). Cross-national panel regressions can address these gaps by incorporating broader factors and distinguishing conflict types, but they often focus on temperature or precipitation anomalies without isolating human- induced climate change (Ko et al., 2024). Attempts to bridge this gap with available data on greenhouse gas emissions face limitations due to missing data, particularly in conflict and humanitarian contexts. Additionally, capturing the interplay between violence, trust, and resilience is challenging, as these dynamics are non-linear and shaped by complex interactions (Barron et al., 2023; Pecorari & Cuesta, 2024). This is especially true when analyzing the spectrum of violence, ranging from interpersonal to intergroup violence (Baysan et al., 2019). Causal inference is more feasible for climatic variability than for slow-moving factors like socio- economic development, state capability, or intergroup inequality. Furthermore, climate-conflict literature has been less concerned with process and case study analyses to generate hypotheses for systematic testing. Most previous analyses have also concentrated on contexts where climate variability has led to conflict, rather than on resilient, cooperative, and peaceful outcomes, as highlighted in ethnographic studies. Literature reviews also face limitations—such as reconciling findings across diverse methods, focusing on average effects while overlooking contextual variation, and being subject to publication bias (Burke et al., 2024b). Moreover, while prior work shows that different types of conflict respond to climate events at varying spatial and temporal scales (Hsiang, Burke & Miguel, 2013), it remains unclear whether the geographic scope of a climate anomaly causally influences the likelihood of conflict (Burke et al., 2024a). Others argue that there is a tendency to underestimate the ability of human societies to adapt to changing circumstances (Brown & Nicolucci-Altman, 2022) while studies that consider climate change tend to be more likely to investigate its link to conflict, rather than the peaceful conditions many societies maintain despite climate vulnerabilities (Simangan et al., 2023). These methodological constraints limit the extent to which we can draw firm conclusions about current linkages or future implications using causal approaches. For this reason, the study adopts simpler correlation-based methods to identify and describe basic patterns, providing insights into the research questions and assessing the validity of the initial hypotheses. Furthermore, this paper relies on four core sources of data for assessing the relationships between social sustainability, climate change, and fragility, conflict, and violence. The first source of data is the Social Sustainability Global Database, SSGD (Cuesta, Madrigal & Pecorari, 8 2024). Following Barron et al. (2023), the database defines social sustainability as being comprised of four distinct components. These include social inclusion, social cohesion, resilience, and process legitimacy. For each of these four components, the SSGD includes a number of distinct indicators, each of which capture slightly different theoretical concepts. We selected two indicators for each pillar of social sustainability. Albeit somewhat arbitrary, this ensures that our measures capture unique contours of the pillar while also ensuring that there were as few missing observations as possible—making the empirical exercise meaningful. It is important to note that social sustainability—and especially resilience—is defined differently across various conceptual frameworks. In the Annexes, we examine alternative measures, including indices, to assess resilience and its impact on results. We draw from Cuesta, Madrigal & Pecorari (2024), who identify conceptually distinct dimensions within each pillar of social sustainability. For instance, for Social Inclusion, they identify access to markets, access to basic services, access to human capital, and the extent to which laws and regulations affect women in society as important subcomponents of inclusion. Each of these subcomponents has in turn unique indicators measuring the underlying concept. Thus, when selecting indicators, we chose questions from distinct areas within each of the four pillars. For Social Inclusion we chose indicators within the areas of access to markets and the distinct legal treatment of women in several domains such as workplace, mobility, safety, marriage or pensions, to cite some. Similarly, for Resilience, we selected indicators from the areas of food and financial insecurity; these two areas are key predictors of resilience when populations are exposed to shocks. 2 For Social Cohesion, we used widely recognized indicators of violent events and interpersonal trust. For Legitimacy, we selected rule of law and control of corruption—proxies for perceived fairness in authorities' decisions, consistent with Barron et al. 2023’s definition of legitimacy. Throughout our analyses, we focus on Wave 2 of the Social Sustainability Global Database, which creates a single measure for the temporal period between 2019-2022. Each of the unique indicators chosen is presented in Table 1. Table 1: Social sustainability core indicators, per component. Concept Indicator Observations Unit Period Source Social Inclusion The share of population 123 Country/Territory 2019-2022 SSGD who owns a bank account Social Inclusion Women, Business and the 190 Country/Territory 2019-2022 SSGD Law score Resilience % of population who have 66 Country/Territory 2019-2022 SSGD gone without enough food Resilience % of population that saves 62 Country/Territory 2019-2022 SSGD money Social Cohesion Number of violent events 197 Country/Territory 2019-2022 SSGD Social Cohesion % of population who think 78 Country/Territory 2019-2022 SSGD most people can be trusted Process Rule of law 203 Country/Territory 2019-2022 SSGD Legitimacy Process Control of corruption 203 Country/Territory 2019-2022 SSGD Legitimacy 2 In Annex D we explore an alternative measure of resilience, an index on resilience to climate change, from ND-GAIN. 9 Source: Cuesta, Madrigal & Pecorari (2024)’s SSGD. The second source of data is a list of countries characterized as being in states of Conflict, Violence, and Fragility. Throughout our analyses, we distinguish between national and subnational conflicts. We define national conflicts as those with widespread, high-intensity conflict affecting a significant proportion of their population. Following the World Bank’s FCS classification for inclusion as a conflict country, countries must meet two conflict-related criteria. First, countries must have over two battle related deaths per 100k of the population according to the ACLED dataset. Second, countries must have over one battle related death per 100k of the population according to the UCDP dataset. Additionally, we classify countries as having national conflict if they also have more than 20 percent of the country affected by conflict. 