## clnea2023001 ## Source details **Canonical URL:** [clnea2023001](https://www.imf.org/-/media/files/publications/staff-climate-notes/2023/english/clnea2023001.pdf) ## Other formats - [Markdown version](/-/media/files/publications/staff-climate-notes/2023/english/clnea2023001.pdf.md) - [Structured JSON version](/-/media/files/publications/staff-climate-notes/2023/english/clnea2023001.pdf.json) --- ### Key findings and framing - The note demonstrates that in fragile and conflict-affected states (FCS), climate vulnerability and underlying fragilities—namely conflict, heavy dependence on rainfed agriculture, and weak capacity and policy buffers—exacerbate each other, amplifying negative impacts on people and economies. - FCS are home to nearly 1 billion people. - About one in five countries around the world are considered to be FCS. - FCS host 43 percent of the global poor living on less than $2.15 per day while only accounting for 12 percent of the world’s population. - Climate vulnerability–fragility nexus: - Climate shocks worsen fragility by increasing competition for scarce resources, lowering rainfed crop yields, and straining public resources. - Fragility (conflict, reliance on rainfed farming, inadequate infrastructure, weak policy buffers) worsens climate vulnerability by limiting emergency response and adaptation capacity. - Several IMF Country Engagement Strategies (Democratic Republic of Congo, Guinea-Bissau, Iraq, Somalia, South Sudan, Yemen) note climate shocks exacerbate fragility and conflict. ### Exposure of FCS to climate shocks - Composite exposure index (average of standardized values of multiple indices) shows FCS are significantly more exposed to climate change than other countries. - Extreme heat: - FCS start from already higher average temperatures and will face significantly higher increases. - UNDP Human Climate Horizons forecasts: by 2040–59 the median FCS will face 61 days of temperatures above 35 degrees Celsius per year (up from 30 days in 1986–2005), compared to only 15 days for other countries (up from 2 days in 1986–2005) under RCP 8.5. - High and rising temperatures endanger human health and reduce labor supply and hours worked in high-risk sectors such as agriculture and construction. - Extreme weather events: - The median FCS has faced disruptive extreme weather events in one out of every four years since 1980. - Floods most frequent, followed by storms and droughts. - Other exposure factors in FCS: - Heavy reliance on climate-dependent sectors (particularly agriculture). - Precarious urban infrastructure, limited access to safe drinking water and sanitation, some FCS face rising sea levels affecting urban centers and ports. - Emissions context: - Carbon dioxide emissions per capita: median FCS 0.5 tons in 2021; other countries 3.5 tons in 2021. - Uncertainty: - Scenarios used for illustration: RCP 8.5, RCP 4.5, and RCP 2.6 from IPCC Sixth Assessment Report. ### Near-term and longer-term impacts on macroeconomic outcomes and food security - General conclusion: - Climate shocks hurt FCS economies in the near term and erode growth and development in the longer term, more so than in other countries. - Near-term effects: GDP losses, food production shortfalls, high inflation, deterioration of the external position. - Longer-term: worsening drought conditions erode real GDP per capita growth and exacerbate hunger, widening income gaps between FCS and other countries. - Near-term impacts of extreme weather events: - Negative impact is larger and more persistent in FCS vs non-FCS. - After three years, cumulative GDP losses reach about 4 percent in FCS, compared to 1 percent in non-FCS. - Consumption does not rebound even after three years in FCS. - Investment starts to recover after two years in FCS versus one year in non-FCS. - Extreme weather events reduce exports and widen the current account deficit. - Longer-term drought effects (SPEI deviations from trend): - Drought conditions would cut real GDP per capita growth every year by 0.2 percentage point in RCP 2.6. - Drought conditions would cut real GDP per capita growth every year by 0.4 percentage point in RCP 8.5. - By 2060, real GDP per capita in FCS would be 5 percent lower in the high emissions scenario compared to the low emissions scenario (RCP 8.5 vs RCP 2.6). - Key channels: lower crop productivity, reduced food production, weaker investment. - Effects are asymmetric: increases in SPEI (wetter conditions) show no significant longer-term impacts. - Food security: - Food production in fragile states is two times more sensitive to drought conditions over the longer term than other countries. - Food represents 42 percent of total consumer expenditure for the median FCS, compared to 23 percent in other countries. - High emissions scenario: confluence of lower food production and higher prices would push 2 percentage points more of fragile states’ population—about 50 million more people—into hunger by 2060. ### Humanitarian and conflict impacts - Humanitarian casualties and displacement: - Three times as many people in FCS are affected every year by extreme weather events than in other countries (CRED EM-DAT data). - Following extreme weather events, the share of undernourished population increases significantly in FCS from already high levels. - Close to 10 percent of internal displacement in FCS is directly linked to disasters; more than twice the share of the population is displaced in FCS than in other countries. - Close to 95 percent of refugees, 86 percent of internally displaced people, and 20 percent of migrants globally have originated in FCS countries (2022 aggregates reported in source figures). - Climate shocks and conflict intensity: - Georeferenced analysis for 106 countries over 2013–2022: climate shocks do not affect onset of new conflict but exacerbate intensity where conflict exists. - Estimated effects: - A 1 percent increase in temperature is associated with a 0.1 percent increase in conflict intensity (conflict-related deaths/population). - Annualized projection: under a high emissions scenario and all else equal, by 2060 conflict deaths as a share of the population for a median FCS could increase by 8.5 percent, and up to 14 percent for countries facing an extreme temperature increase. - Countries facing an extreme drop in precipitation could see conflict intensity increase of up to 8 percent. - Pathways: resource scarcity, food security deterioration, displacement and migration pressures, economic shocks from damaged infrastructure and reduced productivity. - Note: results underestimate destructive impacts of conflict that do not result in higher deaths. ### Agriculture dependence, irrigation, and vulnerability in FCS - Agriculture indicators: - Value added of agriculture: 22 percent of GDP in FCS in 2021; 6 percent in non-FCS. - Employment in agriculture: 43 percent in FCS in 2019; 14 percent in non-FCS in 2019. - Irrigation and rainfed farming: - AQUASTAT: only 3 percent of cultivated areas in FCS are equipped for irrigation, compared to 11 percent in non-FCS. - Satellite-based farm “greenness” analysis (30 FCS, 1984–2021): - Rainfed farms are significantly affected by rainfall volatility, groundwater, and flood frequency. - Irrigated farms are not significantly affected by these climate variables when irrigation functions effectively. - Rainfed farms lose 11 percent of their vegetation when a rainy season disappoints (precipitation falls by one standard deviation). - Fragility exacerbates agricultural climate vulnerability: - Even existing irrigation often underperforms and substantial irrigated farmlands are unused: estimates suggest 43 percent of area of an irrigated farmland is typically unused for the average of 19 irrigation schemes in FCS. - Well-performing irrigation schemes: farmland vegetation stable and uncorrelated to rainfall (example: Lebanon). - Poorly performing schemes: vegetation highly correlated with climate variables (example: Lake Assad scheme in Syria). - Rainfed farms often show an inverted U-curve where vegetation drops with little or heavy rainfalls (example: West Bank and Gaza). ### Case-study evidence on irrigation damage, abandonment, and poor design - Conflict-related destruction and abandonment: - Northern Iraq: Jazeera irrigation scheme and other farmlands almost completely disappeared under ISIS occupation. - Mali: Office du Niger has underperformed since security crisis in 2012; vegetation losses linked to increased flooding and farmer flight. - Inadequate maintenance and capacity constraints: - Gravity surface irrigation requires constant maintenance; lack of maintenance cited as main reason for underperformance. - Examples: Mozambique Limpopo River plain—rehabilitation in 2010s improved resilience after 2000 floods; Sudan Gezira scheme performs poorly due to sediment accumulation and clogging; South Sudan Aweil scheme abandoned, maintenance last undertaken in the 1980s. - Poor design and policy implementation: - Ethiopia Lower Awash Plain: expansion in late 2000s lacked proper drainage, exacerbated salinity hazards and reduced vegetation. - Libya GMMR: unable to provide stable water supply to Benghazi irrigation scheme; caused sensitivity to local groundwater. - Land reforms and input subsidies can either reduce or aggravate sensitivity depending on design and implementation (Ethiopia, Zimbabwe examples). - Weak oversight and governance: - Alternative financing (PPPs, SOEs) often face inadequate oversight; private sector involvement does not guarantee adequate investment (Central African Republic, Democratic Republic of the Congo palm plantations). ### Policies to facilitate immediate response to climate shocks (near-term mitigation) - Build buffers and strengthen institutional capacity: - Countries with larger fiscal buffers—higher fiscal balance and lower public debt—see faster recovery from extreme weather events. - It takes three years for FCS with a relatively higher reserves-to-imports ratio to return to predisaster incomes, compared to five years for FCS with a relatively low ratio. - Ex ante instruments: contingent budget lines and prearranged contingent loans (for example, from international financial institutions). - Improve spending efficiency and public investment