Global Drought Outlook

Trends, Impacts and Policies to Adapt to a Drier World

Report

17 June 2025

3. Impacts and costs of droughts

Copy link to 3. Impacts and costs of droughts

Abstract

This chapter examines the economic, social, and environmental impacts of droughts, highlighting their widespread and interconnected consequences. It shows how droughts affect agriculture while also disrupting economic sectors such as energy and transportation, leading to significant economic damage and macroeconomic instability. Beyond economic effects, the chapter also explores how extreme droughts can affect ecosystems and exacerbate social tensions. Drawing on scientific evidence and novel data analysis by the OECD, it underscores the far-reaching consequences of droughts and the need for proactive resilience strategies.

3.1. Introduction

Copy link to 3.1. Introduction

Droughts have profound and wide-ranging impacts on the environment, the economy, and society. They impose a considerable burden on vegetation and wildlife, disrupting ecosystem processes and threatening biodiversity. In the economic sphere, droughts can severely undermine the performance of water-intensive sectors such as agriculture, power generation, and fluvial transport. In regions heavily dependent on agriculture, drought-induced water scarcity weakens macroeconomic performance, disproportionately affecting the most vulnerable socioeconomic groups. Beyond economic consequences, severe drought episodes pose a threat to food security and water sustainability, which are fundamental to social well-being and economic stability. These conditions may generate discomfort and social unrest, potentially jeopardising political stability and social cohesion. Consequently, the impacts of droughts often transcend national borders, contributing to increased migration flows and fuelling conflicts.

The balance of ecosystems are intrinsically tied to water availability and thus particularly vulnerable to the effects of droughts. Droughts disrupt vegetation in critical ecosystems such as forests and wetlands, compromising their ability to capture and store carbon. In turn, this weakens their role in climate regulation. Droughts also threaten native species, often forcing them to migrate or adapt to survive. Compounding this, drought conditions can foster the proliferation of invasive species, further disrupting ecological balance and threatening biodiversity in key ecological corridors.

The adverse effects of droughts on ecosystems translate into significant economic consequences. Agriculture is particularly affected, with droughts causing substantial losses in both crop quantity and quality. These losses drive up food prices and ripple through other areas of the economy. Droughts also disrupt industrial processes that depend on water, resulting in increased production costs for critical sectors such as fluvial transport, power generation, manufacturing, and mining. Given the central role of these sectors to many economies, prolonged and intense droughts can have far-reaching economic repercussions. Macroeconomic effects include inflationary pressures, recessions, slower long-term GDP growth, job losses, and fiscal deficits as governments allocate resources to emergency relief efforts and infrastructure restoration.

Droughts affect not only tangible economic indicators but also societal cohesion and geopolitical dynamics, often in ways that are difficult to quantify. By threatening food and water security – two cornerstones of social stability – droughts exacerbate social pressures and inequalities. Prolonged droughts significantly reduce human well-being, amplify income and spatial disparities, and may force communities to relocate. These dynamics can weaken political institutions, erode social trust, and reduce civic engagement, contributing to political instability. Although scientific evidence remains inconclusive regarding the exact magnitude of these effects, growing research suggests that they are significant and may persist over time. At the international level, droughts can exacerbate competition over transboundary water resources and may contribute exacerbating cross-border migration, potentially intensifying geopolitical tensions.

The impacts and costs of droughts are projected to intensify under climate change. Rising global temperatures are expected to increase the frequency, duration, and severity of droughts in many regions, exacerbating existing vulnerabilities (see Chapter 2). As extreme drought events become more common, agricultural losses are likely to rise, food price volatility may increase, and disruptions to energy production and industrial processes will become more severe. The socioeconomic consequences will also worsen, with heightened risks of displacement, inequality, and political instability. These escalating effects underscore the urgency of integrating drought resilience into climate adaptation strategies, ensuring that policies account for the growing risks posed by climate change.

In this context, understanding and quantifying the environmental, social, and economic implications of droughts is crucial to shaping proactive and informed policy responses. Unlike rapid-onset disasters such as floods, droughts unfold slowly, presenting unique challenges for climate change mitigation and adaptation. Their extended duration and cascading effects can lead to complex, far-reaching consequences that are still not fully understood, especially when compared to the more immediate impacts of rapid-onset events. Therefore, enhancing the understanding of these impacts is essential. Such knowledge can guide the development of more targeted adaptation strategies, enabling policymakers to prioritise measures that reduce vulnerability and build long-lasting resilience (see Chapter 4). Furthermore, quantifying the economic and social impacts of droughts can inform the design of adaptation policies that allocate resources more equitably and efficiently. Finally, disseminating data on drought exposure and vulnerability can foster civic engagement, strengthen trust in institutions, and support collective efforts to address these challenges.

This chapter serves these objectives by bringing together evidence on the impacts of droughts on ecosystems (Section 3.2), the economy (Section 3.3), and society (Sections 3.3 and 3.4). Section 3.2 examines how drought-related variables, such as reduced soil moisture, lower precipitation, and increased heat stress, affect vegetation and animal biodiversity. Section 3.3 provides a deep dive into the economic costs of droughts by compiling reported losses and damages from past drought events, as documented in scientific and institutional literature. It also explores the role of key drought indicators on GDP, agricultural income, and the productivity of water-intensive sectors. Finally, Section 3.4 reviews recent literature on the societal impacts of drought-induced water scarcity, including its effects on social unrest, political stability, and international migration.

3.2. Impacts of droughts on ecosystems

Copy link to 3.2. Impacts of droughts on ecosystems

Droughts are among the most severe environmental stressors, as they disrupt ecosystems by altering precipitation patterns, soil moisture, and surface or groundwater levels. These prolonged dry periods have far-reaching consequences and can severely affect vegetation, wildlife, and water quality. This section explores the main environmental impacts of droughts, examining how droughts reshape and disrupt the delicate balance of natural systems.

3.2.1. The impact of droughts on vegetation

Droughts cause significant changes in both the lifecycle and morphology of plants. First, intense droughts can shorten the lifetime of several species by reducing their likelihood of survival during drought episodes. Second, droughts often result in a decrease in the overall size of plants. Numerous studies suggest that the effects of water scarcity on plant size are non-linear and vary considerably across species. Furthermore, as water becomes scarcer, plants tend to reallocate biomass from their stems and leaves to their roots to enhance water absorption (Eziz et al., 2017[1]).1

The longer droughts last, the more severe their impact on vegetation. Both plant biomass (i.e. their overall size) and survival rates decrease non-linearly as droughts become longer. A meta-study by Garssen, Verhoeven and Soons (2014[2]) shows that drought episodes exceeding 30 days cause significant reductions in plant size. In most of the studies they examine, at least 50% of plant biomass is lost during droughts that last between 40 and 80 days. Moreover, drought episodes longer than one month can substantially reduce the probability of plant survival, especially if drought intensity is high. For example, a plant exposed to a mild 30-day drought has 75% of the survival probability of a plant not exposed to drought conditions. Under severe drought conditions, this figure falls to 32%.

Herbaceous plants are much more sensitive to droughts than woody plants. Wilschut et al. (2022[3])2 examined the above-ground biomass (i.e. the biomass of stems and leaves) of plants exposed to droughts and those that were not. They found that the above-ground biomass of exposed plants falls short of those that were not exposed, and that reduced precipitation increases that difference. Most importantly, the study indicates that interaction effects between temperature and drought conditions are substantially stronger in herbaceous plants than in woody plants. The finding that woody plants are less vulnerable to droughts is empirically consistent with the findings of the research conducted by the OECD in the context of this report, presented below.

Soil moisture declines (see Chapter 2) may reduce plant health and biomass across all types of ecosystems. New econometric analyses conducted by the OECD for the purpose of this report show that, while croplands are most severely affected by low soil moisture, forests and wetlands are also significantly impacted.3 This finding aligns with existing literature suggesting that woody plants are less vulnerable to droughts than herbaceous plants (Wilschut et al., 2022[3]). Figure 3.1 illustrates the correlation between soil moisture anomalies and vegetation productivity across croplands, forests, and wetlands.

The effects of droughts on vegetation can vary from year to year and may persist over time. Drought impacts on vegetation productivity were found to be stronger during the period 2006-2010 and weaker in the years that followed. Vegetation levels can be influenced by soil moisture shocks that occurred up to two years prior. This pattern seems to hold especially in forests and wetlands, where vegetation cycles are long and less affected by human activity. While the delayed effects of past-year soil moisture losses are significantly weaker – approximately ten times smaller – than same-year effects, this gap narrows in forests and wetlands. In these ecosystems, the ratio of same-year to past-year effects drops to 6 in forests and 4 in wetlands, indicating greater vulnerability to persistent drought impacts. Unlike cultivated plants, which are typically harvested within a one-year period, vegetation in forests and wetlands affected by drought is more likely to remain in place and continue exhibiting stress. Soil moisture shocks that occurred two or more years prior are not found to have a significant effect on current vegetation productivity.

Figure 3.1. Soil moisture anomalies and their impacts on vegetation productivity
Copy link to Figure 3.1. Soil moisture anomalies and their impacts on vegetation productivity

In addition to soil moisture, precipitation and its variability substantially affect croplands, influencing crop growth, yield stability, and the timing of planting and harvesting (Figure 3.2). They also exacerbate drought stress when rainfall patterns become erratic.4 However, the effect of additional rainfall varies by region: it is much stronger in areas that receive insufficient rainfall, weaker in relatively wet regions, and negative in areas with high precipitation. The type of cultivated crops and their water requirements also play a key role. These findings align with several studies detecting non-linear effects of precipitation on crop volumes, particularly Damania, Desbureaux and Zaveri (2020[6]). Lastly, rainfall variability may be more important than total rainfall, as heavy precipitation events have strong negative impacts on both forest and cropland vegetation. In fact, ten days of heavy precipitation can be as damaging as a substantial reduction in soil moisture.

Figure 3.2. The relationship between precipitation and vegetation productivity
Copy link to Figure 3.2. The relationship between precipitation and vegetation productivity

Correlations between total annual rainfall and above-ground vegetation productivity

The impacts of droughts on crops are worse if combined with heat stress. For example, under extremely dry conditions, an additional degree of temperature may decrease yields in maize and wheat by more than 9% (Matiu, Ankerst and Menzel, 2017[7]). Heat stress, which often accompanies droughts, has a region-specific effect on crops, forests and wetlands. In general, croplands in North Europe are found to be more vulnerable to the exposure in temperatures over 32 $°$ C, than those in South Europe, reflecting systematic differences in the cultivated species and their heat resilience. Severe heat stress has a detrimental effect in the croplands and wetland vegetation of all regions, as well as in the forest vegetation of most of the regions.