3 Finally, in order for countries to be classified as having national conflict, they must meet the above criteria for three of the past five years, spanning the 2019-2024 temporal window. Countries are instead classified as having subnational conflict if they meet the battle death related thresholds for UCDP and ACLED as discussed above and have less than 20 percent or fewer of the country’s area in conflict. This distinction in geographic coverage is intended to capture the fact that at times, the geography of conflict is concentrated in relatively remote regions of the country. This geographic distinction can pose unique development challenges. As with national conflict, for a country to be classified as subnational, it must meet the inclusion criteria for three of the past five years. Countries are classified as experiencing violence if they have had a homicide rate above 10 per 100,000 population in at least three of the past five years, using a hierarchical approach to avoid double counting. Finally, countries are classified as Fragile if they are included in the OECD State of Fragility List in either 2020 or 2022. In Annex A, we present the full list of countries that fall into each category. The third source of data is EM-DAT. EM-DAT contains information on the occurrence and impacts from over 26,000 mass disasters that occurred between 1900 and 2024. While there are a myriad of potential sources attempting to measure and assess the causes of climate change, relatively few provide comprehensive information on the consequences of climate disasters. EM- DAT's emphasis on measuring the human impacts of climate change provides a key source for our analyses. We focus on assessing the climate impacts of two types of events: droughts and floods. Each of these are pervasive problems throughout the world that are only expected to become more severe as climate change worsens. Further, restricting our analyses to these two impacts allows us to conduct a deeper analysis than would be possible were we to instead focus on a broader range of climate events. 4 For both floods and droughts, we use three distinct measures for assessing the impacts of climate events. To start, we simply count the number of unique flood and drought events over the 2019- 2024 period within each country on the SSGD list. Second, we count the number of estimated fatalities for floods and droughts for each country over the 2019-2024 temporal window. Finally, 3We calculate the share of the country affected by conflict at the ADMIN1 level. 4Further, discussions with climate experts suggest that alternative measures of climate impacts—such as extreme temperatures— are particularly prone to measurement error. 10 we count the number of individuals affected by droughts and floods over the same temporal window. Ultimately, this approach provides us with six distinct measures of the impacts of climate change by analyzing two types of events across three measures of severity (see Table 2). In analyzing correlations in subsequent sections, we standardize the number of events, deaths, and affected individuals per 100,000 inhabitants to prevent population size from biasing or distorting the results. Table 2: Summary of measures and data sources for climate impacts. Concept Indicator Observations Unit Period Source Flood impacts Number of floods 222 Country/Territory 2019-2024 EM-DAT Flood impacts Number of deaths from floods 222 Country/Territory 2019-2024 EM-DAT Flood impacts Number of individuals 222 Country/Territory 2019-2024 EM-DAT affected by floods Drought impacts Number of droughts 222 Country/Territory 2019-2024 EM-DAT Drought impacts Number of deaths from 222 Country/Territory 2019-2024 EM-DAT droughts Drought impacts Number of individuals 222 Country/Territory 2019-2024 EM-DAT affected by droughts Source: EMDAT Our resulting dataset is thus organized at the country level and focuses on measures from within the 2019-2024 temporal window. For each country, we have a classification for whether it is affected by conflict, fragility, and violence. We also know the frequency and severity of climate events, which can play a pivotal role in shaping the development challenges facing countries. Finally, we have information on the social development challenges faced by each country in our dataset. Taken together, these new data provide a unique opportunity for providing new insights into the links between social sustainability, conflict, and the impacts of climate change. Before proceeding, it will be helpful to note a few challenges posed by the structure of the data. Perhaps most obviously, there are at times a relatively small number of observations for some of the measures of social sustainability. This is most severe with measures for resilience, where we only have approximately 60 observations. Given this, readers should consider the totality of the evidence linking social sustainability with climate impacts and conflict, rather than focusing on any single result. Further, skeptical readers might have concerns that there is measurement error in our measures of battle related deaths or climate impacts. If, for example, more dangerous places have less news reporting because of conflict, then this could lead to systematic underestimations of the impacts of climate change. It is difficult to estimate the number of people affected by floods if it is difficult for reporters to access those regions in the first place. Throughout our analyses, we directly discuss how concerns about missingness might substantively affect our results. 11 4. Results: FCV and climate impacts How do the impacts of climate change vary between fragile, conflict-affected, and violent settings and more stable contexts? Which types of FCV situations experience the most severe climate impacts? Do these impacts differ by the type of climate-related risk? This section answers these research questions by looking at the correlations between climate change impacts and FCV contexts at both aggregated and disaggregated levels. To start, we can explore the differences in climate change impacts between FCV and non-FCV contexts. The average differences in the total number of climate events and the number of deaths and people affected by those events over the 2019-2024 period is depicted in Figure 1. The impact measures of occurrence and severity are adjusted for population, allowing comparisons that account for differences in population size and, indirectly, for the fact that larger countries tend to have larger populations. Annexes B and C present correlation estimates between FCV context and the impacts of droughts and floods, both with and without controls for population and territorial size, as well as correlations standardized per 100,000 inhabitants. The panels on the left correspond with floods, while the panels on the right correspond with droughts. Within each panel, the bar on the left presents the average numbers for non-FCV contexts, while the bars on the right correspond with FCV contexts. The results show that, once adjusted for population size, FCV and non-FCV contexts experience similar frequencies of floods, while droughts occur more often in non-FCV settings. See Figure 1. However, the impacts of both floods and droughts in terms of deaths and affected people per 100,000 inhabitants are more severe in FCV contexts, with differences that are statistically significant and consistent with H1. Notably, when the analysis is conducted without population adjustment, both the frequency and severity of droughts and floods appear greater in FCV contexts (see Annex D). The figure highlights three key differences in climate impacts between FCV and non-FCV contexts. First, the population-adjusted magnitude of deaths from floods and droughts is broadly similar across contexts, averaging around 0.6 deaths per 100,000 people. However, the gap in population-normalized deaths between FCV and non-FCV contexts is much larger for droughts—about ten times higher—than for floods, which are less than three times higher over the period 2019–2024. Second, these patterns hold when considering the population-adjusted number of people affected by droughts and floods. In FCV contexts over 2019–2024, droughts affected an average of 12,000 people per 100,000, compared with about 5,000 people per 100,000 affected by floods. In non-FCV contexts, droughts affected 4,000 per 100,000 and floods 2,000 per 100,000. This shows that, when adjusted for population, droughts impact more people than floods. Finally, the gap between FCV and non-FCV contexts is slightly larger for droughts than for floods. About twice as many people were affected by floods in FCV contexts compared with non-FCV contexts, whereas droughts affected roughly three times as many people in FCV 12 settings. Overall, the results indicate significant differences in the severity of climate impacts between FCV and non-FCV contexts, with the disparities particularly pronounced for droughts. 