management. - Macroeconomic frameworks should reflect adaptation policies and climate risk mitigation and preparedness. - Strengthen social safety nets: - Well-targeted programs with efficient delivery systems that can be scaled up rapidly and wound down once emergency subsides. - Transfer disaster risk through sovereign insurance: - Most FCS cannot self-insure; sovereign insurance is important where cost-effective. - Fragile states should seek regional insurance pools where these exist; several fragile states are insured members (Burkina Faso, Mali, the Marshall Islands, Niger) or eligible members (Chad, Ethiopia, Micronesia, Papua New Guinea, Solomon Islands, Timor-Leste, Tuvalu, Zimbabwe). - Level of insurance remains small due to cost; donor grants could help achieve a more optimal level of protection. ### Policies to build climate resilience over time (structural and governance reforms) - Embed climate-resilience into peace and security efforts: - Adaptation policies must be carefully prioritized, tailored to local contexts, and support conflict prevention and resolution. - Improve governance and fight corruption: - FCS with below average control of corruption would see lower real GDP per capita growth every year by about 0.2 percentage point in RCP 2.6 and 0.4 percentage point in RCP 8.5 compared to FCS with above average control of corruption. - Governance areas: fiscal governance, rule of law, market regulation. - Upgrading governance frameworks can broaden access to climate financing. - Develop climate-smart agriculture: - Countries with higher agricultural investment are less affected by disasters; improved irrigation reduces vulnerability to droughts and floods. - Recommended investments: irrigation and drainage systems, water management, fertilizer and insecticide use, machinery, anti-erosion measures, improved seeds and livestock; early warning systems; broaden mobile phone availability in rural areas. - Scale up social spending and climate-resilient infrastructure: - Regression results show higher social spending helps mitigate adverse impact of climate change on growth. - Monitor asset conditions, ensure efficient selection, execution, and maintenance of projects, and base choices on credible cost-benefit analyses. - Incorporate climate financing throughout budgeting, ensure transparent procurement, and implement risk management. - Enhance financial inclusion: - Enables climate-resilient investment, access to savings, emergency borrowing, and insurance. - In FCS where mobile money matters, broaden mobile phone availability and strengthen supervision of mobile money operations. ### Financing needs, international support, and IMF role - Financing needs and gaps: - Estimated adaptation costs for FCS: about 1.5 percent of GDP per year (Aligishiev, Massetti, and Bellon 2022), compared to 1 percent of GDP in other countries. - Corresponds to 13 percent of tax revenues in FCS compared to less than 6 percent in other countries. - Aid Atlas reports climate adaptation aid flow commitments to FCS at about US$30 billion cumulative between 2010 and 2020. - Humanitarian costs following climate-related disasters estimated at US$3.5 to US$12 billion per year, which could balloon to US$20 billion per year by 2030. - Total humanitarian aid from OECD countries reached US$25 billion in 2021, of which an estimated 80 percent was directed to FCS. - Need for grants and concessional financing: - Financial support should be grants and concessional financing to avoid pressures on fiscal sustainability. - Multilateral and bilateral financing essential alongside domestic revenue mobilization and reprioritization of spending. - Facilitate access to global climate funds and tailored support: - Barriers: limited donor political will for long-term actions, inadequate technical capacity in FCS, security concerns, varying requirements across climate funds. - FCS need well-defined climate strategies with credible, “bankable” projects linked to development strategy and macroeconomic framework. - Climate funds should provide FCS-focused technical support facilities, dedicated funding windows, and tailored application criteria. - Improvements in quality and availability of data are needed. - Support for sovereign disaster risk insurance: - International partners can help recapitalize regional pools, increase coverage limits, and lower insurance premia. - Direct support can include temporary subsidization of premiums to encourage countries to join regional insurance pools. - Technical assistance and capacity development: - FCS need technical assistance and training to strengthen capacity to absorb and spend climate finance effectively. - IMF support: tailored policy advice, financial assistance, and capacity development under the IMF’s FCS Strategy (IMF 2022a). - IMF has enhanced financial support for climate action through standard facilities, emergency financing, and the Resilience and Sustainability Facility (RSF). - As of July 5, 2023, nine RSF-supported programs have been approved, including for two FCS (Kosovo, Niger). - RSF-supported reforms: strengthen monitoring of climate-related spending, integrate climate risks into fiscal planning, incorporate climate issues into public investment management, strengthen climate-related risk management for financial institutions. - RSF program design for FCS should include assessment of climate vulnerability-fragility links, fiscal policies for efficient response, governance strengthening, and comprehensive capacity development support. ### Scenario analysis and Annex model estimates (selected quantitative projections) - Scenario framework: - High emissions scenario: RCP 8.5. - Low emissions scenario: RCP 2.6. - Projection horizon: 2023–60. - Key projected impacts in 2060 (high emissions vs low emissions): - FCS’ drought-induced per capita income loss is estimated to be about 5 percent higher. - Investment would be lower by 3.5 percent. - Food production would be lower by 7 percent. - The share of food imports in total imports would rise by 2 percent. - Inflation would be higher by 2.5 percent. - An additional 2 percent of fragile states’ growing population—about 50 million people—would be pushed into undernourishment by 2060. - Long-term quantified impacts (point estimates summarized): - Growth rates of real GDP per capita, investment, and crop yields in FCS would decline by about 0.2 percentage point annually due to mean deterioration in drought conditions implied by RCP 2.6. - Average food production would be lower by about 8 percent. - Average inflation would be higher by close to 1 percentage point. - Structural amplifiers of drought-induced losses: - High public debt, low social spending, low trade openness, high water insecurity, weak regulatory quality, weak control of corruption. - Robustness: - Results robust to alternative specifications, restricted samples (per capita incomes below $10,000 US dollars), and alternative FCS groupings. ### Methods, data, and key metrics used - Empirical innovations to mitigate data constraints: georeferencing, geospatial analysis, high-frequency and subregional datasets, NDVI-based farmland mapping. - Definitions and data sources: - FCS definition: World Bank country list FY2006–FY2024 (IMF adopted methodology as part of FCS strategy approved in 2022). - Exposure index: ND-GAIN exposure subindex, INFORM Risk Index natural hazard exposure subindex, UNDP Human Climate Horizons projected increase in human deaths, World Risk Index climate exposure subindex. - Adaptive capacity index: World Risk Index coping and adaptive capacity subindices and INFORM coping capacity subindex. - Extreme weather events: Centre for Research on the Epidemiology of Disasters, Emergency Events Database. - Near-term empirical method: - Jorda’s local projection method for horizons h = 0, …, 3. - Climate disaster dummy = 1 when annual death plus 0.3 times affected people exceeds 0.01 percent of population. - Sample: all countries excluding advanced economies and small states (population below 1 million), period 2004–20. - Longer-term empirical method: - Dynamic panel autoregressive distributed lag growth model (modification of Dell, Jones, and Olken (2012)); climate variables expressed as deviations from long-term trend using Hodrick-Prescott filter. - Sample: 159 developing and developed countries over 1975–2018 (excluding small island states). - SPEI classification preserved: SPEI varies between +5 and –5; non-drought (SPEI > –0.5), mild drought (–1 < SPEI < –0.5), moderate drought (–1.5 < SPEI < –1), severe drought (–2 < SPEI < –1.5), extreme drought (SPEI < –2). - NDVI and geospatial methods: - NDVI = (NIR − RED) / (NIR + RED); NDVI ranges from –1.0 to +1.0. - Area with NDVI higher than 0.3 considered vegetated farmland (adjustments for palm vegetation and older Landsat generations). - Landsat imagery at 30m × 30m resolution used to generate time-series of farmland vegetation and to separate irrigated vs rainfed farmland. - Selected regression coefficients and medians preserved exactly in annexes (examples): - Fiscal Balance regressions (Disaster coefficients): -1.069, -2.509, -2.678, -2.289. - Disaster#Policy coefficients: 1.235, 1.130, 1.095, 2.073. - Public debt medians: Total Sample 40.7; “High” Group 70.9; “Low” Group 25.2. - Foreign reserves (months of imports) medians: Total Sample 3.2; “High” Group 5.4; “Low” Group 1.6. - Rainfed sample regression outputs: RAIN coefficient 1.6338* (standard error [0.88177]); FLOOD coefficient -14.8085** (standard error [6.71930]); GW coefficient 2.9133* (standard error [1.60900]). - Irrigated sample regression outputs: RAIN coefficient 0.12151 (standard error [0.25050]); FLOOD coefficient -2.3109 (standard error [2.52726]); GW coefficient -0.25172 (standard error [0.92623]). - Observations: irrigated farms 510; rainfed farms 412; number of farmlands irrigated 19; rainfed 16. - Conflict–climate monthly subregional analysis: unit = monthly, 2,848 subregions, 171 countries, total observations 340,842; a 1 percent increase in temperature associated with 0.1 percent increase in conflict intensity; a 1 percent increase in precipitation associated with 0.02 percent decline in conflict intensity. *Source: clnea2023001 — IMF Staff Climate Notes (selected excerpts).