Figure 3.3. Estimated crop losses due to soil moisture anomalies
Copy link to Figure 3.3. Estimated crop losses due to soil moisture anomalies

The expected impact of droughts on crops is substantial, with impacts on wheat, rice and maize being among the most studied. Zhang et al. (2018[9]) collect estimates from 55 and 60 primary studies on rice and wheat, respectively, which examine the effect of droughts on the above-ground biomass, height and yield of rice and wheat respectively. They find that the average loss of biomass due to a drought episode is 25% in wheat and 27.5% in rice, and that the corresponding numbers for yield loss are 25.2% and 25.4%. Controlling for drought intensity, Zhang et al. (2018[9]) show that biomass losses in both crops lie below 13% for mild drought episodes and above 34% for severe episodes. In line with the literature, OECD estimates suggest that droughts may substantially affect the productivity of almost all crops (Figure 3.3). Compared to the year in which soil moisture is at its mean level, producing in a year among the 25% driest implies an average quantity loss of 6.3%. Being among the 10%, 5% and 1% driest years gradually increases that loss to 12%, 15.5% and 21.9% respectively, with direct and profound implications for crop production and food security.

The impact of droughts on crops is largely uncertain and heterogeneous. In the 10% driest seasons, expected losses may range from 5% to 22% depending on the crop. However, the most conservative econometric estimates suggest losses between 3.6% and 15.4%, and the strongest estimates losses in the range 6.4-27.8%. The largest source of uncertainty is the volume of the drought (moisture shock). Moving from the 25% to 5% driest conditions increases the expected losses of the various crops from 2.6-11.4% (former case) to 9.1-39.3% (latter case). The far-reaching implications of droughts on agricultural sector are examined in Section 3.3.1.

To the extent that wood production resembles agriculture, the analysis conducted here for croplands could be replicated to estimate the impact of a drought episode on wood production. However, only a small part of the above ground biomass present in forests and wetlands is commercial. Rather, its primary functions relate to ecosystem service provision and carbon storage. In that sense, the analysis presented in this section could utilise primary data measuring forest and wetland biomass (e.g. per unit of area covered by forests and wetlands), as well as their variance across time. This could enable the conversion of the estimated effects, expressed in terms of standard deviations, to percentages of biomass lost due to a drought episode.

3.2.2. The impact of droughts on fauna

Droughts affect animal communities, with impacts varying widely across ecosystems and species. Reduced access to surface water directly influences reproduction rates and survival probabilities, while indirect effects typically cascade through the food chain. Bottom-up impacts begin with disruptions to vegetation (see Section 3.2.1) and extend to herbivores, omnivores, and carnivores. Additionally, drought-induced declines in water availability and food sources can lead to increased competition, habitat loss, and higher mortality rates, further destabilising ecosystems. The full consequences of a drought episode on fauna may take years to materialise, as the progression of these effects unfolds over time.

A species' sensitivity to droughts appears closely linked to its dependence on water abundance. Aquatic ecosystems are particularly vulnerable to drought conditions, with substantial declines observed in fish stocks and other aquatic fauna during prolonged dry periods. In contrast, terrestrial and arboreal species tend to exhibit greater resilience, though they remain affected by the long-term impacts of sustained droughts. For example, Bodmer et al. (2018[10]) investigated the effects of the 2010 drought on animal populations in the Amazon, focusing on terrestrial, arboreal, and aquatic species in flooded forests. Their study revealed significant declines in aquatic fauna, with fish stocks decreasing by 12% and pink river dolphins by 45%. In contrast, terrestrial and arboreal species showed no significant population declines during this period. Aquatic populations began recovering only after two years of intensive flooding, illustrating the prolonged effects of drought on water-dependent species and ecosystems.

Differences in feeding behaviours significantly influence animal sensitivity to droughts, even among closely related species. For instance, white rhinos are believed to be more vulnerable to droughts than black rhinos (Ferreira, le Roex and Greaver, 2019[11]). The key difference lies in their feeding habits: white rhinos are grazers, feeding primarily on grass and ground-level vegetation, which becomes scarce during droughts. In contrast, black rhinos are browsers, consuming leaves, shoots, and twigs from shrubs and trees, often above ground level. As a result, browsing herbivore species are better adapted to cope with drought-induced food scarcity.

Despite these figures, there remain substantial knowledge gaps on the extent to which a drought episode may affect different species. Prugh et al. (2018[12]) studied how California’s severe drought (2012–2015) affected 423 species, including arthropods, birds, reptiles, and mammals. They found that the drought is highly likely to have reduced the population of 25% of the species they study, and to have reduced the population of 4% of these species. The population changes observed in the remaining 71% of the species were not large enough to be attributed to the drought episode. Overall, there continues to be a lack of studies observing the population dynamics of multiple species before and after a drought, while controlling for factors that may also be subject to change during a drought episode. Such studies may provide important insights on the fragility of animal species under drought episodes of different duration and intensity.

3.2.3. The impact of drought on water quality and land degradation

Impacts on water quality

By reducing freshwater quantity in water bodies, droughts affect the dilution capacity of aquatic environments, facilitating the concentration of pollutants, nutrients, pathogens, salt, and heavy metals in lakes, rivers, and other freshwater bodies (Mosley, 2015[13]). For example, during the 2018 drought in Europe, the concentration of pharmaceutical residues in the Rhine and Meuse rivers increased by up to 30% (Wolff and van Vliet, 2021[14]). Similarly, the 2005-2006 drought in Salamanca (Spain) led to a significant increase in groundwater pollution levels, with a fourfold increase in water samples exceeding drinking water standards for arsenic levels (García-Prieto et al., 2012[15]). Similarly, in Germany and Poland, consecutive droughts exacerbated the impacts of industrial pollution in the Oder River, leading to severe ecological collapse in 2022 (JRC, 2023[16]).

In turn, high levels of water contamination reduce the amount of freshwater available for safe use. This was observed for example in Denmark, where high pollutant and nutrient concentrations have led to the closure of 30% of existing wells (EEA, 2017[17]). Prolonged droughts can also exacerbate salinisation in coastal aquifers, posing risks to human health, aquatic ecosystems, and the reliability of water supplies. For example, high salinity levels in the Colorado River have reduced agricultural yield and damaged infrastructure, causing USD 348 million per year in damages (Miller et al., 2024[18]). Globally, water contamination is projected to intensify water scarcity by 2050, complicating efforts to ensure water security in drought-affected regions (see Chapter 4).

Finally, drought-induced declines in freshwater levels, coupled with rising average and extreme temperatures, are also warming rivers and groundwater reserves. Reduced flow speeds also contribute to increasing river temperatures (Mosley, 2015[13]). Sixteen out of twenty studies examining river temperature changes during droughts in the United Kingdom report increases in maximum and/or average monthly water temperatures – which have risen by as much as 12°C during low-flow periods compared to normal years (White et al., 2023[19]).

Impacts on land degradation

By reducing soil moisture and affecting biodiversity and vegetation cover, drought plays a critical role in accelerating land degradation. Prolonged drought periods leave soils exposed to wind and water erosion, leading to the depletion of organic matter and essential nutrients. Between 2015 and 2019, global land degradation increased by 4%; currently, it affects more than 15% of the world’s land area, with direct impacts on 1.3. billion people (UNCCD, n.d.[20]). These processes undermine soil fertility, reduce water retention capacity, and limit the ability of land to sustain vegetation. Consequently, they exacerbate global water and food security challenges, compounding the issues discussed in Section 3.2.1.

Through these processes, drought can also facilitate desertification. Desertification arises from the combined effects of climatic factors, such as prolonged drought, and unsustainable human activities, including overgrazing, deforestation, and unsustainable land management practices. It can lead to irreversible declines in land productivity, with significant impacts on ecosystems and livelihoods. It accelerates biodiversity loss, intensifies water scarcity, and contributes to climate change by diminishing the land’s ability to sequester carbon.

3.3. The economic impacts of droughts

Copy link to 3.3. The economic impacts of droughts

Droughts impose a series of quantifiable costs on the economic system. Direct economic effects are mostly pronounced in the agricultural sector. The analysis in Section 3.2.1 indicated that precipitation and soil moisture deficits have a substantial effect on vegetation productivity. This section provides insights on how losses of plant biomass translate into reduced crop volume and agricultural income (Section 3.3.1). It also investigates the impacts of droughts on two other water-intensive sectors of the economy: fluvial transport and power generation (Section 3.3.2).

Several questions arise from observing the economic impacts from droughts. A central question is whether these impacts have a significant upward trend, or whether they remain constant or decrease, indicating effective adaptation to climate change. Another question is whether an upward trend in economic impacts is driven by a growing frequency, a growing duration or a growing intensity of drought episodes. This section provides new relevant insights by exploring the evolution of drought-related losses and damages worldwide and in the United States (US) (Section 3.3.3). The section also explores the extent to which droughts have a substantial impact on GDP (Section 3.3.4).

3.3.1. Impacts on the agricultural sector

A large body of literature suggests that the impact of reduced precipitation on crop volume is considerable (Table 3.1). Qin et al. (2023[21]) review more than 1 800 simulations from 68 modelling studies on the impact of climatic conditions on the volume of rice, maize and wheat production. Their meta-estimate from these studies is that a positive precipitation shock of 10% increases crop volume by more than 4%. Challinor et al. (2014[22]), using a similar number of primary studies and estimates, find a slighlty larger effect (above 5%). Wilcox and Makowski (2014[23]) find that a 10% increase in precipitation has an even larger effect on crop volume (7.0 - 7.5%), but their meta-analysis focuses only on wheat. The three meta-studies control for the corresponding effects of temperature. Both Wilcox and Makowski (2014[23]) and Challinor et al. (2014[22]) agree that the effect of an additional degree Celsius in average termperature is negative (-3.3% and -5.0% respectively). Another study by Troy, Kipgen and Pal (2015[24]) offers richer drought-relevant controls, such as dry-spells, precipitation intensity and maximum rainfall, but reports its effects in standard deviations.

Meta-studies agree that adaptation measures in agriculture are effective in mitigating droughts impacts. However, they widely diverge regarding the volume of this contribution: Challinor et al. (2014[22]) find that common adaptation strategies increase yields by up to 15%, while Qin et al. (2023[21]) find this effect to be much larger (64%).

Studies on the impact of drought-specific indicators on crop volumes are remarkably scarce. Kuwayama et al. (2019[25]) is one of the few empirical studies simultaneously accounting for the presence of a drought episode, its intensity and duration alongside temperature and precipitation effects. This allows for distinguishing between the effect of U.S. Drought Monitor index (which contemplates soil moisture, daily streamflow and vegetation health) and the additional effects of temperature and precipitation. The U.S. Drought Monitor is reported in five levels, D0-D4, with D0 describing a mild drought and D4 an exceptional drought (Figure 3.4).