5 We now turn to exploring whether there are different types of climate impacts depending on whether states are faced with national conflict, subnational conflict, violence, or fragility. Figure 2 depicts the average differences in the total number of climate events, and the number of deaths and people affected by those events over the 2019-2024 period. Once again, those numbers are normalized to reflect population differences in the countries and territories analyzed. The panels on the left correspond with floods, while the panels on the right correspond with droughts. However, each panel now breaks down the impacts by FCV context. Figure 2 highlights a number of important differences in climate impacts across FCV contexts. First, the occurrence of floods per 100,000 people in violent contexts more than doubles the occurrence of floods in other FCV contexts, that is, contexts with national, subnational conflict and fragile settings. In the case of droughts, fragile contexts have a larger occurrence than other forms of FCV contexts. Regarding impacts, deaths from floods are concentrated in, first, subnational and, then, national conflict areas. Instead, deaths from droughts concentrate disproportionally on fragile contexts. When considering number of people affected per 100,000 people as a result of floods, fragile contexts join areas with national and subnational conflict as main contributors. This pattern is also observed among people affected by droughts (normalized by population). These patterns confirm that the impacts of climate change vary across types of FCV, with not a clear, dominating type of FCV showing larger (normalized) numbers of death and people affected. These findings support H1 in that the intensity and nature of climatic impacts vary depending on the type of context and the specific climate hazard. 5 When analyzing total, rather than population-adjusted, figures, several results change. Total deaths from floods and droughts in FCV contexts are no longer similar. On average, floods caused 267 deaths over 2019–2024, compared with roughly 32 deaths from droughts during the same period. Second, the pattern of more people being affected by floods than droughts (per 100,000 inhabitants) does not hold for total numbers. In FCV contexts, an average of 3,243,659 individuals were affected by droughts, compared with 1,945,250 by floods over the same period. Thus, while floods result in more deaths, droughts affect a larger number of people overall. Across both total and population-adjusted impacts, the difference between FCV and non-FCV contexts is consistently larger for droughts than for floods (see Annex D). These results underscore the importance of population size, as previously highlighted in Annexes B and C, while confirming marked disparities in climate impact severity between FCV and non-FCV contexts. 13 Figure 11: Differences in climate impacts by FCV and non-FCV settings, 2019-2024 Source: Authors’ estimates Ultimately, this section shows that the consequences of climate change have, and will likely continue, to hit FCV contexts particularly hard. This may reflect that natural hazards are more intense when they occur in FCV and, or that FCV countries have less ability to deal with them, hence experiencing more destructive impacts. However, there is variation in the types of impacts across contexts, with many of the most severe impacts being concentrated in locations enduring violent conflicts. This can create important challenges for development actors. At times, those most vulnerable to the impacts of climate change live in areas controlled by rebel forces. This means that development actors working to increase resilience in these communities are forced to navigate the local complexities of working in rebel-held territory. And as we explore in the next section, the urgency of development, and also the challenges associated with carrying it out, can potentially be exacerbated depending on the levels of social sustainability. 14 Figure 22: Differences in climate impacts by FCV setting, 2019-2024. Source: Authors’ estimates. 5. Results: Social sustainability and FCV How strongly are poor social sustainability outcomes linked to fragile, conflict-affected, and violent settings? Which types of FCV contexts tend to experience the lowest levels of social sustainability? To answer these questions, we look at the correlations between social sustainability impacts and FCV contexts at both aggregated and disaggregated levels. Figure 3 analyzes the relationships between FCV and non-FCV settings for each of the four different components of social sustainability. Each row focuses on a different component of social sustainability including social inclusion, resilience, social cohesion, and process legitimacy. Each panel within each row uses a different measure capturing a different conceptual 15 element of each component. For instance, the top row measures social inclusion leveraging indicators for the share of a population who has a bank account, and the average country-level index for Women, Business, and the Law. 6 Consistent with our second research hypothesis, across all four components of social sustainability the patterns are the same: social sustainability is worse in FCV contexts. FCV contexts are less inclusive, have lower levels of resilience, lower levels of social cohesion, and lower assessments of process legitimacy. For instance, as can be seen on Panel 1, our estimates suggest that only 50% of individuals in FCV settings have access to a bank account; by contrast, roughly 80% of individuals have access in non-FCV settings. Similarly, as demonstrated in Panels 3 and 4, those in FCV settings are much less likely to save money and have enough food. Individuals in FCV settings are also likely to be exposed to higher levels of violence (normalized by population) 7 and are less likely to say individuals can be trusted. While the percentage is relatively low in both non-FCV and FCV settings, only 14% of respondents in FCV settings say that most people can be trusted. The same patterns also recur for process legitimacy: individuals in FCV settings are less confident in the rule of law, and more likely to think that public power is exercised for private gain. 8 Overall, these findings support H2, indicating that social sustainability tends to be lower in FCV contexts. Figure 4 assesses the relationships between each of the four components of social sustainability and the type of FCV context. FCV contexts are once again broken down by whether they have national conflict, subnational conflict, fragility, or violence. As in Figure 3, each row corresponds to a different component of social sustainability. In contrast to the links between FCV contexts and climate impacts—where many of the impacts were concentrated within subnational and national conflict settings—social sustainability is commonly lower across all FCV contexts. For instance, as can be seen in the top row of Figure 4, all FCV contexts have lower levels of social inclusion than non-FCV contexts. Indeed, for five of the eight measures of social sustainability, all FCV contexts have lower levels of sustainability than non-FCV contexts. Interestingly, fragile settings persistently have some of the lowest levels of social sustainability across all measures except—naturally—for the number of violent events when adjusted for population size. Taken together, the evidence shows that social sustainability is lower in FCV contexts, confirming H2. These patterns are fairly uniform across FCV contexts, meaning that many of the social sustainability challenges facing development actors are similarly pervasive across all settings of fragility, conflict, and violence. The challenges make adaptation and mitigation work harder; they also increase the importance of ensuring that climate action and mitigation is done in socially sustainable ways. 6 The WBL score assesses how laws and regulations—such as those pertaining to mobility, workplace rights, pay, marriage, and parenthood, among other regulations—affect women’s economic opportunity in 190 economies. 