* ### Introduction ### clnea2023001 - Introduction ### Key findings and framing - The note demonstrates that in fragile and conflict-affected states (FCS), climate vulnerability and underlying fragilities—namely conflict, heavy dependence on rainfed agriculture, and weak capacity and policy buffers—exacerbate each other, amplifying negative impacts on people and economies. - FCS are home to nearly 1 billion people. - About one in five countries around the world are considered to be FCS. - FCS host 43 percent of the global poor living on less than $2.15 per day while only accounting for 12 percent of the world’s population. - FCS are highly exposed to climate change and lack the means or capacity to adapt, creating a climate vulnerability–fragility nexus where: - Climate shocks worsen fragility by increasing competition for scarce resources, lowering rainfed crop yields, and straining public resources. - Fragility (including conflict, reliance on rainfed farming, inadequate infrastructure, and weak policy buffers) worsens climate vulnerability by limiting emergency response and adaptation capacity. - Several IMF Country Engagement Strategies (Democratic Republic of Congo, Guinea-Bissau, Iraq, Somalia, South Sudan, Yemen) underscore that climate shocks exacerbate fragility and conflict. ### Exposure of FCS to climate shocks - A composite exposure index (average of standardized values of multiple indices) shows FCS are significantly more exposed to climate change than other countries. - Extreme heat: - FCS will face significantly higher increases in temperature, starting from already higher average temperatures than non-FCS. - According to UN Development Programme Human Climate Horizons forecasts, by 2040–59 the median FCS will face 61 days of temperatures above 35 degrees Celsius per year (up from 30 days in 1986–2005), compared to only 15 days for other countries (up from 2 days in 1986–2005) under a high emissions scenario (RCP 8.5). - High and rising temperatures endanger human health and reduce labor supply and hours worked in high-risk sectors such as agriculture and construction. - Extreme weather events: - The median FCS has faced disruptive extreme weather events in one out of every four years since 1980, leaving little time to fully recover before the next disaster. - Floods have been the most frequent type of disaster, followed by storms and droughts. - Other exposure factors in FCS: - Heavy reliance on climate-dependent sectors (particularly agriculture). - Precarious urban infrastructure, including populations pushed into flood- and landslide-prone areas. - Limited access to safe drinking water and sanitation. - Some FCS face rising sea levels affecting urban centers and major ports. - Emissions context: - Carbon dioxide emissions per capita in FCS are much lower than in other countries: 0.5 tons for the median FCS in 2021 compared to 3.5 tons for other countries. - Uncertainty: - The note notes considerable uncertainty around global emissions trajectories and long-term climate and macroeconomic modeling, and for illustrative purposes uses RCP 8.5, RCP 4.5, and RCP 2.6 scenarios from the IPCC Sixth Assessment Report. ### Impact of climate shocks on macroeconomic outcomes and food security - General conclusion: - Climate shocks hurt FCS economies in the near term and will take a toll on growth and economic development in the longer term, more so than in other countries. - Near-term effects include GDP losses, food production shortfalls, high inflation, and deterioration of the external position. - Over the longer term, worsening drought conditions will erode real GDP per capita growth and exacerbate hunger, widening income gaps between FCS and other countries. - Near-term impacts of extreme weather events (droughts, floods, storms): - The negative impact of extreme weather events is both larger and more persistent in FCS relative to non-FCS. - After three years, cumulative GDP losses reach about 4 percent in FCS, compared to 1 percent in non-FCS. - Possible channels for GDP losses include disruption of economic activity, destruction of productive capital and infrastructure, lower agricultural productivity (damages to crops and livestock), and diversion of resources toward reconstruction. - Consumption does not rebound even after three years in FCS, possibly due to lack of social safety nets, limited financial inclusion, and other means of consumption smoothing. - Investment starts to recover after two years in FCS, compared to one year in non-FCS, possibly reflecting slower post-disaster reconstruction amid limited resources and capacity. - Extreme weather events reduce exports and widen the current account deficit; the sharp drop in exports may be related to lack of diversification (for example, reliance on agricultural exports) and weak infrastructure. ### Methods, data, and analytical approach (summary) - To mitigate severe data constraints for FCS, the note uses innovative data and methodologies including georeferencing and geospatial analysis to build and analyze high-frequency and subregional datasets on FCS, in addition to country-level macroeconomic data and case studies. - Definitions and data sources referenced: - FCS definition based on the World Bank’s country list from FY2006 to FY2024; the IMF adopted the World Bank methodology, thresholds, and criteria as part of the FCS strategy approved in 2022. - Exposure index constructed from Notre Dame Global Adaptation Initiative exposure subindex, INFORM Risk Index natural hazard exposure subindex, UNDP Human Climate Horizons projected increase in human deaths, and the World Risk Index climate exposure subindex. - Adaptive capacity index derived from World Risk Index coping and adaptive capacity subindices and INFORM Risk Index’s coping capacity subindex. - Extreme weather event counts sourced from Centre for Research on the Epidemiology of Disasters, Emergency Events Database. - Note on scope: - The empirical focus in the note on extreme weather events excludes earthquakes and epidemics; a "large disaster year" is defined by deaths and affected people as a proportion of the country’s population. - Coverage of extreme weather events in FCS is likely incomplete given severe data constraints. *Source: clnea2023001 - Introduction (IMF Staff Climate Notes, 2023).* ### 4. Exports ### 4. Exports ### Longer-term Effects of Worsening Climate Conditions - Droughts significantly impact economic activity in FCS over extended periods and are expected to worsen with climate change. - Worsening drought conditions—proxied by the declines of the Standardized Precipitation-Evapotranspiration Index (SPEI) from its long-term trend—have a significant long-term impact on FCS, while no significant long-term impact is found for non-FCS. - Estimated long-term GDP effects in FCS: - Drought conditions would cut real GDP per capita growth every year by 0.2 percentage point in a low emissions scenario (RCP 2.6). - Drought conditions would cut real GDP per capita growth every year by 0.4 percentage point in a high emissions scenario (RCP 8.5). - By 2060, real GDP per capita in FCS would be 5 percent lower in the high emissions scenario compared to the low emissions scenario. - Key channels through which droughts affect long-term growth in FCS: - Lower crop productivity. - Reduced food production. - Weaker investment. - The effects of changes in the SPEI are not symmetric: no significant results were found for the longer-term impact of increases in SPEI levels from its long-term trend (associated with heavy rainfall, floods, and storms). - Worsening drought conditions and food security: - Food production in fragile states is found to be two times more sensitive to drought conditions over the longer term than other countries. - Worsening drought conditions are associated with persistent upward pressure on inflation in FCS, where food represents a large share of consumption. - Data: food represents 42 percent of total consumer expenditure for the median FCS, compared to 23 percent in other countries. - The confluence of lower food production and higher prices in a high emissions scenario would push 2 percentage points more of fragile states’ population—about 50 million more people—into hunger by 2060. - Additional factors that could aggravate climate effects on GDP include a compression of total factor productivity. ### Humanitarian and Conflict Impact of Climate Shocks - Humanitarian casualties and displacement: - Centre for Research on the Epidemiology of Disasters, Emergency Events Database data show that three times as many people in FCS are affected every year by extreme weather events than in other countries. - Following extreme weather events, the share of undernourished population increases significantly in FCS from already high levels. - Close to 10 percent of internal displacement in FCS is directly linked to disasters, with more than twice the share of the population being displaced in FCS than in other countries. - Close to 95 percent of refugees, 86 percent of internally displaced people, and 20 percent of migrants globally have originated in FCS countries (2022 aggregates reported in source figures). - Climate shocks and conflict intensity: - Empirical analysis using georeferenced weather and conflict data for 106 countries over 2013–2022 does not find that climate shocks affect the onset of new conflict, but shows that, where conflict exists, climate shocks exacerbate its intensity. - Estimated effects: - A 1 percent increase in temperature is associated with a 0.1 percent increase in conflict intensity (number of conflict-related deaths/population). - Annualized projection: in a high emissions scenario and all else equal, by 2060 conflict deaths as a share of the population for a median FCS could increase by 8.5 percent, and up to 14 percent for countries facing an extreme temperature increase. - Similarly, countries facing an extreme drop in precipitation could see conflict intensity increase of up to 8 percent. - Note: these results underestimate destructive impacts of conflict that do not result in higher deaths. - Identified pathways linking climate change to conflict: - Resource scarcity (freshwater, arable land, forests, fisheries). - Food security