Study	Information	Controls	Main findings
Qin et al. (2023[21])	Inputs: Meta-analysis of 68 published modelling studies, each reporting multiple results under various climate scenarios, giving rise to 1842 simulations Crops: rice, maize and wheat	Precipitation	An increase of 1% in average precipitation is associated with an increase of 0.43% in crop yield (elasticity = 0.43)
		Drought-relevant controls	An increase of 1% in maximum temperature is associated with an increase of 4.21% in crop yield (elasticity = 4.21). Effect of minimum temperature is insignificant
		Other controls	Adaptation measures increase crop volume by 64%
Wilcox and Makowski (2014[23])	Inputs: Meta-analysis of 90 simulation-based studies Crops: wheat	Precipitation	An increase of 1% in total precipitation is associated with 0.70-0.75% increase in crop yield. A decrease of 1% in total precipitation is associated with a decrease of up to 0.90% in crop yield
		Temperature	An increase of 1 °C in average temperature is associated with a decrease of crop volume by 3.3%
		Other controls	An increase of atmospheric concentration of CO₂ by 100 parts per million (ppm) is associated with an average yield increase 8%. Sowing adaptation increases yield by up to 6%
Challinor et al. (2014[22])	Inputs: Meta-analysis of 1700 published simulation-based estimates Crops: wheat, rice and maize	Precipitation	An increase of average rainfall by 1% is associated with an increase in the crop volume by 0.53%. A 1°C increase of average temperature decreases crop volume by 4.9%. An increase of atmospheric concentration of CO₂ by 100 ppm increases crop volume by 6%. Adaptations (change in variety, sowing dates, irrigation, residue management) increase simulated yields by 7-15%
Troy, Kipgen and Pal (2015[24])	Inputs: Crop yield data from the United States Department of Agriculture Spatial coverage: United States Temporal coverage: 1948-2020 Crops: corn, soy, wheat, rice	Rainfall: Dry-spells^(a), precipitation intensity^(b), max 5-day precipitation^(c), average precipitation	Substantial deviation-to-deviation^(f) relations are visualised for: dry-spells on corn, soy and spring wheat (≈-1.0), reduction in total precipitation on corn and soy (<-1.0), max 5-day precipitation on corn and soy (between 0 and +1.0). Soil moisture and drought-specific indexes are not considered
Troy, Kipgen and Pal (2015[24])		Temperature: minimum, maximum, average, heat stress^(d), heatwaves^(e)	Substantial deviation-to-deviation^(f) are visualised for heatwaves and heat stress on corn and soy (between -1 and -2)
Kuwayama et al. (2019[25])	Empirical analysis of the effect of droughts on farm income. Inputs: U.S. Drought Monitor Index Spatial coverage: United States (3 080 counties) Temporal coverage: 2001-2013 Crops: corn, soy	Drought-specific: U.S. Drought Monitor Index	Point elasticities of production with respect to exposure to drought conditions vary in the range (-0.012,-0.002), but reduce to the range (-0.002, 0.000) when rainfall and average precipitation conditions are considered. Reduced rainfall by one standard deviation reduces corn and soy production by 5.4% and 15.4% respectively^(g). Irrigation drastically reduces precipitation impacts on soy production, and renders precipitation and drought-indexes statistically insignificant on corn production
		Rainfall: Precipitation
		Temperature: days of moderate (10-30 °C) and extreme heat (30+ °C)

Notes: ^(a)Dry-spells: maximum number of days without rain; ^(b)Precipitation intensity: average precipitation in days with rain; ^(c)Maximum rainfall in a 5-day period; ^(d)Heat stress: total number of days with temperature above 25 $°$ C; ^(e)Heatwaves: number of consecutive days with temperature at least 5 $°$ C above the mean climatology; ^(f)Deviation-to-deviation estimates refer to the effect of one standard deviation in the value of $x$ , on the value of variable $y$ (measured in standard deviations); ^(g)Effect calculated using the estimates, sample means and standard deviations reported in Kuwayama et al. (2019[25]).

Source: Author’s own, based on Qin et al. (2023[21]), Wilcox and Makowski (2014[23]), Challinor et al. (2014[22]), Troy, Kipgen and Pal (2015[24]), Kuwayama et al. (2019[25]).

Figure 3.4. Agricultural production loss from drought conditions in the United States
Copy link to Figure 3.4. Agricultural production loss from drought conditions in the United States

Soil moisture shortages have a strongly negative impact on agricultural production. In dryland areas of the United States, drought conditions have reduced corn and soy production by 2.2-2.6% compared to normal years.5 If these drought conditions worsened significantly, these losses could increase to 6.9-10.2%. Under a broader definition including precipitation and temperature shocks, the historical cost of droughts increases to 9-10% of production, and the future cost under substantially worsened conditions reaches 27% of production.

Figure 3.5. Agricultural production loss from precipitation shocks
Copy link to Figure 3.5. Agricultural production loss from precipitation shocks

Positive precipitation shocks in one region cannot make up for the production loss in regions experiencing decreased precipitation (Figure 3.5). A considerable increase in precipitation is predicted to increase corn crop volume by 1.4%. A decrease of the same magnitude in precipitation is predicted to reduce corn crop volume by 5.4%. In soy, the difference is even larger: increased precipitation increases crop volume by 6.1%, while decreased precipitation decreases precipitation volume by 15.3%. Larger shocks can also generate larger asymmetries. For example, a very large increase in rainfall may increase soy crop volume by 3%, while a severe drop in rainfall can eliminate 40% of it. Consequently, the anomalies that droughts introduce to the hydrological cycle can substantially lower agricultural production. The above findings suggest that this may hold even if the water scarcity introduced during a drought episode is later offset by a period of excessive rainfall.

In areas where irrigation systems are widely available, the negative effects of droughts are substantially smaller. Droughts and precipitation shocks do not seem to significantly reduce corn production. Irrigation does not seem eliminate the negative impact of soil moisture and rainfall deficits in soy production. However, the reported effects of these deficits are substantially smaller. This is a strong indication that irrigation technologies can be particularly effective when it comes to the adaptation of the agricultural sector to climate change. The potential of irrigation and other adaptation measures is examined in Chapter 4 of this report.

3.3.2. Beyond agriculture: impacts of droughts on other economic sectors

While agriculture is often the most visible sector affected by droughts, drought impacts extend far beyond farming systems, disrupting industries that rely on water for production, cooling, or transportation. This section explores these broader economic consequences, with a particular focus on the observed and projected impacts of drought on energy systems and fluvial transport.

Impacts on energy production

Droughts have been shown to cause significant negative impacts on hydropower production. They reduce the surface water in lakes, rivers and other bodies that supply hydroelectric plants with water. Water is required to set turbines in motion and to cool down the steam. Declines in its availability during droughts can force these plants to operate at lower capacities or shut down temporarily. Economic repercussions may include temporary shocks in energy prices, especially in regions that are heavily dependent on hydroelectric power. Consequently, this may have environmental repercussions, as the excess energy demand may be met with electricity produced by fossil fuels. In the long term, repeated droughts may necessitate investments in alternative energy sources or improved water management strategies.

Observed hydrological drought episodes had notable impacts on hydropower generation. The severe drought in California (2012-2016) lowered water levels in major reservoirs, leading to a 48% decrease in hydroelectric power generation compared to the 2011-2020 average (U.S. Energy Information Administration, 2024[27]; U.S. Energy Information Administration, 2022[28]). Consequently, California had to rely more on natural gas, which implied increased electricity costs and CO₂ emissions. Similarly, Eyer and Wichman (2018[29]) find that an increase of the Palmer Drought Severity Index (PDSI) value by one standard deviation6 is predicted to cause a 27% decrease in the electricity generated by hydroelectric stations. The authors estimate the monthly social cost for each plant that experiences a one-standard deviation decrease in water availability to be USD₂₀₁₅ 330 000. Similar impacts were observed during the 2002-2003 drought in the Nordic countries, which lowered water inflows to hydropower reservoirs, causing a significant decline in hydropower production. Furthermore, Rodriguez and Madrigal (2014[30]) mention cases of water-related disruptions in the operations of hydroelectric power stations in North America, South Africa, India and Australia, suggesting that more than 50% of the world’s energy companies faced water-related business impacts.7

By impacting hydropower production, droughts threaten electricity affordability and decarbonation strategies. Gleick (2017[31]) estimates the direct economic cost on electricity users at USD₂₀₁₆ 2.45 billion and reports a 10% increase in the CO₂ generated by the state’s power plants. Eriksson, del Valle and De La Fuente (2024[32]) estimate that the replacement of hydropower by fossil fuels due to drought-induced water scarcity in Latin America and the Caribbean increased fine particulate matter by more than 5%. Between 2014 and 2017, Brazil experienced one of its worst droughts, severely impacting the functioning of hydroelectric stations. Hydropower generation, which accounts for 64% of the electricity mix (Cuartas et al., 2022[33]), was drastically reduced. This led to energy rationing and increased use of more expensive and polluting thermal power plants. Beyond hydropower, nuclear power production can also be heavily affected by drought. Linnerud, Mideksa and Eskeland (2011[34]) estimate a 2% loss in the production capacity of nuclear power plants for each degree Celsius of warming during droughts and heatwaves.

Impacts on fluvial transport

Fluvial transport is particularly sensitive to droughts. Severe droughts may lower water levels in rivers, hampering fluvial navigation in different ways. Lower water levels force boats to dock far from riverbanks, making passenger mobility and logistic operations more difficult. Routes may be adjusted and navigation may slow down in order to reduce the risk of running aground. Such adjustments lead to delays and additional costs. The Mississippi river in North America, Amazon in South America, and Rhine in Europe are among the most drought-sensitive fluvial systems. This sensitivity is due to the substantial volume of commodity trade that takes place using their waters and the strategic position of several supply chain hubs along their shores. Europe relies on 40 000 kilometres of waterways to accommodate supply chains that may not be sufficiently elastic to the choice of transport mode. The waterways of the Mississippi river in the United States are used to transport more than 450 million tonnes, according to the United States Department of Transportation (2019[35]).

Drought-related water scarcity may substantially lower fluvial trade volumes, but the magnitude is region-specific. For example, the severe drought conditions that affected the Panama Canal in recent years have forced authorities to restrict ship transit and cargo volumes, causing a 49% reduction in monthly traffic between December 2021 and January 2024 (UNCTAD, 2024[36]). The European Drought Risk Atlas (Rossi et al., 2023[37]) estimates that droughts reduce the volume of expected trade by up to 2.5% in several European countries, and up to 5% in Poland and Croatia. The same report finds that during extreme events, including droughts of very high intensity, the losses increase up to 10% of the expected trade volume in Western Europe and can reach up to 40% in Central and Eastern Europe. The authors attribute this difference to the smaller basins available in rivers traversing these countries, a morphological difference that implies more bottlenecks and therefore larger parts of the network affected.