7 This is also true when using the total number of violent events, that is, without normalizing by 100,000 people. Individuals in FCV settings continue to be more likely to be exposed to higher total levels of violence. See Annex D. 8 Control of corruption captures perceptions of the extent to which public power is exercised for private gain, including both petty and grand forms of corruption, as well as "capture" of the state by elites and private interests. 16 Figure 33: Differences in social sustainability by FCV and non-FCV settings, 2019-2024. Source: Authors’ estimates. 17 Figure 44: Differences in social sustainability by FCV context, 2019-2024. Source: Authors’ estimates. 6. Results: Social sustainability and climate impacts Are areas with lower social sustainability more likely to experience the negative impacts of climate change? In this section we answer this question by assessing the correlations between social sustainability and climate impacts. To do so, for each unique measure of social sustainability, we created a dummy variable for whether a given country was above the mean. If a country was above the mean for a given indicator, they are coded as a 1 while countries below 18 the mean are coded as 0. We then calculated the average differences between places that were high, as opposed to low, for each of our six different measures of climate impacts. The results for these differences are presented in Figure 5. Each point represents the average difference between places with high versus low levels for each of our indicators for social sustainability, while the solid lines denote 95% confidence intervals. The top row focuses on the differences in counts of climate events, while the second and third rows demonstrate the differences in deaths and individuals affected by those events. The first column corresponds with floods and the second corresponds with droughts. Each color denotes a different measure for social sustainability. There are several interesting findings that emerge from the analysis. First, evidence for our third hypothesis—that low social sustainability is common in FCV contexts and in areas experiencing more severe climate impacts—is somewhat mixed The relationships between climate impacts and social sustainability are less stark than those between countries’ FCV status and both social sustainability and climate change. The overwhelming bulk of the estimated differences are not statistically significant. Second, the dimension of social sustainability that is most strongly correlated with climate impacts are the number of violent events normalized by population size. Given the strong correlation between countries’ FCV status and climate impacts, the latter finding is unsurprising. After all, violence is a core component of what makes a country an FCV context. 9 This section demonstrates that climate change is likely to impact areas with higher levels of violent events, while population-adjusted climate impacts are similarly severe in areas with both high and low levels of social sustainability in dimensions other than violence. Consequently, H3 cannot be accepted across most dimensions of social sustainability, though it holds for violent events. This should not be interpreted as implying that designing and implementing climate interventions is easier because impacts are expected to be comparable across areas with different levels of social sustainability. Moreover, the strong link between FCV contexts and climate threats suggests that social sustainability may continue to decline in these areas, compounding challenges for development projects. These findings underscore the urgency of interventions that simultaneously address the drivers of conflict and climate change in socially sustainable ways, including prioritizing actions that effectively reduce violence. 9 While most results remain consistent when using total, rather than population-adjusted, climate impacts, the correlation between total climate impacts and process legitimacy becomes notably strong. For five of the six measures of climate impacts, the differences between high and low levels of both rule of law and control of corruption are statistically significant at the 0.1 level (see Annex D). 19 Figure 55: Difference in climate impacts by high versus low social sustainability, 2019-2024. Source: Authors’ estimates. 7. Conclusions This paper demonstrated several important patterns in the relationships between climate impacts, FCV contexts, and social sustainability. First, the paper showed that there were strong links between FCV contexts and the impacts of climate events. These impacts were most heavily concentrated in conflict settings, including those with both national and subnational conflict. Second, the paper showed that there were strong links between FCV contexts and social sustainability. This is unsurprising, given that we know that low levels of social sustainability can serve as a cause and effect of conflict. Finally, the paper found that certain dimensions of social sustainability—particularly levels of violence and insecurity—tend to worsen in areas more severely affected by climate events. These results confirm our first two research hypotheses, while leaving the third inconclusive. 20 This paper contributes to the growing evidence linking climate change and conflict by highlighting a gap in prior research: the role of social sustainability. While it is broadly observed that climate threats are most acute in conflict-prone areas, existing studies rarely consider how factors like legitimacy, inclusion, and cohesion intersect climate and conflict dynamics. Our analysis addresses this gap by showing that the effects of climate on conflict and violence vary significantly depending on the types of conflict, as well as fragility and other aspects of social sustainability beyond resilience. This is important, because any development interventions targeted either toward conflict, climate, or both, must have a deep understanding of how the contextual features of the setting affect whether the project succeeds or fails. A key implication for the success of World Bank projects is the need for ongoing monitoring and a stronger evidence base. Our findings underscore the importance of disaggregating outcome measures—especially those related to climate resilience, levels of fragility, conflict, and violence. Relying on aggregate results risks overlooking the distinct impacts that emerge in different FCV contexts. Understanding the social dimension of the climate–conflict link has other direct policy implications beyond project monitoring. Strengthening state capacity in climate governance— alongside areas like fiscal management, digital infrastructure, or security—not only improves climate resilience but can also reduce conflict risks. Effective state responses to climate-related grievances can help prevent instability. For instance, in contexts of weak governance, "green grabbing"—the appropriation of land for conservation, carbon offsets, or renewable energy—can fuel conflict, especially when it intersects with longstanding grievances over resource extraction (Hunsberger, Corbera & Vaddhanaphuti, 2019; Nordensvard, Gilbertson & Marinelli, 2016; Myers et al., 2021) or revalues land in ways that alienate local communities (Fairhead, Leach & Scoones, 2012). Our findings also have implications for climate adaptation strategies, both within and outside the World Bank. While national income and wealth are well-established factors mediating the effects of climate and conflict, long-term adaptation must go further—beyond jobs and income generation—to address deeper social vulnerabilities. National Adaptation Plans, for instance, can promote inclusive planning and tackle root causes of exclusion and mistrust (Church & Crawford, 2020; Sacchetto, Stern & Taylor, 2020). In northern Darfur’s Wadi el Ku region, a project successfully brought together previously hostile communities to build water retention dams, pairing infrastructure investment with support for community-led water management systems (Nielsen et al., 2022). Ignoring the complex social dimensions of climate change impacts can create a vicious cycle, where climate actions inadvertently worsen existing challenges. Addressing these dimensions effectively—while tackling both climate and conflict risks—lays the foundation for what is often termed "transformative adaptation" (Fook, 2015) or "socially just climate adaptation" (Malloy & Ashcraft, 2020). To ensure climate actions do no harm and, where possible, support social sustainability and peace, more evidence is urgently needed on the intersections between conflict, climate, and social sustainability, along with insights into what approaches have worked or failed. 