deterioration. - Displacement and migration pressures. - Economic shocks from damaged infrastructure and reduced productivity. ### Climate Vulnerability in Agriculture - Agriculture dependence and vulnerability in FCS: - In 2021, the value added of the agriculture sector represented 22 percent of GDP in FCS, compared to 6 percent in non-FCS. - In 2019, 43 percent of employment in FCS was in agriculture, compared to 14 percent for non-FCS. - Agriculture in FCS is highly sensitive to variations in temperature and precipitation and is one of the sectors most vulnerable to climate shocks. - Irrigation and rainfed farming: - AQUASTAT figures: in FCS only 3 percent of cultivated areas are equipped for irrigation, compared to 11 percent in non-FCS. - Novel satellite-based analysis of farm “greenness” (proxy for agricultural production) across 30 FCS over 1984–2021 shows: - Rainfed farms in FCS are significantly affected by volatility in rainfall and groundwater and by the frequency of floods. - Irrigated farms are not significantly affected by these climate variables when irrigation functions effectively. - Rainfed farms stand to lose 11 percent of their vegetation when a rainy season disappoints (that is, precipitation falls by one standard deviation). - The results underscore that lack of irrigation infrastructure and heavy reliance on rainfed farms make agricultural production and the entire economy more vulnerable to climate shocks in FCS. ### Fragility Exacerbates Climate Vulnerability of Agriculture in FCS - Irrigation performance and unused capacity: - Even when irrigation exists in FCS, many schemes underperform and substantial irrigated farmlands are left unused. - Estimates suggest that 43 percent of area of an irrigated farmland is typically unused for the average of 19 irrigation schemes in FCS. - Performance patterns: - Well-performing irrigation schemes show farmland vegetation stable and uncorrelated to rainfall levels (example: Lebanon). - Poorly performing irrigation schemes make farmland vegetation highly correlated with climate variables and thus vulnerable (example: Lake Assad scheme in Syria). - Rainfed farms tend to show an inverted U-curve where farmland vegetation drops with little or heavy rainfalls (example: West Bank and Gaza). *Source: Diallo and Lee (forthcoming); Annex 2 on methodology; figures and empirical results as presented in the source content.* ### 1. West Bank and Gaza: Gaza Strip ### 1. West Bank and Gaza: Gaza Strip ### Case-study findings on irrigation, vegetation, and fragility - Case studies illustrate how different sources of fragility amplify the impact of climate shocks by impairing scarce irrigation infrastructure and destabilizing agricultural production (Figure 17, Annex 3). - Damage and abandonment of irrigation systems by conflict: - Conflicts wreck irrigation systems through the direct impact of battles or the displacement of farmers that impairs their maintenance, increasing the areas’ vulnerability to climate shocks. - Example: In northern Iraq, the Jazeera irrigation scheme and other vast farmlands almost completely disappeared when the area was occupied by the Islamic State in Iraq and Syria, which actively destroyed irrigation infrastructure in their battles. - Example: Mali’s largest irrigation scheme (Office du Niger) has been underperforming significantly since the security crisis in 2012. Vegetation losses have been linked to increased flooding (including because of a deterioration of the drainage system), as farmers fled the area due to security risks. - Inadequate maintenance because of lack of resources and capacity: - Gravity surface irrigation, the most common type of irrigation system in FCS, requires constant maintenance to dredge canals and drainages and repair dikes and pumping equipment. - Lack of maintenance is cited as a main reason why irrigation in FCS is severely underperforming and underutilized. - Examples: Mozambique’s Limpopo River plain remained highly vulnerable to climate shocks following major floods in 2000; rehabilitation projects in the 2010s supported by development partners increased resilience. In Sudan, the Gezira irrigation scheme performs poorly mainly due to accumulation of sediment and clogging of canals and pipes, making farmland vegetation sharply correlated to precipitation. In South Sudan, the Aweil irrigation scheme was abandoned in absence of maintenance, which was last undertaken in the 1980s. - Unworkable projects or policies because of poor design, planning, or implementation: - Improperly designed irrigation projects can become unusable, with problems of water distribution (over- and underirrigation) and quality (leaching of pollutants), making farmlands more sensitive to climate shocks. - Examples: Ethiopia’s Lower Awash Plain irrigation scheme lacked proper drainage to prevent salinity hazards; an expansion in the late 2000s exacerbated salinity hazards and farmland vegetation became smaller than before the expansion. Libya’s Great Man-Made River (GMMR) project was unable to provide stable water supply to the irrigation scheme in the Benghazi area and caused it to become more sensitive to local groundwater levels than nearby farms using makeshift wells. - Policy examples: Land reforms and input subsidies can either reduce or aggravate sensitivity to climate shocks depending on design and implementation. Chen and others (2017) find land reforms in Ethiopia after the 2000s facilitated rentals, reduced resource misallocation, and increased productivity—unlike land policies between the 1970s and 1990s that included abrupt expropriation and frequent redistributions. In Zimbabwe, the Fast-Track Land Reform Program initially damaged agriculture production, although its redesign supported an increase in tobacco production. Studies on input subsidies highlight effectiveness depends on coverage and implementation (Gignoux and others 2022 on Haiti 2014 subsidy program; Theriault, Smale, and Haider 2017 on the fertilizer subsidy program in Burkina Faso). - Weak oversight and governance over projects: - Alternative financing schemes (public-private partnerships or state-owned enterprises [SOEs]) are used by resource-constrained countries but have often faced inadequate oversight and governance. - Examples: Palm plantations in the Central African Republic and the Democratic Republic of Congo show private sector or SOE involvement does not guarantee adequate investment to deliver positive outcomes. ### Empirical indicators and figures cited - Figure annotations and metrics preserved from source: - R² = 0.0653 - R² = 0.2045 - R² = 0.1989 - Time-series and event mentions preserved: - Major floods in 2000 (Mozambique Limpopo River plain). - Maintenance in Aweil last undertaken in the 1980s. - Security crisis in 2012 (Mali Office du Niger). - Expansion in the late 2000s (Ethiopia Lower Awash Plain). ### Macro-critical policies for climate adaptation in FCS - Urgency and approach: - It is urgent that FCS implement policies for climate adaptation. - FCS need adaptation policies that both facilitate immediate response to climate shocks and build climate resilience over time. - Adaptation policies must follow a multidimensional approach that is carefully prioritized and tailored to specific contexts and local communities and that supports conflict prevention and resolution. ### Policies to facilitate immediate response to climate shocks - Build buffers and strengthen institutional capacity to facilitate robust emergency responses: - Empirical evidence shows countries with larger fiscal buffers—higher fiscal balance and lower public debt—see a faster recovery from extreme weather events (Figure 18, Annex 2). - It takes three years for FCS with a relatively higher reserves-to-imports ratio to return to predisaster incomes, compared to five years for FCS with a relatively low ratio. - Ex ante instruments recommended: contingent budget lines and prearranged contingent loans (for example, from international financial institutions that disburse immediately after disasters). - FCS need to improve spending efficiency and strengthen public investment management (Aydin and others 2022). - Macroeconomic policies should be supported by frameworks, including a medium-term fiscal framework, that reflect adaptation policies and climate risk mitigation and preparedness (Duenwald and others 2022; IMF 2016). - Strengthen the social safety net to protect the most vulnerable: - A stronger social safety net in FCS would help vulnerable households cope with climate shocks in contexts of high informality and poverty. - Social safety nets require well-targeted programs with efficient delivery systems that can be scaled up rapidly when a disaster strikes, and wound down once the emergency subsides. - Transfer disaster risk through sovereign insurance, where cost-effective: - Most FCS are unable to adequately self-insure against disasters, especially larger ones, by building policy buffers and resilient infrastructure. - Risk transfer through sovereign insurance is an important tool for financing disaster risks. - Fragile states should seek to transfer risk to regional insurance pools (where these exist) because these facilities are more affordable than market insurance due to risk pooling and most provide protection against extreme climate-related disasters such as droughts and floods. - Regional facilities set up with World Bank assistance exist in the Caribbean, Africa, the Pacific, and Southeast Asia; several fragile states are insured members (Burkina Faso, Mali, the Marshall Islands, Niger) or eligible members (Chad, Ethiopia, Micronesia, Papua New Guinea, Solomon Islands, Timor-Leste, Tuvalu, Zimbabwe). - The level of insurance remains small due to cost; donor grants could help achieve a more optimal level of protection. *Source: clnea2023001 - 1. West Bank and Gaza: Gaza Strip (IMF Staff Climate Notes).