The estimated costs associated with fluvial transport disruptions are remarkable. In Europe, droughts lowered the water levels of the Rhine River in 2018 and 2022, disrupting the voluminous trade that takes in it. Ships were authorised to sail with cargo capacity below 25%, steeply increasing the cost of fluvial trade and causing ripple effects on industrial production and the supply chain. For instance, 30 consecutive days of water level below a critical threshold (0.78 metres) is estimated to reduce freight transport by 24% over two months (Kara, van Reeken-van Wee and Swart, 2023[38]). Ademmer, Jannsen and Meuchelböck (2023[39]) estimate that 30 days of inland waterway disruptions in Rhine decrease industrial production by 1%, which roughly corresponds to a reduction of 0.3% in German GDP, i.e. more than USD₂₀₂₁ 15 billion. Therefore, it is possible that the 2018 and 2022 droughts may have had a substantially negative effect on German GDP, in particular due to their duration. The numbers are comparable to those presented for the United States. In the context of the ongoing 22-year mega-drought, disruptions in the fluvial transport system of Mississippi have been estimated to cause total losses and damages of USD 20 billion.8

3.3.3. Losses and damages

While the sectoral impacts presented in the previous sections highlight the direct costs of droughts, they represent only a part of a broader picture. Figures presented in this section may encompass direct and indirect costs due to damages on income-generating entities (e.g. crops), human health (e.g. due to water scarcity), and losses of physical capital and land (e.g. buildings due to soil erosion). Losses and damages reflect wider quantified impacts that span various aspects of economic activity with profound welfare impacts, such as production, labour supply and physical asset values. Due to the strong links between capital stock, productivity, employment and income, losses and damages bear significant relevance for the overall economic performance of a country. However, drought-induced losses and damages should not be confused with macroeconomic effects, and in particular with the negative impact that a drought may have on the level and the growth rate of GDP. The latter effects are examined in detail in Section 3.3.4.

Ad-hoc evidence and challenges in measuring and comparing drought costs

Measuring, benchmarking, and comparing damages from droughts is a challenging task. Losses and damages measured as percentages of national GDP may simply reflect the size of a country’s economy, rather than the extent of the losses. Damages expressed in direct monetary terms expressed at the national level or in a time-average manner may hide considerable impacts from events that were spatially limited or did not last for a long time. Moreover, historically aggregated or time-averaged costs in long periods may not fully reflect the toll of droughts. The European Commission reports a long-run annual cost that ranges between EUR 2 billion and EUR 9 billion (European Commission, n.d.[40]), a figure that translates to 0.014-0.062% of the European Union (EU)’s GDP.

Time-averaged costs during drought-intense years can provide a clearer picture of the economic shocks induced during an episode. For example, Wheaton et al. (2008[41]) report total economic costs of Canadian Dollar (CAD) 2.13 and 3.65 billion due to the drought that affected Canada in 2001 and 2002 respectively. These represent 0.19% and 0.31% of the Canadian GDP during those years. Ziolkowska (2016[42]) reports that the total economic cost of the drought episode of 2011 for Texas (United States) was USD 16.9 billion, or 1.26% of the state’s GDP. Felbermayr and Gröschl (2014[43]) follow disaster events and measure their intensities using geophysical and meteorological information. They estimate that the droughts in Syria (1983) and Guatemala (1998) caused a GDP per capita loss of 0.16% and 0.27%, respectively. The same study reports a high-intensity drought (above the 95^th percentile) is statistically associated with a GDP loss of 0.34%.

When expressed relative to GDP, drought costs may appear much larger in less advanced economies or countries of low population. For example, damages from drought episodes in the Caucasus and Central Asian countries, as reported by Duenwald et al. (2022[44]), all exceed 5% of GDP.

Expressing the costs relative to the surface of the affected area adds further precision to the estimates. The primary effects of droughts can be spatially concentrated in an area that is much smaller than the administrative unit in which total economic costs are measured. This is particularly true for regional episodes within large countries. For instance, the damages of 0.19-0.31% of Canadian GDP reported in Wheaton et al. (2008[41]) increase to 0.6-1.2% of the combined GDP of Alberta, Saskatchewan and Manitoba, i.e. the provinces that were mostly affected by the 2001-2002 droughts. Zooming in further in Saskatchewan, the damages climb to 1.6-2.6% of the province’s GDP. Another example of how spatial aggregation can compress substantial local effects of droughts is the 2008 episode recorded in Spain. Martin-Ortega, González-Eguino and Markandya (2012[45]) report an economy-wide cost of EUR 1.61 billion, i.e. less than 0.1% of the country’s GDP in 2008. However, almost all of the impacts were manifested around Barcelona, where the economic costs amounted to 0.5% of the regional GDP.

Systematic evidence

The analysis that follows uses systematic information (datasets) on drought losses and damages to assess:

if drought episodes become more costly with time
if an upward trend in costs could be attributed to longer and more frequent episodes
if an upward trend could be attributed to more intense episodes
whether the recorded damage of the average drought episode in an OECD country differs significantly from that in a developing country.

The following findings were obtained with analysis conducted by the OECD for the purposes of this report. All technical details are available in Tikoudis, Gabriel and Oueslati (2025[4]).

Figure 3.6. The economic costs of drought episodes
Copy link to Figure 3.6. The economic costs of drought episodes

The economic cost of drought episodes has been increasing, at least since 2000, at an annual rate that exceeds 3% and could be as high as 7.5% (Figure 3.6). With the most conservative estimates, the average drought episode today costs at least twice what it did in 2000, and in 2035 it will cost a national economy almost 40% more than it does today. Overall, no significant differences are observed between OECD and non-OECD countries. The level of losses and damages from an episode occurring in an OECD country cannot be statistically differentiated from those occurring in a non-OECD country. Furthermore, there is no significant evidence that the rate at which droughts become more costly differs between OECD and non-OECD countries.

The upward trend is driven more by a longer duration, rather than an increasing intensity of drought episodes. The estimates indicate that the duration of droughts may have increased at an annual rate of 6%, implying that drought episodes are now four times longer than they were in 2000. Current data do not allow for systematic assessment of the degree to which drought intensity evolves.

At least in the US, there is limited evidence that droughts have become more intense. The incidence of droughts in US counties has substantially increased in the last 20 years. Today, it is much more likely that a US county would be facing a drought episode that causes losses and damages, than it was 30 years ago. Figure 3.7 displays the share of US counties that reported positive losses and damages due to a drought episode at any given month in the period 1996-2024. The increasing trend translates also to longer episodes, as counties are now more likely to report losses and damages for a larger number of consecutive months. On the other hand, the level of losses and damages reported by US counties during a month of a given episode do not have a clear trend (Figure 3.8). Instead, they had a decreasing trend between 2001 and 2017, before starting to increase again. This pattern could potentially reflect a simultaneous intensification of droughts and a proliferation of adaptation measures, and its complexity renders any forward projection speculative.

Figure 3.7. The relative frequency of recorded losses and damages in the United States
Copy link to Figure 3.7. The relative frequency of recorded losses and damages in the United States

Share of US counties recording damages from a drought episode in the period 1996-2024, by month

Figure 3.8. Trends in drought-related costs in the United States
Copy link to Figure 3.8. Trends in drought-related costs in the United States

The cost-intensity of a given drought episode does not have a distinct trend

3.3.4. Macroeconomic effects

There is currently limited evidence on the direct impact drought episodes exert on GDP levels and growth rates. A study from the International Monetary Fund (Fuje et al., 2023[47]) finds that a weighted Palmer Drought Severity Index value below -2.0, which indicates a moderate-to-severe drought episode, during the three most critical crop-growing months is linked to a 1.4 percentage point slowdown in the GDP growth rate of the same year. The effect is short-term and is confined to developing economies, with no significant long-term effects observed in subsequent years. Felbermayr and Gröschl (2014[43]) show that, all else equal, if rainfall lies 50% below its long-run monthly mean for at least 3 consecutive months, per capita growth rate of GDP is expected to be lower by $1.3$ percentage points.

There is a voluminous literature on the impact of precipitation on GDP. This literature examines the continuous evolution of GDP and several climatic variables that strongly correlate with drought episodes. Apart from rainfall, these variables include mean temperature, days of heavy or extreme precipitation and heat stress. The significant advantage of this literature is that climatic variables are observed with much higher spatiotemporal frequency than drought episodes.

The impact of precipitation on national GDP is much more likely to go undetected than that of temperature. The study by Damania, Desbureaux and Zaveri (2020[6]) highlights how spatial aggregation may blur the substantial impact rainfall has on local GDP.9 The study links local GDP, temperature and precipitation data, reporting that the within-country variation of precipitation is two times that of temperature. It demonstrates that averaging precipitation at the country level may render the impact of national rainfall average on national GDP statistically insignificant. Therefore, the spatial aggregation of precipitation may be the reason why the impact of rainfall shocks on GDP goes undetected in studies at the country level. In sharp contrast, estimating the impact of local rainfall on local GDP allows a robust relationship to emerge. The key findings reported below stem from the estimates of Damania, Desbureaux and Zaveri (2020[6]), which is one of the studies overcoming this issue.

Figure 3.9. Negative precipitation shocks and their potential effects on regional GDP
Copy link to Figure 3.9. Negative precipitation shocks and their potential effects on regional GDP

A severe dry shock has a pronounced negative impact on the GDP growth rate of the affected region. A significant reduction in annual rainfall – equivalent to one standard deviation – can halve the region's GDP growth rate. If the shock is temporary, the growth rate recovers to its normal level. However, if the rainfall deficit becomes permanent, the region's GDP by 2050 could be 30% lower than it would have been under normal conditions.

Smaller rainfall shocks also have noticeable impacts. A 100-millimetre reduction in annual rainfall can lower GDP growth by 0.2 percentage points in arid zones but has minimal effect in regions with high precipitation (above 2.5 metres annually). In very wet areas, this reduction may positively impact GDP growth, especially if it reflects fewer episodes of extreme rainfall (Figure 3.9). A substantial wet shock has a positive, but smaller effect. While a substantial dry shock halves an area's GDP growth rate, a substantial wet shock of the same magnitude —one standard deviation— may increase that GDP growth rate by one third.

Rainfall deficits have a stronger negative impact on the GDP of dry and temperate regions whose economy depends on agriculture. The drier an area is, the more detrimental the effect of a decrease in annual rainfall is (Figure 3.9). Furthermore, the larger the dependence of the local GDP on agriculture, the larger the sensitivity of the GDP to negative precipitation shocks, at least in dry and temperate areas. In very humid areas, negative precipitation shocks may increase GDP, e.g. by reducing the probability of flooding. In a temperate region with annual rainfall of about 1000 millimetres, a drought episode reducing rainfall to 900 millimetres could lower GDP growth by 0.10–0.15 percentage points.

Precipitation impacts on GDP are more pronounced in developing countries. In these economies, agriculture constitutes a larger share of GDP and income, making their economic performance more dependent on climatic conditions. Damage to agricultural crops, such as reduced yields during droughts, translates directly into measurable economic losses. In contrast, developed economies have more diversified economic structures, which reduce the relative importance of agricultural shocks to overall GDP. Another reason why GDP is impacted in different ways by rainfall is the adoption of adaptation strategies, such as irrigation systems. The gap in resilience measures between developing and developed countries is significant and well documented in the work of Barrios, Bertinelli and Strobl (2010[48]). As examined earlier in this chapter and further discussed in Chapter 4, the uptake of irrigation systems mitigates the economic impacts of reduced rainfall during drought episodes, making GDP effects less detectable (see Section 3.3.1).