21 In Annex C we show the robustness of these findings. We show that the results generally hold when removing controls for population and territorial size. We also show that when we use alternative measures for components of social sustainability—such as ND-GAIN—we again observe results consistent with our hypothesis testing using single indicators (See Annex E). There are several methodological and data limitations that are important to acknowledge. First, there are important measurement challenges for assessing all three outcomes. If there are biases or difficulties in news reporting, this could affect estimates of the number of individuals affected by climate and conflict. Second, there are potential issues with missing data. Some measures of social sustainability have relatively few observations. We sought to assuage this concern by relying on multiple measures and indicators. Finally, there are issues with temporal coverage across the datasets. As noted in the Annexes, we varied these and showed that our results are largely robust to small tweaks to the time periods under consideration. Finally, there are several areas of future research that naturally follow from this paper. First, in the future, scholars and practitioners can explore the links between climate impacts, social sustainability, and fragility, conflict, and violence by conducting within-conflict analyses. There has been an interesting wave of micro-level research exploring the links between conflict and climate; this work could be strengthened by considering how climate and conflict affect social sustainability, and how this will in turn affect the unique development challenges in these contexts. Second, future research can continue to theoretically and empirically refine the measures of social sustainability. Finally, future research can continue to explore the links beyond conflict, social sustainability, and climate impacts expanding beyond droughts and floods. References Adger, W. N., N. W. Arnell, and E. L. Tompkins. 2014. “Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects.” In Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects, edited by C. B. Field, V. R. Barros, D. J. Dokken, et al., 755–791. Cambridge: Cambridge University Press. Agliardi, E., M. Casari, and A. Xepapadeas. 2019. “Introduction: Special Issue on the Economics of Climate Change and Sustainability (Part A).” Environment and Development Economics 24: 555–559. Anderson, C. A. 2001. “Heat and Violence: Current Directions in Psychological Science.” Current Directions in Psychological Science 10 (1): 33–38. Alvaredo, F., L. Chancel, T. Piketty, E. Saez, and G. Zucman. 2018. “The Elephant Curve of Global Inequality and Growth.” AEA Papers and Proceedings 108 (May): 103–108. Bai, Y., and J. K. Kung. 2011. “Climate Shocks and Sino-Nomadic Conflict.” Review of Economics and Statistics 93 (3): 970–981. 22 Barnett, J. 2006. “Climate Change, Insecurity, and Injustice.” In Fairness in Adaptation to Climate Change, edited by W. Adger, J. Paavola, S. Huq, and M. Mace, 115–130. Cambridge: MIT Press. Barron, P., L. Cord, J. Cuesta, S. A. Espinoza, G. Larson, and M. Woolcock. 2025. “Social Sustainability and the Development Process: What is it, Why Does it Matter, and How Can it be Enhanced?” Oxford Development Studies. Barron, P., L. Cord, J. Cuesta, S. A. Espinoza, G. Larson, and M. Woolcock. 2023. Social Sustainability in Development: Meeting the Challenges of the 21st Century. Washington, DC: World Bank. Basheer, M., and N. A. Elagib. 2024. “Armed Conflict as a Catalyst for Increasing Flood Risk.” Environmental Research Letters 19: 104034. Baysan, C., M. Burke, F. Gonzalez, S. Hsiang, and T. Miguel. 2019. “Economic and Non- Economic Factors in Violence: Evidence from Organized Crime, Suicides and Climate in Mexico.” Journal of Economic Behavior and Organization 168: 434–452. Brück, T., M. d’Errico, and R. Pietrelli. 2019. “The Effects of Violent Conflict on Household Resilience and Food Security: Evidence from the 2014 Gaza Conflict.” World Development 119: 203–223. Buhaug, H. 2010. “Climate Not to Blame for African Civil Wars.” Proceedings of the National Academy of Sciences 107 (38): 16477–16482. Buhaug, H., H. Hegre, and H. Strand. 2010. “Sensitivity Analysis of Climate Variability and Civil War.” Peace Research Institute Oslo (PRIO) 17: 1–17. Buhaug, H., and N. von Uexkull. 2021. “Vicious Circles: Violence, Vulnerability, and Climate Change.” Annual Review of Environment and Resources 46 (1): 545–568. Burke, M. B., E. Miguel, S. Satyanath, J. A. Dykema, and D. B. Lobell. 2009. “Warming Increases the Risk of Civil War in Africa.” Proceedings of the National Academy of Sciences 106 (49): 20670–20674. Burke, M. B., E. Miguel, S. Satyanath, J. A. Dykema, and D. B. Lobell. 2009. “Warming Increases the Risk of Civil War in Africa.” Proceedings of the National Academy of Sciences 106: 20670–20674. Burke, M., J. Ferguson, S. Hsiang, and E. Miguel. 2024a. “New Evidence on the Economics of Climate and Conflict.” NBER Working Paper 33040, October 2024. Burke, M., J. Ferguson, S. Hsiang, and E. Miguel. 2024b. “Will Wealth Weaken Weather Wars?” AEA Papers and Proceedings 114: 65–69. 23 Burke, M., S. M. Hsiang, and E. Miguel. 2015. “Climate and Conflict.” Annual Review of Economics 7: 577–617. Carleton, T., M. Greenstone, S. Hsiang, A. Hutgren, A. Jina, R. Kopp, and A. Rode. 2017. “Food Security and Social Stability in Africa: New Estimation Methods for Data-Driven Climate Impact Projections in Data-Sparse Regions.” International Growth Centre. Caso, N., D. Hilhorst, and R. Mena. 2023. “The Contribution of Armed Conflict to Vulnerability to Disaster: Empirical Evidence from 1989 to 2018.” International Journal of Disaster Risk Reduction 95: 103881. Church, C., and A. Crawford. 2018. Green Conflict Minerals: The Fuels of Conflict in the Transition to a Low-Carbon Economy. International Institute for Sustainable Development. Clement, V. W. C., K. K. Rigaud, A. de Sherbinin, B. R. Jones, S. Adamo, J. Schewe, N. Sadiq, and E. S. Shabahat. 2021. Groundswell Part II: Acting on Internal Climate Migration. Vol. 1 of 2. Washington, DC: World Bank Group. Collier, P. 1999. “On the Economic Consequences of Civil War.” Oxford Economic Papers 51: 168–183. Cuesta, J., L. Madrigal, and N. Pecorari. 2024. “Social Sustainability, Poverty and Income: An Empirical Exploration.” Journal of International Development 36 (3): 1789–1816. Das Gupta, M. 2014. “Population, Poverty, and Climate Change.” World Bank Research Observer 29 (1): 83–108. Desposato, S. W., G. Wang, and J. Y. Wu. 2021. “The Long-Term Impact of Mobilization and Repression on Political Trust.” Comparative Political Studies 54 (14): 2447–2474. Fairhead, J., M. Leach, and I. Scoones. 