* ### 1. Fiscal Balance 2. Public Debt 3. Foreign Reserves ### 1. Fiscal Balance 2. Public Debt 3. Foreign Reserves ### Policies to Build Climate Resilience Over Time - Embed climate-resilience into efforts to improve peace and security. - Poorly designed climate interventions can compound existing inequalities and exacerbate conflict risk (Cappelli and others 2023). - Security efforts that ignore climate adaptation miss underlying threats that worsen social tensions and feed into conflict (Global Center for Adaptation 2022). - Example: African Union emphasis on comprehensively assessing the climate, peace, and security nexus to link early warning systems and adaptation measures with violent conflict prevention (African Union 2022). - Improve governance and fight corruption, including to facilitate access to financing. - Empirical results: FCS with below average control of corruption would see lower real GDP per capita growth every year by about 0.2 percentage point in a low emissions scenario (RCP 2.6) and 0.4 percentage point in a high emissions scenario (RCP 8.5), compared to FCS with above average control of corruption. - Costs of corruption in the climate context: misdirected and poor-quality investments, slow/inadequate responses to extreme weather events, loss and misuse of revenues needed to build climate resilience, environmental damage and resource overuse. - Relevant governance areas: fiscal governance, rule of law, market regulation (IMF 2018). - Upgrading governance frameworks can reduce perceptions of corruption and broaden access to climate financing. - Develop climate-smart agriculture to build resilience and reduce food insecurity. - Empirical results: countries with higher agricultural investment are less affected by disasters; improved irrigation reduces vulnerability to droughts and floods (Annex 2). - Recommended investments and measures: improve irrigation and drainage systems, water management, fertilizer and insecticide use, machinery, anti-erosion measures, improved seeds and livestock; establish early warning systems; broaden mobile phone availability in rural areas. - Scale up social spending and climate-resilient infrastructure investments that are carefully designed and implemented. - Regression results show higher social spending helps mitigate the adverse impact of climate change on growth (Annex 2). - Protecting human capital (health and education) is indispensable for inclusive growth and poverty reduction in fragile countries. - FCS need more and high-quality climate resilient infrastructure while enhancing infrastructure governance and institutional quality. - To keep adaptation affordable: monitor asset conditions, ensure efficient selection, execution, and maintenance of projects, and base choices on credible cost-benefit analyses (Aligishiev, Massetti, and Bellon 2022). - Incorporate climate financing throughout the budgeting process, ensure transparent procurement, and implement risk management. Initial efforts may be basic given FCS capacity constraints. - Enhance financial inclusion to encourage private investment and enable households to smooth shocks. - Financial inclusion enables climate-resilient investment, access to savings, emergency borrowing, and insurance. - Reforms should enhance financial stability and develop well-functioning financial markets (IMF 2021). - In FCS where mobile money matters, broaden mobile phone availability and strengthen supervision of mobile money operations. ### Urgent Need for International Support - Financing needs and gaps - Estimated adaptation costs for FCS: about 1.5 percent of GDP per year (Aligishiev, Massetti, and Bellon 2022), compared to 1 percent of GDP in other countries. - Corresponds to 13 percent of tax revenues in FCS compared to less than 6 percent in other countries. - Aid Atlas reports climate adaptation aid flow commitments to FCS at about US$30 billion cumulative between 2010 and 2020. - Humanitarian costs following climate-related disasters estimated at US$3.5 to US$12 billion per year (International Federation of Red Cross and Red Crescent Societies 2019), which could balloon to US$20 billion per year by 2030. - Total humanitarian aid from OECD countries reached US$25 billion in 2021, of which an estimated 80 percent was directed to FCS. - Need for grants and concessional financing - Financial support should be provided in the form of grants and concessional financing to avoid creating pressures on fiscal sustainability. - Multilateral and bilateral financing (grants and concessional financing) will be essential alongside domestic revenue mobilization and reprioritization of spending. - Facilitate access to global climate funds and tailored support - Barriers to scaling climate finance in FCS: limited donor political will for long-term climate actions, inadequate technical capacity/resources in FCS to navigate funding mechanisms, and security concerns threatening project delivery. - Varying requirements across climate funds create bottlenecks (Belianska and others 2022). - FCS need well-defined climate strategies with credible, “bankable” projects linked to development strategy and macroeconomic framework (IMF 2019, 2023; Fouad and others 2021). - Climate finance funds should provide FCS-focused technical support facilities, dedicated funding windows, and application criteria tailored to FCS. - Improvements in the quality and availability of data are needed. - Support for sovereign disaster risk insurance - International partners can help recapitalize regional pools to lower reinsurance costs, increase coverage limits, and lower insurance premia for sovereigns (Cebotari and Youssef 2020). - Direct support can include temporary subsidization of premiums to encourage countries to join regional insurance pools. - Technical assistance and capacity development - FCS need technical assistance and training to strengthen capacity to absorb and spend climate finance effectively. - IMF support: tailored policy advice, financial assistance, and capacity development under the IMF’s FCS Strategy (IMF 2022a). - IMF has enhanced financial support for climate action through standard facilities, emergency financing, and the Resilience and Sustainability Facility (RSF). - As of July 5, 2023, nine RSF-supported programs have been approved, including for two FCS (Kosovo, Niger). - RSF-supported reforms examples: strengthen monitoring of climate-related spending, integrate climate risks into fiscal planning, incorporate climate issues into public investment management, and strengthen climate-related risk management for financial institutions (IMF 2022c). - Designing RSF programs for FCS should include assessment of climate vulnerability-fragility links, fiscal policies to respond efficiently to climate events, governance strengthening, and comprehensive capacity development support. ### Box 1 — Urbanization and Increased Flood Risks in Kinshasa, Democratic Republic of the Congo - Kinshasa context and exposure - World Bank (2019) estimates: population of Kinshasa increased 30-fold between 1960 and 2017. - Around 65 percent of urban population falls below the poverty line. - Between 2001 and 2021, the highest daily precipitation in Kinshasa reached 158 millimeters, and there was an 80 percent chance of occurrence of daily precipitation exceeding 99 millimeters in any given year. - Urban expansion and flood risk - Densely developed area estimated from satellite imagery: 21 square kilometers in 1994 expanded to 100 square kilometers by 2022. - Flood simulations (HEC-RAS, 24-hour simulations) using historically observed rainfall scenarios show a greater percentage of the area developed after 1994 will experience deadly inundation (exceeding two meters deep) than the area developed before 1994. - With 158 millimeters of daily rainfall (which has a 5 percent probability of occurring every year), the number of people exposed to life-threatening flood risks in these areas is estimated at 62,000 in 2022, compared to 9,000 in 1994. - Simulation rainfall scenarios (historically observed rainfalls in Kinshasa) - 4.11 millimeters/hour (99 millimeters/day with 80 percent chance in any given year) - 4.41 millimeters/hour (106 millimeters/day with 50 percent chance) - 4.70 millimeters/hour (113 millimeters/day with 25 percent chance) - 6.60 millimeters/hour (158 millimeters/day with 5 percent chance) - Population density assumption used in exposure estimate: 28,000 people per square kilometer (World Bank 2019). ### Annex 1 — Definition and Coverage of Fragile and Conflict-Affected States (FCS) - Definition and coverage - Definition of FCS is based on the World Bank Classification of Fragile and Conflict-Affected Situations from FY2006 to FY2024. - A total of 61 countries have been classified as FCS at least once during the sample period. - 17 countries have been considered FCS throughout the sample period. - As of FY2024, 39 countries are classified as FCS. - Country coverage differs across empirical analyses due to focus and data availability for each section. *Source: Diallo and Lee (forthcoming); IMF Staff Climate Notes (selected excerpts).* ### Annex Table 1.2. Typology of Fragile and Conflict-Affected States, FY2020–FY2024 ### Annex Table 1.2. Typology of Fragile and Conflict-Affected States, FY2020–FY2024 ### Classification and country lists - Source classification: World Bank Classification of Fragile and Conflict-Affected Situations. - Note on taxonomy: - Countries separated into (1) countries with high levels of institutional and social fragility and (2) countries affected by violent conflict. This distinction was introduced in the FY2020 list; classification is not available prior to FY2020. - Countries whose classification changed during FY2020–FY2024 are shown by the latest classification. - Countries classified as FCS throughout the sample period are marked with asterisk. - FCS = fragile and conflict-affected states. - Conflict (listed by income group) - Low income - Afghanistan* - Burundi* - Guinea - Burkina Faso - Chad* - Madagascar - Central African Republic* - Eritrea* - Malawi - Congo, Dem. Rep.