Dry shocks may boost the GDP of wet areas if they reduce extreme rainfall episodes. The study by Kotz, Levermann and Wenz (2022[49]) accounts for heavy and extreme precipitation. They find that significant decrease in annual rainfall (one standard deviation) reduces GDP growth rates at all levels of annual rainfall. However, the effect of fewer episodes of heavy and extreme rainfall is predominantly positive, as fewer such episodes translate to lower flooding risks.

The literature is not conclusive on how long the negative GDP effect of a dry shock may last. Berlemann and Wenzel (2018[50]) find that a negative precipitation anomaly of one standard deviation keeps GDP growth rates 0.05 to 0.15 percentage points below their baseline levels for up to 14 years. This implies that 14 years after a moderate rainfall deficit shock, the GDP level of a country whose baseline growth rate (i.e. in absence of the shock) is $2 %$ , would be 1.1% below its baseline level.

Dry shocks may be less detrimental if green water is abundant. The work by Zaveri, Damania and Engle (2023[51]) indicates that the negative impact of a dry shock in areas of high forest cover is less than 50% of the impact dry shocks can have in areas of low forest cover. This finding suggests that forests could possibly act as a natural adaptation mechanism to droughts.

3.4. Beyond the economy, beyond borders

Copy link to 3.4. Beyond the economy, beyond borders

Persistent and intense drought episodes produce far-reaching effects that extend beyond the economic domain and are often difficult to confine within national borders. They may disrupt existing migration patterns or trigger new ones, forcing populations to relocate to more resource-abundant areas. This relocation can place significant strain on resources and infrastructure in receiving regions, potentially fuelling social tensions and instability. In drought-affected areas, the social costs of fundamental needs like drinking water and food can rise sharply, creating lasting disruptions to social well-being and economic stability. In extreme cases, recurring droughts of prolonged duration and high intensity may weaken political institutions, foster instability, and contribute to internal violence and armed conflicts (OECD, 2023[52]).

Moreover, the impacts of droughts can cross borders. Competition for scarce resources such as water and arable land may intensify tensions both within and between nations. Historical evidence suggests that, in some instances, prolonged droughts have contributed to precipitate violent conflicts, as communities and countries compete for dwindling resources. In other areas, the unsustainable use of water resources has played a key role in exacerbating cross-border issues. These interconnected effects emphasise the need for policymakers to approach droughts not merely as environmental challenges but as pressing geopolitical concerns.

3.4.1. Droughts and migration flows

Migration and displacement figure among the most concerning effects of climate-related disasters. The increasing interest in climate-relevant migration is well reflected in the frequency with which the word migration appeared in the text of past IPCC assessment reports. While migration was mentioned only twice in the First Assessment Report of 1990, this number raised to 185 in the 5^th assessment report in 2014.10 The interest in the impacts of climate change on migration has also been inscribed into a voluminous scientific literature, with publications accumulating at an annual rate of 18.5% between 2003 and 2020.11 The volume of environmental migration literature is also well reflected in the various meta-studies on the field, including systematic literature reviews, meta-analyses and bibliometric studies. These meta-studies are summarised in Table 3.2.

Meta-study	Information
Black et al. (2013[53])	Qualitative synthesis of the evidence accumulated until 2013. Conceptualisation of mobility, displacement and climate change
Millock (2015[54])	Systematic literature review focusing on empirical and theoretical environmental migration studies with a strong economic component
Berlemann and Steinhardt (2017[55])	Literature review
Hoffmann et al. (2020[56])	Meta-analysis utilising 1803 estimates of climate-relevant impacts on migration from 30 studies published between 2006 and 2019 Central meta-estimate: 1.0 standard deviation change in the environmental conditions increases migration by 0.021 standard deviations Drought-relevant control variables: precipitation level, precipitation variability Relevant findings: Estimates of precipitation effects are systematically weaker than these of temperatures and rapid-onset events and precipitation anomalies (by 0.015 to 0.018 standard deviations)
Beine and Jeusette (2021[57])	Meta-analysis utilising 1355 estimates from 51 studies attempting to explain (i) why some studies do obtain significant results while others not, (ii) the probability that a study detects a direct effect and (iii) the probability to detect a significant positive displacement effect. Some of the most important findings are: (a) Studies that focus on developing countries are 19% more likely to detect a positive displacement effect;^a (b) Studies that control for rainfall levels and rainfall variability are 12-17% and 19-24% more likely to detects migration effects;^b (c) No rainfall variable makes a study more likely to detect positive migration effects (emigration);^b (d) Studies controlling for droughts are less likely by 0-5% to detect migration effects.^c

Notes: (a) Beine and Jeusette (2021[57]) Tables 5-9; (b) Beine and Jeusette (2021[57]) Table 16; (c) Beine and Jeusette (2021[57]) Table 19.

Source: Author’s own, based on Black et al. (2013[53]), Millock (2015[54]), Berlemann and Steinhardt (2017[55]), Hoffmann et al. (2020[56]) and Beine and Jeusette (2021[57]).

There are few conclusions to extract from the existing literature on the impacts of climate change on migration and displacement, with relevance to the specific role of droughts. The meta-analysis by Hoffmann et al. (2020[56]), which uses more than 1800 estimates from 30 country-level studies on the impact of various climate factors (including drought-relevant variables) on migration. Across studies, migration is reported in various forms that are not directly comparable (nominal flows, relocation probability or migration probability odds). To overcome these comparison barriers, the study standardises its primary estimates.12 The meta-analysis suggests that the average impact of climate factors on migration across studies is seemingly low, i.e. $0.021$ standard deviations. To the extent this finding is valid for droughts, it suggests that people tend to migrate in response to warmer and drier climate conditions, though such response is rather weak.

Understanding the link between droughts and migration remains challenging due to data gaps, limited bibliographic evidence and methodological diversity. Only a narrow subset of the existing literature on the impacts of climate change on migration refers to droughts (Table 3.3). A small amount of studies control for drought-specific variables, such as precipitation and soil moisture, and an even smaller subset reports statistically significant effects. Comparability across findings is further hampered by the diversity of methodological approaches. Some studies employ aggregate empirical models to estimate how conditions in origin and destination regions, such as unemployment rates or water scarcity, influence migration flows. Other studies use microdata and event-history models to isolate the migration impact of drought episodes from individual (e.g. education, age) and household (e.g. family size) characteristics. Aggregate models face additional challenges, as migration flows often alter the demographic and skill composition of a region, which they assume to be exogenous. Migration is studied at different levels, with some studies focusing on international migration, while others explore internal movements, such as rural-to-rural or rural-to-urban relocations.

Primary study	Methodological Information	Key findings
Studies controlling for *rainfall level (one-side)*¹
Backhaus, Martinez-Zarzoso and Muris (2015[58])	Gravity model estimated with data from 142 immigration-origin countries between 1995 and 2006	An 10% decrease in precipitation decreases migration flows by 0.55 percentage points
Marchiori, Maystadt and Schumacher (2012[59])	Theoretical migration model estimated with migration flows between Sub-Saharan African countries	The annual weather-induced international migration rate is estimated at 0.03%, and 53% of this (0.016%) is attributed to rainfall anomalies. Within the sampling period of 40 years (1960–2000) it accumulates to 0.64%
Barrios, Bertinelli and Strobl (2006[60])	Econometric model predicting urbanisation rates at the country level using data from the United Nations’ World Urbanisation Prospects	The elasticity of urbanisation with respect to rainfall is estimated to be between -0.3 and -0.6 in Sub-Saharan African countries
Studies controlling for *rainfall level (both sides)* ²
Bohra-Mishra, Oppenheimer and Hsiang (2014[61])	Econometric study following 7185 Indonesian households from 13 provinces for over 15 years	1% increase in precipitation (from the mean) affects the interprovincial migration rate by -1.8% to +1.6%; 1% decrease in precipitation increases the interprovincial migration rate by to 0.6-4.0%
Henry, Schoumaker and Beauchemin (2003[62])	Retrospective migration survey recording the complete locational history of more than 8500 individuals in Burkina Faso	For male population, the odds of rural-to-rural migration increase by more than 200% in annual rainfall decreases from a level exceeding 0.9 metres to a level between 0.2 and 0.5 metres. For female population, the change is smaller (70%), but the shock also decreases the odds of international migration by up to 50%
Studies controlling for *rainfall anomaly and/or rainfall variability*
Henry, Schoumaker and Beauchemin (2003[62])	Retrospective migration survey recording the complete locational history of more than 8500 individuals in Burkina Faso	For male population, the odds of rural-to-rural migration increase by almost 60% if annual rainfall falls 15% below its historical annual mean. The odds of international migration decrease by 30%. No statistically significant effects are detected for female population
Coniglio and Pesce (2015[63])	Gravity model estimated on bilateral international migration flows from emerging and developing countries toward OECD countries in the period 1990–2001	Level of rainfall does not explain migration to OECD countries. A one standard deviation increase in rainfall variability is associated with a 13.7% increase in average bilateral migration. Floods (positive rainfall anomalies) cause larger migration outflows than droughts (negative rainfall anomalies)
Mastrolillo (2016[64])	Gravity model for migration flows between 52 zones in South Africa (1997-2011)	A 10% increase in the occurrence of negative precipitation anomalies increases inter-regional migration flows by 2.2 percentage points. A 10% increase in the occurrence of positive precipitation anomalies increases inter-regional migration flows by 1.0 percentage point
Thiede, Gray and Mueller (2016[65])	Event history model estimated on 21 million observations from 25 censuses conducted in 8 South American countries	-1.0 standard deviation in precipitation increases emigration probability by 7% at young age but has no impact at mid-age or older people
Studies controlling for *Standardised Precipitation Index (SPI)*
Dallmann and Millock (2017[66])	Gravity model estimated on inter-state migration flows in India	An additional month in which a drought is at least moderate SPI<-1.0 increases inter-regional migration probability by 1.3 percentage points. An additional drought episode that is at least “moderate” increases migration probability by 1.7 percentage points. Decreasing SPI by 1.0 point within a drought episode increases migration probability by 0.8 percentage points. The mean migration rate in the sample is 0.2%
Gray and Mueller (2012[67])	Event history model estimated with panel data from Ethiopian Rural Household Survey (1 500 households, 15 rural communities, 15-years)	A severe drought almost doubles labour mobility rates, and almost triples the probability of out-of-district immigration
*Studies controlling for soil moisture levels***
Mastrolillo (2016[64])	Gravity model for migration flows between 52 zones in South Africa (1997-2011)	A 1.0 percentage point increase in soil moisture decreases migration flows by 5.0 percentage points
Mueller, Gray and Kosec (2014[68])	Event history model estimated with individual level data collected for Pakistan Panel Survey (1986–1991, 2001, 2011)	+ 1.0 standard deviation in soil moisture decreases probability of internal migration by 27-29%
Henderson, Storeygard and Deichmann (2017[69])	Econometric model of urbanisation estimated with Census data of 29 African countries collected from various sources	A decrease in the growth rate of moisture by 1.0 standard deviation increases the growth rate of urbanisation by up to 1.5 times. The finding is conditional to the presence of industries in the district
Studies with indirect drought-related controls
Feng, Krueger and Oppenheimer (2010[70])	Instrumental variables regression. Stage 2: gravity equation with crop yield explaining immigration; Stage 1: crop yield is explained by climatic variables, including rainfall	A 10% decrease in crop yield increases the emigration rates by 2%
Ezra and Kiros (2001[71])	Event history model estimated with data from 2000 Ethiopian households	All else equal, migration from Ethiopia was 30% lower in 1987-90, compared to 1984, which was characterised by a drought-induced famine

Type of control: (1) the study does not distinguish the effects of rainfall increases or decreases, (2) the impact of rainfall increments is differentiated from that of rainfall reductions.