2012. “Green Grabbing: A New Appropriation of Nature?” Journal of Peasant Studies 39 (2): 237–261. Field, J. 2018. “Divided Disasters: Examining the Impacts of the Conflict–Disaster Nexus for Distanced Crises in the Philippines.” Disasters 42 (S2): S265–S286. Fook, T. C. T. 2015. “Transformational Processes for Community-Focused Adaptation and Social Change: A Synthesis.” Climate and Development 9 (1): 5–21. Froese, R., and J. Schilling. 2019. “The Nexus of Climate Change, Land Use, and Conflicts.” Current Climate Change Reports 5 (1): 24–35. Gates, S., H. Hegre, H. M. Nygård, and H. Strand. 2012. “Development Consequences of Armed Conflict.” World Development 40 (9): 1713–1722. 24 Ghimire, R., and S. Ferreira. 2015. “Floods and Armed Conflict.” Environment and Development Economics 21: 23–52. Ghobarah, H. A., P. Huth, and B. Russet. 2003. “Civil Wars Kill and Maim People Long after the Shooting Stops.” American Political Science Review 97 (2): 189–202. Gyau-Boakye, P. 2001. “Environmental Impacts of the Akosombo Dam and Effects of Climate Change on the Lake Levels.” Environment, Development and Sustainability 3 (1): 17–29. Hatz, S. 2019. “Israeli Demolition Orders and Palestinian Preferences for Dissent.” The Journal of Politics 81 (3): 1069–1074. Hegre, H., and N. Sambanis. 2006. “Sensitivity Analysis of Empirical Results on Civil War Onset.” Journal of Conflict Resolution 50: 508–535. Hoffmann, R. 2022. “Contextualizing Climate Change Impacts on Human Mobility in African Dry Lands.” Earth’s Future 10: e2021EF002591. Homer-Dixon, T. F. 1999. Environment, Scarcity, and Violence. Princeton: Princeton University Press. Hsiang, S., and M. Burke. 2014. “Climate, Conflict, and Social Stability: What Does the Evidence Say?” Climatic Change 123 (1): 39–55. Hsiang, S., M. Burke, and E. Miguel. 2013. “Quantifying the Influence of Climate on Human Conflict.” Science 341 (6151): Article 1235367. Hunsberger, C., E. Corbera, and C. Vaddhanaphuti. 2019. “Climate Change Policies, Land Control and Conflict in Southeast Asia.” Canadian Journal of Development Studies Policy Brief Series 1. Ide, T., A. Kristensen, and B. Henrikas. 2021. “First Comes the River, Then Comes the Conflict? A Qualitative Comparative Analysis of Flood-Related Political Unrest.” Journal of Peace Research 58: 83–97. IPCC. 2022. “Contribution of Working Group II to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change.” In Climate Change 2022: Impacts, Adaptation and Vulnerability. Cambridge: Cambridge University Press. IUCN, UNEP, and WWF. 1980. World Conservation Strategy: Living Resource Conservation for Sustainable Development. Gland, Switzerland: IUCN. Iyigun, M., J. Mueller, and N. Qian. 2024. “The Intensifying Effects of Prolonged Climate Change on Conflict, 1400–1900 CE.” AEA Papers and Proceedings 114: 80–83. 25 Jaramillo, L., A. Cebotari, Y. Diallo, R. Gupta, Y. Koshima, C. Kularatne, D. J. Dae Lee, S. Rehman, K. Tintchev, and F. Yang. 2023. Climate Challenges in Fragile and Conflict-Affected States. IMF Staff Climate Note 2023/001. Washington, DC: International Monetary Fund. Jones, S., K. Oven, and B. Wisner. 2016. “A Comparison of the Governance Landscape of Earthquake Risk Reduction in Nepal and the Indian State of Bihar.” International Journal of Disaster Risk Reduction 15: 29–42. Kassegn, A., E. Ebrahim, and F. Yildiz. 2021. “Review on Livelihood Diversification and Food Security Situations in Ethiopia.” Cogent Food & Agriculture 7 (1): 1882135. Koubi, V. 2019. “Climate Change and Conflict.” Annual Review of Political Science 22: 18.1– 18.18. Ko, J., H. F. Lee, and C. K. Leung. 2024. “War and Warming: The Effects of Climate Change on Military Conflicts in Developing Countries (1995–2020).” Innovation and Green Development 3: 100175. Lee, H. F., D. Zhang, P. Brecke, and Q. Pei. 2019. “Climate Change, Population Pressure, and Wars in European History.” Asian Geographer 36 (1): 29–45. Lupu, N., and L. Peisakhin. 2017. “The Legacy of Political Violence across Generations.” American Journal of Political Science 61 (4): 836–851. Mach, K., C. Kraan, N. Adger, H. Buhaug, M. Burke, et al. 2019. “Climate as a Risk Factor for Armed Conflict.” Nature 571 (7764): 193–197. Mach, K., M. D. Mastrandrea, T. E. Bilir, and C. B. Field. 2016. “Understanding and Responding to Danger from Climate Change: The Role of Key Risks in the IPCC AR5.” Climatic Change 136: 427–444. Mackay, R., A. Mavisakalyan, and Y. Tarverdi. 2024. “Trust a Few: Natural Disasters and the Disruption of Trust in Africa.” Journal of Behavioral and Experimental Economics 113: 102288. Malloy, J. T., and C. M. Ashcraft. 2020. “A Framework for Implementing Socially Just Climate Adaptation.” Climatic Change 160 (1): 1–14. Marktanner, M., E. Mienie, and L. Noiset. 2015. “From Armed Conflict to Disaster Vulnerability.” Disaster Prevention and Management 24 (1): 53–69. Mekonen, D. A., and A. B. Berlie. 2021. “Rural Households’ Livelihood Vulnerability to Climate Variability and Extremes: A Livelihood Zone-Based Approach in the Northeastern Highlands of Ethiopia.” Ecological Processes 10 (1): 55. Mera, G. A. 2018. “Drought and Its Impacts in Ethiopia.” Weather and Climate Extremes 22: 24– 35. 26 Minale, A. S., Z. Y. Alemayehu, and A. G. Adam. 2024. “Climate-Change-Driven Conflict: Insights from North Wollo, Northeast Ethiopia.” Sustainable Environment 10 (1): 2361563. Mobjörk, M., F. Krampe, and K. Tarif. 2020. Pathways of Climate Insecurity: Guidance for Policy Makers. SIPRI Policy Brief. Miguel, E., S. Satyanath, and E. Sergenti. 2004. “Economic Shocks and Civil Conflict: An Instrumental Variables Approach.” Journal of Political Economy 112 (4): 725–753. Milanovic, B. 2024. “The Three Eras of Global Inequality, 1820–2020 with the Focus on the Past Thirty Years.” World Development 177: 106516. Mollica, R., K. McInnes, N. Sarajlić, J. Lavelle, I. Sarajlić, and M. Massagli. 1999. “Disability Associated with Psychiatric Comorbidity and Health Status in Bosnian Refugees Living in Croatia.” Journal of the American Medical Association 282 (5): 433–439. Mulugeta, K. 2011. “The Causes of Ethio-Somali War of 2006.” https://repositorio.uam.es/handle/10486/677558. Myers, R., C. Luttrell, R. Harjanthi, M. R. Fisher, M. Menton, P. Läderach, and E. Wollenberg. 2021. Climate Change Mitigation in Forests: Conflict, Peacebuilding, and Lessons for Climate Security - Position Paper. CGIAR. Nordensvard, J., T. Gilbertson, and D. Marinelli. 2016. Social Implications of Carbon Markets – The Case of Carbon Offsets and Plantar in Brazil. ND GAIN. n.d. ND-GAIN Country Index. Notre Dame Global Adaptation Initiative. Nielsen, F., A. Uras, E. Y. Musa, and A. H. M. Osman. 2022. “Natural Resources Management, Environmental Governance, and Peacebuilding in Darfur: Creating a Win-Win Situation When Farmers and Pastoralists Were Incentivized to Find Shared Solutions.” In O. Brown and G. Nicolucci-Altman (eds.), Ecosystem for Peace: A Compendium of Ideas. Njiru, B. N. 2012. “Climate Change, Resource Competition, and Conflict Amongst Pastoral Communities in Kenya.” In J. Scheffran, M. Brzoska, H. G. Brauch, P. M. Link, and J. Schilling (eds.), Climate Change, Human Security and Violent Conflict: Challenges for Societal Stability, 513–527. Springer. Nunn, N., and N. Qian. 2011. “The Potato’s Contribution to Population and Urbanization: Evidence from a Historical Experiment.” Quarterly Journal of Economics 126 (2): 593–650. Pecorari, N., and J. Cuesta. 2024. “Citizen Participation and Political Trust in Latin America and the Caribbean: A Machine Learning Approach.” European Journal of Development Research 36: 1227–1252. 27 Peters, K., and M. Budimir. 2016. When Disasters and Conflict Collide: Facts and Figures. ODI Briefing. https://cdn.odi.org/media/documents/When_disasters_and_conflicts_collide_facts_and_figures.p df. Petrova, K., and E. L. Rosvold. 