* - Gambia, The - Sierra Leone - Ethiopia - Guinea-Bissau* - Tajikistan - Mali - Haiti* - Togo - Mozambique - Liberia - Niger - Somalia* - South Sudan* - Sudan* - Syrian Arab Republic - Yemen, Rep. - Lower middle income - Cameroon - Comoros* - Angola - Myanmar* - Congo, Rep. - Cambodia - Nigeria - Kiribati - Cote d'Ivoire - Lao PDR - Djibouti - Micronesia, Fed. Sts. - Mauritania - Papua New Guinea - Nepal - São Tomé and Príncipe - Uzbekistan - Solomon Islands* - Vanuatu - Timor-Leste - Zimbabwe* - Upper middle income - Armenia - Kosovo* - Bosnia and Herzegovina - Azerbaijan - Lebanon - Georgia - Iraq - Libya - Tonga - Ukraine - Marshall Islands - West Bank and Gaza* - Tuvalu - Venezuela, RB ### Near-term macroeconomic analysis of extreme weather events (methods and key thresholds) - Empirical method: Jorda’s local projection method to estimate response of macroeconomic variables to extreme weather events; estimated for each h = 0, …, 3 (h = 0 is the year of the shock). - Dependent variable: log of variable of interest (for example, GDP per capita). - Key binary variables: - Climate disaster dummy = 1 when a country’s annual death plus 0.3 times the number of affected people exceeds 0.01 percent of its population; = 0 otherwise. - FCS dummy indicates FCS classification. - Event selection: climate-related disasters only (drought, flood, storm); earthquakes and epidemics excluded. - Sample coverage for near-term analysis: all countries excluding advanced economies and small states (population below 1 million) for the period 2004–20; FY2006 FCS classification is assessed on 2004 data (two-year difference for other fiscal years). - Main near-term findings (qualitative): - Drought, floods, and storms have heterogeneous impacts. - Much of the near-term GDP contraction is driven by floods and storms, with storms having a large immediate impact on investment. - Floods have a deflationary impact. - Droughts adversely affect food production and push up inflation, increasing the prevalence of undernourishment. - Results robust to inclusion of different lag specifications and broadly in line with other literature. ### Policy buffers and near-term mitigation - Second-part sample: countries classified as FCS at least once between FY2006 and FY2022; sample extended to 1980–2020 for these countries. - Approach: interaction terms estimating impact of climate shocks conditional on policy buffers over a five-year horizon; results shown by countries above or below the median of the policy variable. - Key findings on policy buffers: - Fiscal buffers—as measured by a higher fiscal balance and lower public debt—are associated with a faster recovery (Annex Table 2.2). - External buffers—as measured by foreign reserves in months of imports—also matter. - Policy buffers are more important for FCS than non-FCS (Diallo and Lee, forthcoming). - Mobilizing domestic resources swiftly is crucial for FCS given limited access to financial markets. - High debt is more detrimental given low debt-carrying capacity in FCS. - Selected regression coefficient excerpts (preserved exactly as presented) - Fiscal Balance regressions (Disaster coefficients): -1.069, -2.509, -2.678, -2.289 - Disaster#Policy coefficients: 1.235, 1.130, 1.095, 2.073 - Public Debt regressions (Disaster coefficients): 0.279, 0.156, 0.0967, 0.861 - Disaster#Policy coefficients (Public Debt): -1.427, -2.130, -2.583, -2.557 - Reserves (Imports) regressions (Disaster coefficients): -0.482, -1.417, -1.912, -1.242 - Disaster#Policy coefficients (Reserves): 0.765, -0.400, 0.105, 1.079 - Agricultural Investment regressions (Disaster coefficients): -0.286, -1.731, -1.431, 1.225 - Disaster#Policy (Agricultural Investment): 0.940, 1.649, 0.395, -0.322 - Observations and groups are reported (examples preserved): Observations 743 697 652 605; Number of groups 47 47 47 47; Observations 613 613 580 541; Number of groups 43 43 43 43; Observations 724 680 637 592; Number of groups 45 45 45 45; Observations 558 530 499 465; Number of groups 37 37 37 37; Observations 184 184 184 184; Number of groups 28 28 28 28. ### Variable medians used in near-term heterogeneity analysis (Annex Table 2.1) - Primary fiscal balance (Overall balance excluding net interest payment, percent of GDP (IMF WEO)) - Median of the Total Sample: –2.3 - Median of “High” Group: –0.1 - Median of “Low” Group: –4.7 - Public debt (General government gross debt, percent of GDP (IMF WEO)) - Median of the Total Sample: 40.7 - Median of “High” Group: 70.9 - Median of “Low” Group: 25.2 - Agricultural investment (Agricultural expenditures, percent of agricultural GDP (FAOSTAT)) - Median of the Total Sample: 3.5 - Median of “High” Group: 6.8 - Median of “Low” Group: 1.6 - Foreign reserves (Foreign exchange reserves, in months of imports (IMF IFS)) - Median of the Total Sample: 3.2 - Median of “High” Group: 5.4 - Median of “Low” Group: 1.6 - Trade openness (Size of exports and imports in relation to GDP (WDI)) - Median of the Total Sample: 62.7 - Median of “High” Group: 91.4 - Median of “Low” Group: 45.0 ### Longer-term macroeconomic impact of drought (methods and quantified effects) - Empirical method: dynamic panel autoregressive distributed lag growth model (modification of Dell, Jones, and Olken (2012)); climate variables expressed as deviations from long-term trend using Hodrick-Prescott filter following Kahn and others (2019). - Sample: 159 developing and developed countries over 1975–2018 (excluding small island states). - Drought proxy: deviations of the Standardized Precipitation-Evapotranspiration Index (SPEI) from its long-term trend. - SPEI range and classification preserved exactly: - SPEI varies between +5 and –5 and classifies soil moisture conditions as follows: non-drought (SPEI > –0.5), mild drought (–1 < SPEI < –0.5), moderate drought (–1.5 < SPEI < –1), severe drought (–2 < SPEI < –1.5), and extreme drought (SPEI < –2). - Long-term quantified impacts (point estimates summarized from Annex Table 2.4): - Growth rates of real GDP per capita, investment, and crop yields in FCS would decline by about 0.2 percentage point annually due to the projected mean deterioration in drought conditions implied by the Intergovernmental Panel on Climate Change low-emission scenario (RCP 2.6). - Average food production would be lower by about 8 percent. - Average inflation would be higher by close to 1 percentage point. - The share of food imports in total imports and undernourished persons in the total population would increase (no numeric bounds provided in the excerpt). - Asymmetry in SPEI effects: - Positive deviations of SPEI (wetter conditions, heavy rainfall) produce statistically insignificant longer-term growth effects in this analysis. - Possible reasons noted: floods and storms may be more localized and shorter in duration; reconstruction after floods and storms may offset losses; benefits of flooding for recessionary agriculture in FCS. ### Variables, policy interactions, and mitigation channels (Annex Table 2.3 and model specification) - Key dependent variables defined for long-term analysis (preserved terminology): - Real per capita GDP growth: Annual percent change in real per capita GDP (IMF WEO). - Inflation: Annual percent change in end-year consumer price index (IMF WEO), winsorized to exclude extreme inflation and deflation (above 100 percent). - Investment growth: Log difference of capital stock at constant 2017 national prices multiplied by 100 (Penn World Table). - Crop yield change: Log difference of cereal crop yield index smoothed using the Hodrick-Prescott filter; multiplied by 100 (World Development Indicators). - Food production: Natural logarithm of food production index (World Development Indicators). - Undernourishment: Annual difference in undernourished people as a percentage of the total population (World Development Indicators). - Food imports: Annual difference in food imports as a percentage of total imports (World Development Indicators). - Independent variables and transformations: - Drought = (SPEI – SPEI_t)*100 where SPEI_t is trend from Hodrick-Prescott filter (SPEI database). - Primary balance to GDP: Average annual difference in the general government primary balance as a percentage of nominal GDP (World Development Indicators). - Public debt to GDP: Government gross debt as a percentage of nominal GDP (IMF WEO). - Water insecurity index: Water security score on the ND-GAIN Global Adaptation Index scaled between 0 (low) and 1 (high). - Social expenditure to GDP, Trade to GDP, Regulatory quality index, Control of corruption index as defined in Annex Table 2.3. - Categorical policy variables used in interactions (thresholds set to FCS averages, preserved exactly): - Public debt = unity if public debt to GDP exceeds 60 percent (the FCS average). - Social expenditure = unity if average social expenditure exceeds 1 percent of GDP (the FCS average). - Trade openness = unity if the sum of imports and exports exceeds 73 percent of GDP (the FCS average). - Water insecurity = unity if the water insecurity score exceeds 0.37 (the FCS average). - Model for policy interactions: drought variable interacted with FCS dummy and with policy categorical variables to gauge marginal effects. ### Interpretation and literature context - Results consistent with recent literature finding: - Droughts have stronger medium- to long-term impacts than floods on growth, especially in sub-Saharan Africa. - Other cited results: moderate to extreme droughts reduce GDP per capita growth between 0.4 and 0.9 percentage point on average (Zaveri, Damania, and Engle 2023); run-off changes reduce short-term GDP growth by 0.4–0.6 percent (Russ 2020); future warming projected to increase annual food and headline inflation by 0.9–3.2 and 0.3–1.2 percentage points per year by 2035, respectively (Kotz and others 2023). - Policy implications emphasized: - Strengthening fiscal buffers (improving primary balance, lowering public debt) and external buffers (higher foreign reserves in months of imports) can mitigate near-term macroeconomic impacts of climate shocks in FCS. - Structural policies—social expenditure, trade openness, improvements in regulatory quality and control of corruption, and addressing water insecurity—are explored as potential mitigators of long-term welfare impacts from droughts via interactions with SPEI and FCS status. - Rapid domestic resource mobilization is particularly crucial for FCS because of limited access to international financial markets; high public debt is especially damaging given low debt-carrying capacity. *Source: Annex Table 1.2 and associated annexes in the provided IMF Staff Climate Notes content.