Studies controlling for drought-related variables with insignificant or partial effects: Beine and Parsons (2015[72]), Cattaneo and Peri (2016[73]), Cai et al. (2016[74]), Drabo and Mbaye (2015[75]), Findley (1994[76]), Gröschl and Steinwachs (2017[77]), Koubi et al. (2016[78]).

Source: Author’s own, based on primary studies and back-of-the-envelope calculations detailed in Tikoudis, Gabriel and Oueslati (2025[4]).

In a dedicated global study on the impacts of drought on migration, the World Bank reports that water scarcity induces out-migration, but in a way that is highly context-specific (Zaveri, Damania and Engle, 2023[51]). The study combines microdata from 189 different census, which contain 442 million migration cases in 64 countries in the period 1960-2015, with high-granularity weather data. The analysis distinguishes the effects of rainfall deficits from those of other climate factors and individual characteristics. Rainfall shocks are found to have 40% of the explanatory power education has, and almost 10% of the explanatory power of age, which is the strongest predictor. The likelihood of migrating due to a dry shock is about five times higher for individuals with incomes above the median compared to those with lower incomes. This pattern reverses for wet shocks, highlighting potentially contrasting distributional effects of droughts and floods.

Some studies show that water scarcity may prompt displacement (OECD, 2016[79]) and hamper migration flows out of the country that experiences it (Backhaus, Martinez-Zarzoso and Muris, 2015[58]). Part of the literature attributes this finding to the existence of a poverty trap effect. While there are several arguments that may support this hypothesis, studies that confirm it may be subject to considerable methodological limitations.

The effects seem to be mostly confined to developing countries or low-income provinces depending on agriculture. For example, Barrios, Bertinelli and Strobl (2010[48]) follow rainfall and urbanisation rates in 78 countries between 1960 and 1990. Controlling for several covariates and unobserved heterogeneity, they find that a 1% increase in rainfall decreases the urbanisation rate by 0.3-0.6%. The effect is confined to sub-Saharan African countries and is not significantly different from zero for other developing countries. Currently, there is limited evidence for the effect of drought episodes on relocation trends within OECD countries, or cross-country migration flows between them.

Finally, there is some evidence that rural relocation is more responsive to droughts than international migration is. For example, from the estimates of Henry, Schoumaker and Beauchemin (2003[62]) it can be inferred that droughts in Burkina Faso affect within-country rural-to-rural migration flows by 20-300% more than they do affect migration flows to another country.

3.4.2. Droughts, environmental security and conflicts

It is well documented that climate disasters may limit accessibility to important resources, and droughts constitute no exception. During drought episodes water is scarcer and dryland becomes less arable. Conflict may arise not only because water and food become scarcer, but also because some population groups and geographic locations are in a more vulnerable position than others. Environmental migration and displacement constitutes another channel via which drought episodes may contribute to tensions and conflict in regions not directly affected by droughts (OECD, 2023[52]). Violence can take various forms, from property crime at the individual level to armed conflict between organised groups. It is theoretically possible that drought episodes negatively affect political stability and social order, and that they may fuel tensions between neighbouring nations, even if they do not affect them at the same time.

Plenty of case-study and anecdotal evidence supports the hypothesis that droughts can contribute to conflicts, though this evidence is not conclusive. A visual analysis of conflict locations reveals their spatial coincidence with dry shocks and negative anomalies in rainfall and soil moisture. However, a substantial portion of conflicts take place in wet locations. Climate Diplomacy (n.d.[80]) presents more than 138 case studies on clashes and conflicts that occurred around the world and can be attributed to climate change. Out of these 138 incidences, 105 relate to water scarcity. Further filtering reveals that 70% of the conflicts that are jointly related to climate change and water scarcity simultaneously constitute conflicts driven by local or international competition, food security concerns and migration. Table 3.4 summarises a sample of conflicts whose occurrence could be attributed, at least partially, to persistent or gradually worsening anomalies in precipitation and temperature. All explored incidences pertain to developing countries, with the majority of them having occurred or currently occurring in Sub-Saharan Africa.

The geographic concentration of drought-related conflicts in certain regions, such as Sub-Saharan Africa, poses significant challenges to conducting rigorous scientific analyses. Resource scarcity frequently arises in Sub-Saharan Africa during drought episodes, creating a consistent overlap between drought conditions and scarcity-related unrest. This overlap makes it difficult to separate the direct effects of climate disasters—such as increased mistrust in institutions and social panic—from the indirect effects driven by intensified competition for increasingly scarce resources. Moreover, resource scarcity in Sub-Saharan Africa often coincides with economic and political institutions unique to the region, which differ substantially from those found in developed economies. These overlapping factors complicate efforts to isolate the specific role droughts play in triggering conflicts. Similar challenges in detecting causal relationships also apply to other developing regions, where resource scarcity and institutional vulnerabilities often coexist with drought conditions, further complicating the analysis.

Countries involved	Period	Key findings
Local competition for resources
Mali	Since 2012	Northern Mali faces warming and shifts in rainfall patterns that have resulted in crop losses and political exclusion of local communities. By exacerbating inequalities, droughts are believed to increase support for separatist groups and recruitment for armed extremist group
Yemen	Since 1990	Numerous local conflicts occurring at various levels (individuals, tribal groups, villages). Internal migration and land sales further exacerbate conflicts, as cohabiting tribes with diverging interests vie for access to dwindling water resources
Sudan	Since 2003	The severe droughts recorded in 1970s and 1980s may have contributed to hostility between local groups and the government, which culminated with the civil war in Darfur
Nigeria, Niger, Chad, Cameroon	Since 2009	The rise of terrorist groups is preceded by a complex nexus of policy failures and the occurrence of recurrent droughts in the Chad and Niger
Niger	Since 1944	Ethnic violence at a local scale reflected is reflected upon conflicts of herders and farmers, which tend to exacerbate under drought and famine conditions
South Sudan	Since 1944	Increasing variability of rainfall in South Sudan may be linked to various conflicts between communities
Cross-border competition for resources
Ethiopia, Kenya	Since 1944	Since 1960, droughts occurring with higher frequency and intensity in the Omo-Turcana basin. Droughts intensify resource competition of between communities located at different sides of the border causing 600 deaths due to conflicts recorded between 1989 and 2011
Tajikistan, Kyrgyzstan, Uzbekistan	Since 1991	The recorded disputes may be exacerbated by upward trends in water consumption which occur simultaneously with temperature rises and a decrease in the average rainfall

Source: Author’s own, based on data from Climate Diplomacy (n.d.[80]).

As a result of these methodological barriers, the statistical evidence on the effect of climate change on conflict remains largely inconclusive.13 One of the earliest studies in the field (Burke et al., 2009[81]) predicted a significant effect of temperature (but not of rainfall) on the incidence of civil war in Sub-Saharan African countries. Two closely related studies (Hsiang, Burke and Miguel, 2013[82]; Hsiang and Burke, 2014[83]) collected estimates from various scientific studies from disciplines including psychology, archaeology, paleo-climatology, political science and econometrics. Using meta-analytic techniques, the latter study postulates that significant temperature or precipitation shocks (i.e. one standard deviation) increase interpersonal conflict by 4% and intergroup conflict by 11%. However, several other experts in the field (Buhaug et al., 2014[84]) expressed concerns about the methodological barriers that could limit the validity of the study by Hsiang, Burke and Miguel (2013[82]). The lack of consensus is well reflected in Chapter 12 of the IPCC’s Fifth Assessment Report dedicated to human security (Agder et al., 2014[85]). This stresses that: “Some of these (i.e. studies) find a weak relationship, some find no relationship, and collectively the research does not conclude that there is a strong positive relationship between warming and armed conflict”.14 In agreement with this statement, several literature reviews stress this lack of consensus. However, one of the latest reviews on the field (Koubi, 2019[86]) concludes that there is substantial consensus on the role of climate in the onset of conflict, but the impact is subject to several conditions. Notably, less developed regions that rely on agriculture are susceptible to conflict arising from climatic conditions, especially when political marginalisation is present.

Some analyses support the hypothesis that droughts contribute to conflict, but only when the necessary preconditions for the outbreak of violence are present (Table 3.5). Harari and Ferrara (2018[87]) use unique data spanning 2 700 cells ( $1 ° \times 1 °, a p p r o x . 12000 {k m}^{2})$ , 46 African countries and 24 years. Their estimates indicate that a transitory dry shock that occurred three years ago increases the probability of conflict today by 3 percentage points. If the same dry shock endured until now, its effect would grow to 6.3 percentage points. The magnitude of the estimate is large, as a location in the sample had a 17% probability of experiencing some kind of conflict at any point in time. The study predicts that a location that was in a state of conflict during the previous year has an additional probability of 12-34 percentage points of being in a state of conflict in the current year as well. Also, Harari and Ferrara (2018[87]) find significant spatial spill-over effects. A location is more likely by 2.3 to 4.5 percentage points to experience conflict if a neighbouring location is currently in a state of conflict. Most importantly, Harari and Ferrara (2018[87]) show that the effects are confined to growing seasons only, as dry shocks occurring outside the time windows that are critical for crop growth have no effect on conflict incidence. While the preconditions highlighted by Harari and Ferrara (2018[87]) are crop failure and loss in agricultural production, the study by Almer, Laurent-Lucchetti and Oechslin (2017[88]) stresses the importance of water scarcity and the presence of multiple ethnic groups as similar preconditions to conflict.