2024. “Mitigating the Legacy of Violence: Can Flood Relief Improve People’s Trust in Government in Conflict-Affected Areas? Evidence from Pakistan.” World Development 173: 106372. Petrova, K. 2022. “Floods, Communal Conflict and the Role of Local State Institutions in Sub- Saharan Africa.” Political Geography 92: Article 102511. Paavola, J. 2017. “Health Impacts of Climate Change and Health and Social Inequalities in the UK.” Environmental Health 16 (1): 113. Pottier, A., M. Fleurbaey, A. Mejean, and S. Zuber. 2021. “Climate Change and Population: An Assessment of Mortality Due to Health Impacts.” Ecological Economics 183: Article 106967. Raleigh, C. 2011. “The Search for Safety: The Effects of Conflict, Poverty and Ecological Influences on Migration in the Developing World.” Global Environmental Change 21S: S82– S93. Roy, A. 2024. “A Panel Data Study on the Effect of Climate Change on Life Expectancy.” PLOS Climate 3 (1): Article e0000339. Sacchetto, C., N. Stern, and C. Taylor. 2020. Priorities for Renewable Energy Investment in Fragile States. International Growth Centre (IGC). Sakaguchi, K., A. Varughese, and G. Auld. 2017. “Climate Wars? A Systematic Review of Empirical Analyses on the Links Between Climate Change and Violent Conflict.” International Studies Review 19 (4): 622–645. Salehyan, I. 2008. “From Climate Change to Conflict? No Consensus Yet.” Journal of Peace Research 45 (3): 315–326. Simangan, D., S. Bose, J. L. Candelaria, F. Krampe, and S. Kaneko. 2023. “Positive Peace and Environmental Sustainability: Local Evidence from Afghanistan and Nepal.” In Environment and Security. Stapleton, S. O., R. Nadin, C. Watson, and J. Kellett. 2017. Climate Change, Migration and Displacement: The Need for a Risk-Informed and Coherent Approach. Overseas Development Institute and United Nations Development Programme. Stern, N. 2007. The Economics of Climate Change: The Stern Review. Cambridge: Cambridge University Press. 28 Theisen, O. M., N. P. Gleditsch, and H. Buhaug. 2013. “Is Climate Change a Driver of Armed Conflict?” Climatic Change 117 (3): 613–625. Van de Vuurest, P., and L. E. Escobar. 2023. “Climate Change and Infectious Disease: A Review of Evidence and Research Trends.” Infectious Diseases of Poverty 12 (1): 51. Van Schaik, L., C. Born, E. Sellwood, and S. de Bruin. 2019. Making Peace with Climate. Background Paper for the Global Commission on Adaptation. Walch, C. 2018. “Disaster Risk Reduction Amidst Armed Conflict: Informal Institutions, Rebel Groups, and Wartime Political Orders.” Disasters 42 (S2): S239–S264. Warnecke, A., D. Tänzler, and R. Vollmer. 2010. “Climate Change, Migration and Conflict: Receiving Communities under Pressure?” International Journal of Research in Environmental Studies 5: 18–35. Woodward, G., D. M. Perkins, and L. E. Brown. 2010. “Climate Change and Freshwater Ecosystems: Impacts Across Multiple Levels of Organization.” Philosophical Transactions of the Royal Society B: Biological Sciences 365 (1549): 2093–2106. World Bank. 2021. From COVID-19 Crisis Response to Resilient Recovery: Saving Lives and Livelihoods while Supporting Green, Resilient and Inclusive Development. Prepared by the World Bank Group for the Virtual April 9, 2021 Development Committee Meeting. Zhang, D., J. Zhang, H. F. Lee, and Y. He. 2007. “Climate Change and War Frequency in Eastern China over the Last Millennium.” Human Ecology 35 (4): 403–414. Zhang, S., D. Zhang, J. Li, and Q. Pei. 2020. “Secular Temperature Variations and the Spatial Disparities of War in Historical China.” Climatic Change 159 (4): 545–564. 29 Annex A: FCV Categories (2019-2024) This Annex outlines the methodology used for the classification of countries into mutually exclusive groups based on fragility, conflict, and violence (FCV) criteria. The grouping process is based on quantitative thresholds applied to multiple datasets: Overview of the Grouping Criteria The grouping process resulted in four mutually exclusive categories: National Conflict, Subnational Conflict, Violence, and Fragility. The mutually exclusive nature of these groups ensures clarity and avoids overlaps. Additionally, a hierarchical approach was used to prioritize National Conflict over Subnational Conflict, Violence, and Fragility in classification. Data Sources The classification methodology integrates data from multiple sources: • ACLED (Armed Conflict Location & Event Data Project): Used for tracking conflict intensity and fatalities. • UCDP (Uppsala Conflict Data Program): Used for conflict-related deaths and subnational conflict dynamics. • UNODC (United Nations Office on Drugs and Crime): Used for intentional homicide rates. • Fragility Data: Includes countries classified as extremely and non-extremely fragile in 2020 and 2022, based on lists of countries provided by the OECD. Classification Criteria a. National Conflict Group Countries with widespread, high-intensity conflict affecting a significant proportion of their population are classified as National Conflict based on: Thresholds at the Subnational (ADMIN1) Level: • ACLED: ACLED_BRD_per100K > 2 • UCDP: UCDP_BRD_per100K > 1 • Affected area: The share of administrative regions (ADMIN1) affected by conflict is calculated as:   1     min 1   ℎ  =    1   ℎ  • Classification Rule: A country is classified as National Conflict if more than 20% of its ADMIN1 areas are affected by conflict. • Persistence: At least 3 of the past 5 years must meet the above criteria. b. Subnational Conflict Group Countries experiencing high-intensity conflict affecting localized populations (e.g., subnational regions) are classified as Subnational Conflict based on the affected_admin1_share method. This method evaluates the share of administrative regions affected by conflict, ensuring a localized perspective on conflict intensity. 30 Thresholds at the Subnational (ADMIN1) Level: • ACLED: ACLED_BRD_per100K > 2 • UCDP: UCDP_BRD_per100K > 1 • Affected area: The share of administrative regions (ADMIN1) affected by conflict is calculated as above. • Classification Rule: A country is classified as Subnational Conflict if 20% or fewer of the ADMIN1 areas are affected by conflict. • Persistence: At least 3 of the past 5 years must meet the above criteria. • Exclusion Rule: Countries classified as National Conflict are excluded. c. Violence Group Countries are classified as violent if they meet the following conditions: • Homicide Rate: If the country has had 3 or more years with a homicide rate per 100K population greater than 10 in the past five years. • Exclusion Rule: Countries classified as National Conflict or Subnational Conflict are excluded. d. Fragility Group Countries classified as fragile but not already included in the above groups remain in this category. Hierarchical Exclusion Criteria To ensure mutually exclusive groups, a hierarchical procedure was applied: 1. National Conflict Group: Countries meeting National Conflict criteria are prioritized and excluded from Subnational Conflict and Violence groups. 2. Subnational Conflict Group: Remaining countries meeting Subnational Conflict criteria are classified here, excluding them from the Violence group. 3. Violence Group: Countries meeting Violence criteria but not classified as National or Subnational Conflict are placed in this group. 4. Fragility: Countries classified as fragile but not already included in the above groups remain in this category. This updated classification structure improves clarity and ensures that national conflicts are properly distinguished from subnational conflicts. By adjusting the hierarchy to prioritize National Conflict over Violence, we better reflect the operational and humanitarian significance of widespread conflict. Table A1 below presents the countries and territories categorized into each of the four groups. 