* ### Annex Table 2.4. Model Estimates: Long-term Impact of Drought ### Annex Table 2.4. Model Estimates: Long-term Impact of Drought Conditions on Fragile States ### Scenario analysis and projected macroeconomic impacts (2023–60) - Scenario framework: - High emissions scenario: RCP 8.5. - Low emissions scenario: RCP 2.6. - Projection horizon: 2023–60. - Key projected impacts in 2060 (high emissions vs low emissions): - FCS’ drought-induced per capita income loss is estimated to be about 5 percent higher. - Investment would be lower by 3.5 percent. - Food production would be lower by 7 percent. - The share of food imports in total imports would rise by 2 percent. - Inflation would be higher by 2.5 percent. - The confluence of lower food production and higher food prices would push an additional 2 percent of fragile states’ growing population—about 50 million people—into undernourishment by 2060. - Comparative context: - Results are consistent with real GDP per capita losses estimated by Kahn and others (2019) for poor countries using temperature deviations from trend. - Note on population assumption: - Population projection for the undernourishment estimate is based on the UN population growth forecast. ### Structural amplifiers and mitigation of drought impacts - Factors that amplify real GDP per capita losses in fragile and conflict-affected states (FCS): - High public debt. - Low social spending. - Low trade openness. - High water insecurity. - Weak regulatory quality. - Weak control of corruption. - Effect of structural improvements: - Structural improvements in the above indicators help mitigate the long-term impact of droughts on real GDP per capita. - These findings align with literature indicating that greater fiscal space strengthens adaptive capacity. ### Robustness checks - Alternative specifications tested include: - A shorter sample period (by about 10 years). - A sample of countries with per capita incomes below $10,000 US dollars. - A sample of FCS countries with below average humidity. - Alternative specifications of the FCS group. - Result of robustness checks: - Estimates for the various dependent variables remain overall significant and with the expected sign across the alternative specifications. ### Conflict intensity and climate links (monthly, subregional analysis 2013–22) - Data and scope: - Unit of observation: monthly at the subregional level (2,848 subregions) across 171 countries (including FCS) over 2013–22; total observations: 340,842. - Conflict data: Uppsala Georeferenced Event Dataset (best estimate of fatalities used). - Population data: Gridded Population of the World (version 4). - Climate data: Climatic Research Unit gridded Time Series (0.5-degree grid). - Nighttime lights: Visible Infrared Imaging Radiometer Suite (monthly, 2013 to present) used as proxy for economic activity. - Main empirical findings: - A 1 percent increase in temperature is associated with a 0.1 percent increase in the intensity of conflict on a monthly basis. - A 1 percent increase in precipitation is associated with a 0.02 percent decline in conflict intensity on a monthly basis. - Model specification highlights: - Dependent variable: conflict intensity measured as number of conflict-related deaths/population (log). - Climate variables included in log: temperature and precipitation. - Nighttime lights (log) used as a control for economic activity. - State and time fixed effects included; standard errors clustered at the country level. - Use of coefficients: - Coefficients are paired with high versus low emissions climate projections to extrapolate climate change impacts on conflict intensity. ### Agriculture and irrigation: rainfed versus irrigated farms in FCS - Novel farm-level dataset: - Cross-country dataset of farmland with vegetation for FCS: 35 farmlands across 30 FCS. - Composition: 19 irrigated farmlands and 16 rainfed farmlands. - Research period: 1984 to 2021 (with missing data in 1980s–1990s for sub-Saharan Africa due to older Landsat generations). - Empirical approach: - Dependent variable: FARM — annual peak area of farmland with normalized difference vegetation index (NDVI) above 0.3. - Climate variables: annual total precipitation (RAIN), frequency of flood events (FLOOD), annual peak groundwater level (GW). - Controls: conflict deaths (CONFLICT), land reform dummy (LR), irrigation expansion (EX). - Specification estimated separately for irrigated and rainfed farms with farmland and year fixed effects. - Key results (Annex Table 2.8): - Rainfed farms: coefficients of RAIN, FLOOD, and GW are statistically significant with economically meaningful signs, controlling for conflict. - Irrigated farms: none of the climate variable coefficients are statistically significant. - Interpretation: irrigated farms are less sensitive to climate shocks than rainfed farms. - Selected reported regression outputs (as presented): - Rainfed sample RAIN coefficient: 1.6338* (standard error [0.88177]). - Rainfed sample FLOOD coefficient: -14.8085** (standard error [6.71930]). - Rainfed sample GW coefficient: 2.9133* (standard error [1.60900]). - Irrigated sample RAIN coefficient: 0.12151 (standard error [0.25050]). - Irrigated sample FLOOD coefficient: -2.3109 (standard error [2.52726]). - Irrigated sample GW coefficient: -0.25172 (standard error [0.92623]). - Observations: irrigated farms 510; rainfed farms 412. - Number of farmlands: irrigated 19; rainfed 16. - R squared: irrigated 0.07830; rainfed 0.3499. - Methodological innovation: - Inclusion of groundwater and flooding measures, in addition to precipitation, to provide a more comprehensive picture of water sources for agriculture in FCS. ### Geospatial methods and NDVI-based farmland mapping - NDVI definition and thresholding: - NDVI = (NIR − RED) / (NIR + RED). - NDVI ranges from –1.0 to +1.0. - Area with NDVI higher than 0.3 is considered to have vegetation (with adjustments for palm vegetation and older Landsat generations). - Data source and resolution: - Landsat imagery from the USGS database with 30m × 30m resolution, used to generate time-series of farmland with vegetation and to separate irrigated vs rainfed farmland and non-farmlands. - Processing steps (high level): - Digitize agricultural land boundaries to exclude natural vegetation. - Update boundaries when irrigation expansions occur. - Apply NDVI thresholding to identify vegetated farmland area. ### Case study highlights: Iraq and Mali (selected findings) - Iraq (Nineveh Province, northern Iraq): - Intensive conflict with ISIS physically destroyed irrigation infrastructure and wiped out farmlands in occupied areas. - Farmland with vegetation improved and stabilized after 2003 except for 2008–09 deterioration; during ISIS occupation in 2014 farmland with vegetation almost completely disappeared. - Post-2017 recapture: farmers were largely unable to resume agriculture due to lack of financing to reconstruct infrastructure; farmland with vegetation remains almost nonexistent. - Mali (Office du Niger irrigation scheme): - Security crisis since 2012 associated with deterioration of Office du Niger irrigation scheme infrastructure and increased flooding within the scheme. - Satellite estimates: around 30 percent of the scheme was unplanted on average between 1999–2011; share of unused farms rose to around 50 percent between 2012–21. - Estimated vegetation losses appear linked to increased flooding caused by deterioration of the drainage system as farmers fled due to security risks. *Source: Tintchev (forthcoming); Rehman (forthcoming); Koshima (forthcoming) — Annex Tables and Boxes as provided.* ### Annex Figure 3.1. Mali—Office du Niger ### Annex Figure 3.1. Mali—Office du Niger ### Mali — Office du Niger: Scheme visuals and metrics - Figure: Scheme: Farmland with Vegetation (Square kilometers). - Note: Vege = vegetation. - Axis ticks shown: 0, 500, 1,000, 1,500, 2,000, 2,500 (Area of farms w vege / Total scheme area). - Legend items include: "* No data" and "*". - Figure: Scheme: Farmland with Vegetation and Floods. - Note: Vege = vegetation. - Two axes/ticks preserved exactly: - Left axis (Area of farms with vege (km2)): 0, 200, 400, 600, 800, 1,000, 1,200, 1,400, 1,600, 1,800. - Right axis (Flood occurence (days)): 0, 10, 20, 30, 40, 50. - Temporal grouping labels: Before 2011After 2012. - Lower tick series presented: 0 1 2 3 0 4 0 5 0 (as in figure annotation). ### Mozambique — Limpopo River plain (irrigated farms sensitivity) - Findings: - Irrigation from the Limpopo River is important in a semiarid climate. - Cyclone Leon-Eline in 2000 caused catastrophic floods inundating the entire lower Limpopo plain. - Major droughts in southern Africa in 2007 devastated farmland with vegetation. - In the 2010s, additional resources from international development partners were mobilized for repair and maintenance work (World Bank 2016; African Development Bank 2012). - Maintenance work increased resilience of irrigation infrastructure to flooding and drought. - When Cyclones Idai and Eloise struck in 2019 and 2021, farmland with vegetation decreased but to a much lesser extent than in 2000. ### Sudan — Gezira irrigation scheme - Background and evolution: - Scheme originally created in the 1920s and expanded from around 4,000 square kilometers to 8,000 square kilometers in the 1960s. - Findings: - Lack of proper maintenance contributed to shrinking by half of the functioning area of the Gezira irrigation scheme. - Farmland with vegetation within the scheme has been sharply correlated to precipitation, behaving like rainfed farms. - Sediment accumulation and clogging of canals and pipes were key malfunction drivers. - Sediment removal increased in the 2000s but removal works were not properly implemented and instead damaged canal systems and facilitated sediment accumulation (Osman 2015). - Farmland with vegetation stagnated at around 3,500 square kilometers in the late 2010s, half of intended coverage, notwithstanding good rainfall. - Figure: Annex Figure 3.3. Sudan—Gezira Scheme: Farmland with Vegetation and Precipitation. - Note: Vege = vegetation. - Axis ticks shown: 0, 1,000, 2,000, 3,000, 4,000, 5,000, 6,000, 7,000, 8,000, 9,000, 10,000 (Area of farms with vege (km2)). - Precipitation axis ticks shown: 0, 200, 400, 600, 800, 1,000 (Precipitation (mm)). ### Ethiopia — Lower Awash Plain irrigation scheme - Findings: - Around 1,600 square