Study	Methodological information	Key findings
Indirect impact of precipitation on conflict and violence, via income
Hidalgo et al. (2010[89])	Brazil, Precipitation anomalies are used as a predictor¹ of agricultural income, which in turn affects land invasions	1.0 standard deviation in rainfall (both sides) increases land invasions by 2.6-4.1 standard deviations
Direct impact of Standardised Precipitation Evapotranspiration Index (SPEI) on conflict
Harari and Ferrara (2018[87])	African continent, 1997-2011. Spatial econometric study following conflicts with high spatial resolution data (2700 cells, 46 countries)	Depending on its duration, a 1.0 standard deviation decrease in SPEI values (i.e. a mild drought) during a growing season increases the likelihood of conflict in the subsequent two years by 0.09 to 0.26 standard deviations
Almer, Laurent-Lucchetti and Oechslin (2017[88])	African continent, 1997-2011. Spatial econometric study following conflicts with high spatial resolution data (2700 cells, 46 countries)	On average, 1.0 standard deviation decrease in SPEI values (i.e. a mild drought) increases the probability of the overall onset of conflict by 8% or 0.002 standard deviations
Abel et al. (2019[90])	World, 2006-2015. Study exploring the nexus of climatic change, conflict and asylum seeking	Droughts had limited or no impact on conflicts, except for North African and Middle East countries during 2010-2015
Direct impact of precipitation on conflict
Sofuoğlu and Ay (2020[91])	18 countries in the Middle East and North Africa region, 1985-2016. Panel causality analysis using data on temperature and precipitation	Causal relationships from temperature to political instability and conflict are detected for at least 15 of 18 countries in the sample. Causal relationships from precipitation to political instability and conflict are detected for 3 out of 18 countries in the sample
Miguel (2005[92])	Tanzania, 1992 - 2002. Study analysing differences in murders rates across villages, controlling for observed and unobserved heterogeneity across them (fixed effects)	Droughts and floods (almost) doubled the incidence of murders and attacks

Source: Author’s own, based on primary studies and back-of-the-envelope calculations detailed in Tikoudis, Gabriel and Oueslati (2025[4]).

References

[90] Abel, G. et al. (2019), “Climate, conflict and forced migration”, Global Environmental Change, Vol. 54, pp. 239-249, https://doi.org/10.1016/j.gloenvcha.2018.12.003.

[39] Ademmer, M., N. Jannsen and S. Meuchelböck (2023), “Extreme Weather Events and Economic Activity: The Case of Low Water Levels on the Rhine River”, German Economic Review, Vol. 24/2, pp. 121-144, https://doi.org/10.1515/ger-2022-0077.

[85] Agder, W. et al. (2014), “Human security”, in Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.

[88] Almer, C., J. Laurent-Lucchetti and M. Oechslin (2017), “Water scarcity and rioting: Disaggregated evidence from Sub-Saharan Africa”, Journal of Environmental Economics and Management, Vol. 86, pp. 193-209, https://doi.org/10.1016/j.jeem.2017.06.002.

[58] Backhaus, A., I. Martinez-Zarzoso and C. Muris (2015), “Do climate variations explain bilateral migration? A gravity model analysis”, IZA Journal of Migration, Vol. 4/1, p. 3, https://doi.org/10.1186/s40176-014-0026-3.

[48] Barrios, S., L. Bertinelli and E. Strobl (2010), “Trends in Rainfall and Economic Growth in Africa: A Neglected Cause of the African Growth Tragedy”, Review of Economics and Statistics, Vol. 92/2, pp. 350-366, https://doi.org/10.1162/rest.2010.11212.

[60] Barrios, S., L. Bertinelli and E. Strobl (2006), “Climatic change and rural–urban migration: The case of sub-Saharan Africa”, Journal of Urban Economics, Vol. 60/3, pp. 357-371, https://doi.org/10.1016/j.jue.2006.04.005.

[57] Beine, M. and L. Jeusette (2021), “A meta-analysis of the literature on climate change and migration”, Journal of Demographic Economics, Vol. 87/3, pp. 293-344, https://doi.org/10.1017/dem.2019.22.

[72] Beine, M. and C. Parsons (2015), “Climatic Factors as Determinants of International Migration”, The Scandinavian Journal of Economics, Vol. 117/2, pp. 723-767, https://doi.org/10.1111/sjoe.12098.

[55] Berlemann, M. and M. Steinhardt (2017), “Climate Change, Natural Disasters, and Migration—a Survey of the Empirical Evidence”, CESifo Economic Studies, Vol. 63/4, pp. 353-385, https://doi.org/10.1093/cesifo/ifx019.

[50] Berlemann, M. and D. Wenzel (2018), “Precipitation and Economic Growth”, SSRN Electronic Journal, https://doi.org/10.2139/ssrn.3275432.

[103] Bernauer, T., T. Böhmelt and V. Koubi (2012), “Environmental changes and violent conflict”, Environmental Research Letters, Vol. 7/1, p. 015601, https://doi.org/10.1088/1748-9326/7/1/015601.

[53] Black, R. et al. (2013), “Migration, immobility and displacement outcomes following extreme events”, Environmental Science & Policy, Vol. 27, pp. S32-S43, https://doi.org/10.1016/j.envsci.2012.09.001.

[10] Bodmer, R. et al. (2018), “Major shifts in Amazon wildlife populations from recent intensification of floods and drought”, Conservation Biology, Vol. 32/2, https://doi.org/10.1111/cobi.12993.

[61] Bohra-Mishra, P., M. Oppenheimer and S. Hsiang (2014), “Nonlinear permanent migration response to climatic variations but minimal response to disasters”, Proceedings of the National Academy of Sciences, Vol. 111/27, pp. 9780-9785, https://doi.org/10.1073/pnas.1317166111.

[84] Buhaug, H. et al. (2014), “One effect to rule them all? A comment on climate and conflict”, Climatic Change, Vol. 127/3-4, pp. 391-397, https://doi.org/10.1007/s10584-014-1266-1.

[81] Burke, M. et al. (2009), “Warming increases the risk of civil war in Africa”, Proceedings of the National Academy of Sciences, Vol. 106/49, pp. 20670-20674, https://doi.org/10.1073/pnas.0907998106.

[74] Cai, R. et al. (2016), “Climate variability and international migration: The importance of the agricultural linkage”, Journal of Environmental Economics and Management, Vol. 79, pp. 135-151, https://doi.org/10.1016/j.jeem.2016.06.005.

[73] Cattaneo, C. and G. Peri (2016), “The migration response to increasing temperatures”, Journal of Development Economics, Vol. 122, pp. 127-146, https://doi.org/10.1016/j.jdeveco.2016.05.004.

[22] Challinor, A. et al. (2014), “A meta-analysis of crop yield under climate change and adaptation”, Nature Climate Change, Vol. 4/4, pp. 287-291, https://doi.org/10.1038/nclimate2153.

[99] Cipollina, M., L. De Benedictis and E. Scibè (2024), “Environmental migration? A systematic review and meta-analysis of the literature”, Review of World Economics, https://doi.org/10.1007/s10290-024-00529-5.

[80] Climate Diplomacy (n.d.), Conflict mitigation and peace building, https://climate-diplomacy.org/ (accessed on 13 December 2024).

[63] Coniglio, N. and G. Pesce (2015), “Climate variability and international migration: an empirical analysis”, Environment and Development Economics, Vol. 20/4, pp. 434-468, https://doi.org/10.1017/S1355770X14000722.

[33] Cuartas, L. et al. (2022), “Recent Hydrological Droughts in Brazil and Their Impact on Hydropower Generation”, Water, Vol. 14/4, p. 601, https://doi.org/10.3390/w14040601.

[66] Dallmann, I. and K. Millock (2017), “Climate Variability and Inter-State Migration in India”, CESifo Economic Studies, Vol. 63/4, pp. 560-594, https://doi.org/10.1093/cesifo/ifx014.

[6] Damania, R., S. Desbureaux and E. Zaveri (2020), Does rainfall matter for economic growth? Evidence from global sub-national data (1990–2014), Academic Press Inc., https://doi.org/10.1016/j.jeem.2020.102335.

[75] Drabo, A. and L. Mbaye (2015), “Natural disasters, migration and education: an empirical analysis in developing countries”, Environment and Development Economics, Vol. 20/6, pp. 767-796, https://doi.org/10.1017/S1355770X14000606.

[44] Duenwald et al., C. (2022), “Feeling the Heat: Adapting to Climate Change in the Middle East and Central Asia”, Departmental Paper, No. 008, International Monetary Fund (IMF).

[5] EEA (2024), Drought impact on ecosystems in Europe, 2000-2023 | European Environment Agency’s home page, https://www.eea.europa.eu/en/analysis/maps-and-charts/drought-impact-on-ecosystems-in-data-viewers (accessed on 13 December 2024).

[17] EEA (2017), “Briefing: Water management in Europe: price and non-price approaches to water conservation”, https://www.eea.europa.eu/downloads/a8dbfc1e447f4bc6b2f3f6330e4c9bc8/1616676916/water-management-in-europe.pdf (accessed on 19 August 2024).

[32] Eriksson, M., A. del Valle and A. De La Fuente (2024), “Droughts Worsen Air Quality by Shifting Power Generation in Latin America and the Caribbean”, Policy Research Working Papers, No. 10760, World Bank, Washington, D.C.

[40] European Commission (n.d.), Water scarcity and droughts, https://environment.ec.europa.eu/topics/water/water-scarcity-and-droughts_en (accessed on 13 December 2024).

[29] Eyer, J. and C. Wichman (2018), “Does water scarcity shift the electricity generation mix toward fossil fuels? Empirical evidence from the United States”, Journal of Environmental Economics and Management, Vol. 87, pp. 224-241, https://doi.org/10.1016/j.jeem.2017.07.002.

[1] Eziz, A. et al. (2017), “Drought effect on plant biomass allocation: A meta‐analysis”, Ecology and Evolution, Vol. 7/24, pp. 11002-11010, https://doi.org/10.1002/ece3.3630.

[71] Ezra, M. and G. Kiros (2001), “Rural Out-migration in the Drought Prone Areas of Ethiopia: A Multilevel Analysis”, International Migration Review, Vol. 35/3, pp. 749-771, https://doi.org/10.1111/j.1747-7379.2001.tb00039.x.

[8] FAO (2024), Crops and livestock products, https://www.fao.org/faostat/en/#data/QCL/metadata (accessed on 13 December 2024).

[43] Felbermayr, G. and J. Gröschl (2014), “Naturally negative: The growth effects of natural disasters”, Journal of Development Economics, Vol. 111, pp. 92-106, https://doi.org/10.1016/j.jdeveco.2014.07.004.

[70] Feng, S., A. Krueger and M. Oppenheimer (2010), “Linkages among climate change, crop yields and Mexico–US cross-border migration”, Proceedings of the National Academy of Sciences, Vol. 107/32, pp. 14257-14262, https://doi.org/10.1073/pnas.1002632107.

[11] Ferreira, S., N. le Roex and C. Greaver (2019), “Species-specific drought impacts on black and white rhinoceroses”, PLoS ONE, Vol. 14/1, https://doi.org/10.1371/journal.pone.0209678.

[76] Findley, S. (1994), “Does Drought Increase Migration? A Study of Migration from Rural Mali during the 1983–1985 Drought”, International Migration Review, Vol. 28/3, pp. 539-553, https://doi.org/10.1177/019791839402800306.

[47] Fuje, H. et al. (2023), “Fiscal Impacts of Climate Disasters in Emerging Markets and Developing Economies”, IMF Working Papers, No. 261, International Monetary Fund (IMF).