31 Table A.1: countries and territories classified as settings of national conflict, subnational conflict, violence, and fragility. National Conflict Subnational Conflict Violence Fragility Afghanistan Burundi Belize Congo, Rep. Burkina Faso Brazil Costa Rica Eritrea Central African Congo, Dem. Rep. El Salvador Equatorial Guinea Republic Cameroon Colombia Dominica Zimbabwe Ethiopia Egypt, Arab Rep. Honduras Libya Iraq Haiti Jamaica Uganda Mexico Kenya Saint Lucia Korea, Dem. People’s Rep. Mali Mozambique Saint Vincent and the Madagascar Grenadines Myanmar Nigeria Venezuela, RB Guinea-Bissau Niger Philippines South Africa Guinea Pakistan Papua New Guinea Bangladesh Sudan Chad Mauritania Somalia Türkiye Liberia South Sudan West Bank and Gaza Syrian Arab Nicaragua Republic Togo Guatemala Ukraine Zambia Yemen, Rep. Comoros Tajikistan Sierra Leone Lao PDR Angola Eswatini Lesotho Tanzania Côte d’Ivoire Solomon Islands Iran, Islamic Rep. Djibouti Cambodia Gambia, The Timor-Leste Turkmenistan Benin Source: Authors 32 Annex B: Results with and without Controls Readers might wonder how the inclusion of control variables affects our substantive takeaways. To probe these linkages, we conduct a robustness check where we include controls for population and territorial size. This helps address the concern that places that are larger or more populous might suffer more severe climate change impacts. For presentational simplicity, throughout these robustness checks we focus on the links between FCV and climate impacts. Before presenting the robustness checks, it will be helpful to first conduct a more formal analysis of the links between FCV status and climate impacts. Tables B.1 and B.2 present regression results probing this relationship. All models use OLS. Across both types of climate change impacts, there are statistically significant differences in the counts and numbers of total individuals affected. While directionally consistent with our predictions, for the measures of total deaths, we fail to reject the null hypothesis of no difference. Table B.1 Unconditional regressions of FCV status on drought’s impacts, 2019-2024. Dependent variable: Drought Count Drought Total Deaths Drought Total Affected (1) (2) (3) FCV 0.515 *** 30.927 2,967,510.000*** (0.108) (23.779) (532,675.300) Constant 0.245*** 2.966 276,148.700 (0.063) (13.821) (309,611.300) Observations 222 222 222 R 2 0.094 0.008 0.124 Note: * p<0.1; **p<0.05; ***p<0.01 Source: Authors’ estimates. 33 Table B.2 Unconditional regressions of FCV status on floods’ impacts, 2019-2024. Dependent variable: Flood Count Flood Total Deaths Flood Total Affected (1) (2) (3) FCV 2.861 ** 155.699 1,313,759.000* (1.233) (107.675) (669,239.300) Constant 3.939*** 112.034* 631,490.800 (0.716) (62.585) (388,987.500) Observations 222 222 222 R 2 0.024 0.009 0.017 Note: * p<0.1; **p<0.05; ***p<0.01 Source: Authors’ estimates. Tables B.3 and B.4 similarly assess the relationship between climate impacts and FCV; however, they now include controls for population size and geographic area. Both measures come from the World Bank estimates. We focus on measures from the year 2018, as this is directly prior to the temporal window we focus on in our analysis. The results are substantively similar to those presented in Tables B.1 and B.2. The one difference is that the estimates for difference in flood deaths between FCV and non-FCV settings are now statistically significant. Table B.3 Conditional regressions of FCV status on drought’s impacts, 2019-2024. Dependent variable: Drought Count Drought Total Deaths Drought Total Affected (1) (2) (3) FCV 0.504 *** 30.802 2,961,219.000*** (0.104) (24.462) (539,226.700) Territory size (2018) 0.00000*** 0.00000 0.120 (0.00000) (0.00001) (0.164) Population (2018) 0.000** 0.00000 0.004** (0.000) (0.00000) (0.002) Constant 0.158** 0.382 67,353.870 (0.064) (15.060) (331,974.800) Observations 215 215 215 R 2 0.204 0.009 0.150 Note: * p<0.1; p<0.05; ***p<0.01 ** Source: Authors’ estimates. 34 Table B.4 Conditional regressions of FCV status on drought’s impacts, 2019-2024. Dependent variable: Flood Count Flood Total Deaths Flood Total Affected (1) (2) (3) FCV 2.714 ** 159.000 ** 1,337,330.000*** (1.051) (73.352) (437,442.900) Territory size (2018) 0.00000 *** -0.0001 *** -0.298** (0.00000) (0.00002) (0.133) Population (2018) 0.00000*** 0.00000*** 0.028*** (0.000) (0.00000) (0.002) Constant 2.456 *** -1.467 -177,805.900 (0.647) (45.159) (269,311.600) Observations 215 215 215 R 2 0.325 0.566 0.603 Note: * p<0.1; **p<0.05; ***p<0.01 Source: Authors’ estimates. 35 Annex C: Results with Climate Impacts Standardized by 100k Population, 2019-2024 An alternative way of accounting for the possible influence of population size on the severity of climate impacts is to create a standardized version of the dependent variable which directly adjusts for population. To do so, we standardize each of our three measures of climate impacts for droughts and floods by 100k population. Regression models are presented in Tables C.1 and C.2. The results are largely consistent with the main findings from the paper. Table C.1 Unconditional regressions of FCV status on droughts’ impacts, 2019-2024. Dependent variable: Droughts per 100k Drought Deaths per 100k Drought Affected per 100k (1) (2) (3) FCV -0.105 0.078 8,623.328*** (0.101) (0.058) (2,588.468) Constant 0.113* 0.001 3,710.296** (0.059) (0.034) (1,525.269) Observations 216 216 216 R2 0.005 0.009 0.049 Note: * p<0.1; **p<0.05; ***p<0.01 Source: Authors’ estimates. Table C.2 Unconditional regressions of FCV status on floods’ impacts, 2019-2024. Dependent variable: Floods per 100k Flood Deaths per 100k Flood Affected per 100k (1) (2) (3) FCV 0.001 0.391 *** 2,728.972* (0.010) (0.105) (1,499.200) Constant 0.036*** 0.188*** 2,167.069** (0.006) (0.062) (883.412) Observations 216 216 216 R2 0.00002 0.061 0.015 Note: * p<0.1; p<0.05; ***p<0.01 ** Source: Authors’ estimates. 36 Annex D. Correlations between FCV context, climate impacts, and social sustainability without population-adjustments Figure 6D1: Differences in total climate impacts by FCV and non-FCV settings, 2019-2024. Source: Authors’ estimates. 37 Figure 7D2: Differences in total climate impacts by FCV setting, 2019-2024 Source: Authors’ estimates. 38 Figure 8D3: Differences in social sustainability by FCV context, 2019-2024. Source: Authors’ estimates. 39 Figure 9D4: Difference in total climate impacts by high versus low social sustainability, 2019-2024. Source: Authors’ estimates. 40 Annex E: Results with an Alternative Measure of Resilience Tables E1-E3 present results using an alternative measure for resilience from ND-GAIN. Consistent with the analysis presented for social sustainability throughout this paper, we focus on the year 2022. 10 Across nearly all specifications, we observe results that are statistically significant and directionally consistent with our expectations. Table E.1: Regressions of FCV on Resilience, 2022 Dependent variable: ND-GAIN Index (2022 (1) (2) FCV -14.332 *** (1.203) Violence -6.970*** (2.592) National -16.618*** (2.043) Subnational -14.156*** (2.301) Fragility -15.455*** (1.554) Constant 55.255*** 55.255*** (0.752) (0.736) Observations 187 187 R 2 0.434 0.466 Note: * p<0.1; p<0.05; ***p<0.01 ** Source: Authors’ estimates. 10 2022 corresponds to the final year in the Social Sustainability Global Database. 41 Table E.2: Unconditional regressions of resilience on drought’s impacts, 2019-2024 Dependent variable: Drought Count Drought Total Deaths Drought Total Affected (1) (2) (3) ND-GAIN Index 0.373*** 20.099 2,471,849.000*** (Low) (0.121) (26.859) (605,370.800) Constant 0.291*** 5.070 141,646.000 (0.089) (19.739) (444,898.900) Observations 187 187 187 R2 0.049 0.003 0.083 Note: * p<0.1; p<0.05; ***p<0.01 ** Source: Authors’ estimates. Table E.3: Unconditional regressions of resilience on floods’ impacts, 2019-2024 Dependent variable: Flood Count Flood Total Deaths Flood Total Affected (1) (2) (3) ND-GAIN Index (Low) 3.545*** 283.601** 1,052,964.000 (1.343) (119.648) (752,765.700) Constant 3.860*** 42.221 682,719.300 (0.987) (87.932) (553,222.300) Observations 187 187 187 R2 0.036 0.029 0.010 Note: * p<0.1; p<0.05; ***p<0.01 ** Source: Authors’ estimates. 42