kilometers of irrigated lands exist in the Awash River Basin. - Drainage is crucial in arid areas to prevent saline groundwater from rising and contaminating farmlands (Criddle and Haise 1957). - The Lower Awash Plain irrigation scheme had a poorly designed drainage system and was largely abandoned by the late 2000s due mainly to salinity hazard exacerbated by population density, wetland degradation, and administrative issues. - An expansion project in 2013 increased the scheme from around 300 square kilometers to 500 square kilometers following completion of the nearby Tendaho Dam. - After the expansion project, the farmland with vegetation shrank rapidly to a lower level than before the expansion as groundwater level increased, bringing more salts. ### South Sudan — Aweil irrigation scheme - Findings: - The Aweil scheme, a gravity irrigation system for rice using water from the Lol River, was originally built in the 1940s and expanded to around 50 square kilometers by the 1960s. - The scheme was abandoned after decades-long neglect of maintenance following the Second Sudanese Civil War starting in 1983. - Canals were buried under sediment and natural vegetation; infrastructure effectively ceased to exist and the scheme is largely abandoned except for a fraction used as rainfed farms. - Figure: Annex Figure 3.4. South Sudan—Aweil Scheme: Farmland with Vegetation (Square kilometers). - Note: Vege = vegetation. ### Libya — GMMR and Benghazi irrigation scheme - Findings: - The GMMR (Great Man-Made River) project takes water from large groundwater aquifers and distributes it through 4,000 kilometers of pipelines. - In the irrigation scheme south of Benghazi, water supply has been unstable; in the 1990s one reservoir was always empty and the other lost water frequently. - Farmland with vegetation plummeted particularly when neither reservoir had water. - A large area was abandoned particularly after the beginning of the Second Civil War in 2014. - Remaining farms became more strongly correlated to local groundwater level (measured by water storage) than groundwater-fed farms 80 kilometers north. - Figure: Annex Figure 3.5. Libya—GMMR: Farmland with Vegetation and Groundwater. - Note: GMMR = Great Man-Made River; vege = vegetation. - Time series ticks: 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020 (Area of farm with vege / Total scheme area). - R² values shown exactly: R² = 0.2236 and R² = 0.0041. - Scatter axis ticks shown: 0, 50, 100, 150, 200, 250, 300, 350, 400 (Area of farms with vege (km2)); 200, 300, 400, 500 (Water storage (kg/m2)). ### Democratic Republic of the Congo — PHC palm plantation - Findings: - Plantations et Huileries du Congo (PHC) is one of the largest and oldest palm plantations, created in the early twentieth century. - After the Congo War in the late 1990s, the foreign company struggled to keep up maintenance, with rapid decline in palm vegetation area in the 2000s. - Ownership changed in 2009 and again in 2020, but palm oil production continued on a downward trend amid governance concerns. - Figure: Annex Figure 3.6. Democratic Republic of the Congo—PHC: Area of Palm Vegetation (Square kilometers). - Note: PHC = Plantations et Huileries du Congo. ### Central African Republic — CENTRAPALM palm plantation - Findings: - Centrafricaine des Palmiers project launched in 1986 to create around 25 square kilometers of palm plantations managed by an SOE. - The SOE was unable to make investments to replace aged or damaged palm trees and upgrade machines and equipment; production declined. - In the 2000s, palm vegetation covered only around one-third of the total plantation area. - The situation deteriorated significantly with the beginning of the civil war in 2012; in recent years the plantation has been abandoned except for a small area likely cultivated by private farmers who purchased palm trees from the SOE in the 1990s. - Figure: Annex Figure 3.7. Central African Republic—CENTRAPALM: Area of Palm Vegetation (Square kilometers). - Note: CENTRAPALM: Centrafricaine des Palmiers. ### Zimbabwe — Metabeleland North Province (Fast-Track Land Reform and tobacco) - Findings: - Fast-Track Land Reform Program began in 2000; commercial farmlands were confiscated and distributed to groups without knowledge or financing to operate large-scale farming, causing near-vanishing of vast crop farms in northwest Zimbabwe in the early 2000s. - Farmlands revived following a shift to tobacco production since 2009. - Together with climate adaptation measures taken by the government, farmland in the region has become less sensitive to changes in rainfall. - Figure: Annex Figure 3.8. Zimbabwe—Metabeleland North Province: Farmland with Vegetation (Square kilometers). - Note: Vege = vegetation. - R² values shown exactly for periods: R² = 0.7719, R² = 0.6353, R² = 0.0229. - Axes ticks shown: 0, 2,000, 4,000, 6,000, 8,000, 10,000, 12,000, 14,000, 16,000, 18,000, 20,000 (Area of farms with vege (km2)); 0, 200, 400, 600, 800, 1,000, 1,200 (Precipitation (mm)). - Period group labels: 1989–99, 2000–08, 2009–21. *Source: Koshima (forthcoming). Note: Vege = vegetation; GMMR = Great Man-Made River.* ### References ### clnea2023001 - References ### Major themes represented in the references - “Climate, Conf lict and Forced Migration.” Global Environmental Change 54: 239–49. - “Human Security.” In: Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects. - Project appraisal and implementation reports on irrigation and resilience: “Project Appraisal Report: Baixo Limpopo Irrigation and Climate Resilience Pro ject.”; “Emergency Resilience Recovery Project in Mozambique.” Report PAD2115. - Large-scale irrigation schemes and performance: “Spatio-temporal Perf ormance of Larg e-Scale Gezira Irrigation Scheme, Sudan.”; “Too Big to Handle, Too Important to Abandon: Ref orming Sudan’s Gezira Scheme.” - Macro-fiscal and fiscal policy treatment of climate adaptation: “Macro-Fiscal Implications of Adaptation to Climate Change.” IMF Staff Climate Note 2022/002; “Planning and Mainstreaming Adaptation to Climate Change in Fiscal Policy.” IMF Staff Climate Note 2022/003; “How to Make the Management of Public Finances Climate Sensitive—‘Green PFM.’” How To Note 22/06. - Climate change impacts on economic activity and macroeconomy: “Climate Change and the Macro-economy: A Critical Review.” Working Paper 706; “Long-Term Macroeconomic Effects of Climate Change: A Cross-Country Analysis.” IMF Working Paper 19/215; “The Macroeconomic Consequences of Disasters.” Journal of Development Economics 88 (2): 221–31. - Climate shocks, conflict, and security: “Climate Change, Human Security and Violent Conf lict.”; “Warming Increases the Risk of Civil War in Af rica.” PNAS 106 (49): 20670–74; “Climate Change and Conf lict.” Annual Review of Political Science 22 (1): 343–60. - Climate impacts on agriculture, water, and food security: “Climate Change and Chronic Food Insecurity in Sub-Saharan Af rica.” IMF Departmental Paper 2022/016; “Irrigation and Water Use Ef f iciency in Sub-Saharan Af rica.”; “Maize Yield Response to Fertilizer under Dif fering Agro-Ecological Conditions in Burkina Faso.” - Insurance, financial protection, and public finance for disasters: “Natural Disaster Insurance for Sovereigns: Issues, Challenges and Optimality.” IMF Working Paper 20/3; “Financial Protection of the State Against Natural Disasters: A Primer.” World Bank Policy Research Working Paper 5429. - Adaptation finance and policy frameworks: “Unlocking Access to Climate Finance for Pacific Island Countries.” IMF Departmental Paper 2021/020; “Proposal to Establish a Resilience and Sustainability Trust.” IMF Policy Paper 2022/013. - Measurement, indices, and remote sensing for agricultural and climate monitoring: “USA Crop Yield Estimation with MODIS NDVI: Are Remotely Sensed Models Better Than Simple Trend Analyses?” Remote Sensing 13: 4227; “A Multiscalar Drought Index Sensitive to Global Warming: The Standardized Precipitation Evapotranspiration Index.” Journal of Climate 23 (7): 1696–1718. - Climate modeling and scenarios: “CMIP6: The Next Generation of Climate Models Explained.” Carbon Brief, December 2; “Differential Climate Impacts for Policy-Relevant Limits to Global Warming: The Case of 1.5°C and 2°C.” Earth System Dynamics 7: 327–51. ### Geographic and sectoral focus evident in citations - Sub-Saharan Africa: multiple IMF Departmental Papers, Working Papers, and research on irrigation, food insecurity, fragility, and conflict. - Middle East and Central Asia: “Feeling the Heat—Adapting to Climate Change in the Middle East and Central Asia.” IMF Departmental Paper 2022/008. - Pacific Island Countries: “Unlocking Access to Climate Finance for Pacific Island Countries.” IMF Departmental Paper 2021/020. - Country-specific IMF reports cited: “Zimbabwe: Article IV Staff Report.” IMF Country Report 2022/112; “Solomon Islands Selected Issues.” IMF Country Report 22/15; “Democratic Republic of Timor-Leste Selected Issues.” IMF Country Report 22/308. ### Types of sources and analytic approaches cited - Peer-reviewed journal articles on climate impacts, conflict, and economic production (Nature; PNAS; Journal of Political Economy; American Economic Review). - IMF working papers, departmental papers, staff climate notes, policy papers, and country reports. - World Bank policy and project documents and working papers. - Reports and working papers from other multilateral and research institutions (IPCC, FAO, Global Center for Adaptation, Brookings). - Remote sensing and applied methods publications (MODIS NDVI, Landsat case studies, drought index development). ### Recurring policy-oriented topics referenced - Integration of climate adaptation into fiscal policy and public financial management. - Building resilience to natural disasters in developing and fragile states. - Access to concessional climate finance for vulnerable regions. - Financial protection mechanisms for sovereigns against natural disasters. - Monitoring and measurement tools for climate impacts on agriculture and economies. *Climate Challenges in Fragile and Conflict-Affected States — IMF STAFF CLIMATE NOTE 2023/001* --- _Source: https://www.imf.org/-/media/files/publications/staff-climate-notes/2023/english/clnea2023001.pdf_