[15] García-Prieto, J. et al. (2012), “Impact of drought on the ecological and chemical status of surface water and on the content of arsenic and fluoride pollutants of groundwater in the province of Salamanca (Western Spain)”, Chemistry and Ecology, Vol. 28/6, pp. 545-560, https://doi.org/10.1080/02757540.2012.686608.

[2] Garssen, A., J. Verhoeven and M. Soons (2014), “Effects of climate‐induced increases in summer drought on riparian plant species: a meta‐analysis”, Freshwater Biology, Vol. 59/5, pp. 1052-1063, https://doi.org/10.1111/fwb.12328.

[31] Gleick, P. (2017), Impacts of California’s Five-Year (2012-2016) Drought on Hydroelectricity Generation, Pacific Institute.

[67] Gray, C. and V. Mueller (2012), “Drought and Population Mobility in Rural Ethiopia”, World Development, Vol. 40/1, pp. 134-145, https://doi.org/10.1016/j.worlddev.2011.05.023.

[77] Gröschl, J. and T. Steinwachs (2017), “Do Natural Hazards Cause International Migration?*”, CESifo Economic Studies, Vol. 63/4, pp. 445-480, https://doi.org/10.1093/cesifo/ifx005.

[87] Harari, M. and E. Ferrara (2018), “Conflict, Climate, and Cells: A Disaggregated Analysis”, The Review of Economics and Statistics, Vol. 100/4, pp. 594-608, https://doi.org/10.1162/rest_a_00730.

[94] Hejazi, M. et al. (2023), “Impacts of water scarcity on agricultural production and electricity generation in the Middle East and North Africa”, Frontiers in Environmental Science, Vol. 11, https://doi.org/10.3389/fenvs.2023.1082930.

[69] Henderson, J., A. Storeygard and U. Deichmann (2017), “Has climate change driven urbanization in Africa?”, Journal of Development Economics, Vol. 124, pp. 60-82, https://doi.org/10.1016/j.jdeveco.2016.09.001.

[62] Henry, S., B. Schoumaker and C. Beauchemin (2003), “The Impact of Rainfall on the First Out-Migration: A Multi-level Event-History Analysis in Burkina Faso”, Population and Environment, Vol. 25/5, pp. 423-460, https://doi.org/10.1023/B:POEN.0000036928.17696.e8.

[89] Hidalgo, F. et al. (2010), “Economic Determinants of Land Invasions”, Review of Economics and Statistics, Vol. 92/3, pp. 505-523, https://doi.org/10.1162/REST_a_00007.

[56] Hoffmann, R. et al. (2020), “A meta-analysis of country-level studies on environmental change and migration”, Nature Climate Change, Vol. 10/10, pp. 904-912, https://doi.org/10.1038/s41558-020-0898-6.

[83] Hsiang, S. and M. Burke (2014), “Climate, conflict, and social stability: what does the evidence say?”, Climatic Change, Vol. 123/1, pp. 39-55, https://doi.org/10.1007/s10584-013-0868-3.

[82] Hsiang, S., M. Burke and E. Miguel (2013), “Quantifying the Influence of Climate on Human Conflict”, Science, Vol. 341/6151, https://doi.org/10.1126/science.1235367.

[16] JRC (2023), An EU analysis of the ecological disaster in the Oder River of 2022 - Publications Office of the EU, https://op.europa.eu/en/publication-detail/-/publication/acae85a4-ae18-11ed-8912-01aa75ed71a1/language-en? (accessed on 5 December 2024).

[38] Kara, A., J. van Reeken-van Wee and A. Swart (2023), Ripple effects - Exploring the impact of low Rhine water levels on the Dutch economy, ABN AMRO.

[49] Kotz, M., A. Levermann and L. Wenz (2022), “The effect of rainfall changes on economic production”, Nature, Vol. 601/7892, pp. 223-227, https://doi.org/10.1038/s41586-021-04283-8.

[86] Koubi, V. (2019), “Climate Change and Conflict”, Annual Review of Political Science, Vol. 22/1, pp. 343-360, https://doi.org/10.1146/annurev-polisci-050317-070830.

[78] Koubi, V. et al. (2016), “Environmental Stressors and Migration: Evidence from Vietnam”, World Development, Vol. 79, pp. 197-210, https://doi.org/10.1016/j.worlddev.2015.11.016.

[25] Kuwayama, Y. et al. (2019), “Estimating the Impact of Drought on Agriculture Using the U.S. Drought Monitor”, American Journal of Agricultural Economics, Vol. 101/1, pp. 193-210, https://doi.org/10.1093/ajae/aay037.

[95] Lada, B. (2022), Human remains found on parched shore of Mississippi River, AccuWeather.

[34] Linnerud, K., T. Mideksa and G. Eskeland (2011), “The Impact of Climate Change on Nuclear Power Supply”, The Energy Journal, Vol. 32/1, pp. 149-168, https://doi.org/10.5547/ISSN0195-6574-EJ-Vol32-No1-6.

[59] Marchiori, L., J. Maystadt and I. Schumacher (2012), “The impact of weather anomalies on migration in sub-Saharan Africa”, Journal of Environmental Economics and Management, Vol. 63/3, pp. 355-374, https://doi.org/10.1016/j.jeem.2012.02.001.

[102] Maretti, M., A. Tontodimamma and P. Biermann (2019), “Environmental and climate migrations: an overview of scientific literature using a bibliometric analysis”, International Review of Sociology, Vol. 29/2, pp. 142-158, https://doi.org/10.1080/03906701.2019.1641270.

[45] Martin-Ortega, J., M. González-Eguino and A. Markandya (2012), “The costs of drought: the 2007/2008 case of Barcelona”, Water Policy, Vol. 14/3, pp. 539-560, https://doi.org/10.2166/wp.2011.121.

[64] Mastrorillo, M. et al. (2016), “The influence of climate variability on internal migration flows in South Africa”, Global Environmental Change, Vol. 39, pp. 155-169, https://doi.org/10.1016/j.gloenvcha.2016.04.014.

[7] Matiu, M., D. Ankerst and A. Menzel (2017), “Interactions between temperature and drought in global and regional crop yield variability during 1961-2014”, PLOS ONE, Vol. 12/5, p. e0178339, https://doi.org/10.1371/journal.pone.0178339.

[92] Miguel, E. (2005), “Poverty and Witch Killing”, The Review of Economic Studies, Vol. 72/4, pp. 1153-1172, https://doi.org/10.1111/0034-6527.00365.

[100] Milán-García, J. et al. (2021), “Climate change-induced migration: a bibliometric review”, Globalization and Health, Vol. 17/1, p. 74, https://doi.org/10.1186/s12992-021-00722-3.

[18] Miller, O. et al. (2024), “Temporal variability in irrigated land and climate influences on salinity loading across the Upper Colorado River Basin, 1986-2017”, Environ. Res. Lett., Vol. 19, https://iopscience.iop.org/article/10.1088/1748-9326/ad18dd.

[54] Millock, K. (2015), “Migration and Environment”, Annual Review of Resource Economics, Vol. 7/1, pp. 35-60, https://doi.org/10.1146/annurev-resource-100814-125031.

[98] Minx, J. et al. (2017), “Learning about climate change solutions in the IPCC and beyond”, Environmental Science & Policy, Vol. 77, pp. 252-259, https://doi.org/10.1016/j.envsci.2017.05.014.

[13] Mosley, L. (2015), “Drought impacts on the water quality of freshwater systems; review and integration”, Earth-Science Reviews, Vol. 140, pp. 203-214, https://doi.org/10.1016/j.earscirev.2014.11.010.

[68] Mueller, V., C. Gray and K. Kosec (2014), “Heat stress increases long-term human migration in rural Pakistan”, Nature Climate Change, Vol. 4/3, pp. 182-185, https://doi.org/10.1038/nclimate2103.

[46] NOAA (n.d.), Storm Events Database, https://www.ncdc.noaa.gov/stormevents/.

[52] OECD (2023), “What role for migration and migrants in climate adaptation?”, https://www.oecd.org/en/publications/what-role-for-migration-and-migrants-in-climate-adaptation_72ea3fb2-en.html.

[79] OECD (2016), International Migration Outlook 2016,, OECD Publishing, Paris, https://doi.org/10.1787/migr_outlook-2016-en.

[101] Priovashini, C. and B. Mallick (2022), “A bibliometric review on the drivers of environmental migration”, Ambio, Vol. 51/1, pp. 241-252, https://doi.org/10.1007/s13280-021-01543-9.

[12] Prugh, L. et al. (2018), “Ecological winners and losers of extreme drought in California”, Nature Climate Change, Vol. 8, https://www.nature.com/articles/s41558-018-0255-1.

[21] Qin, M. et al. (2023), “Response of Wheat, Maize, and Rice to Changes in Temperature, Precipitation, CO2 Concentration, and Uncertainty Based on Crop Simulation Approaches”, Plants, Vol. 12/14, p. 2709, https://doi.org/10.3390/plants12142709.

[30] Rodriguez, D. and M. Madrigal (2014), 4 Ways Water Shortages Are Harming Energy Production, World Bank blogs: The Water Blog.

[37] Rossi, L. et al. (2023), European Drought Risk Atlas, Publications Office of the European Union, Luxembourg.

[26] Schlenker, W. and M. Roberts (2009), “Nonlinear temperature effects indicate severe damages to U.S. crop yields under climate change”, Proceedings of the National Academy of Sciences, Vol. 106/37, pp. 15594-15598, https://doi.org/10.1073/pnas.0906865106.

[97] Šedová, B., L. Čizmaziová and A. Cook (2021), “A meta-analysis of climate migration literature”, CEPA Discussion Papers, No. 29, Centter for Economic Policy Analysis, Potsdam.

[91] Sofuoğlu, E. and A. Ay (2020), “The relationship between climate change and political instability: the case of MENA countries (1985:01–2016:12)”, Environmental Science and Pollution Research, Vol. 27/12, pp. 14033-14043, https://doi.org/10.1007/s11356-020-07937-8.

[104] Theisen, O., N. Gleditsch and H. Buhaug (2013), “Is climate change a driver of armed conflict?”, Climatic Change, Vol. 117/3, pp. 613-625, https://doi.org/10.1007/s10584-012-0649-4.

[65] Thiede, B., C. Gray and V. Mueller (2016), “Climate variability and inter-provincial migration in South America, 1970–2011”, Global Environmental Change, Vol. 41, pp. 228-240, https://doi.org/10.1016/j.gloenvcha.2016.10.005.

[4] Tikoudis, I., M. Gabriel and W. Oueslati (2025), “The toll of droughts: Environmental impacts, economic costs, and international consequences”, OECD Environment Working Papers, No. 260, Vol. OECD Publishing, Paris, https://doi.org/10.1787/2f7891b9-en.

[24] Troy, T., C. Kipgen and I. Pal (2015), “The impact of climate extremes and irrigation on US crop yields”, Environmental Research Letters, Vol. 10/5, p. 054013, https://doi.org/10.1088/1748-9326/10/5/054013.

[35] U.S. Department of Transportation (2019), Commodities on the Mississippi River: 2019, BTS analysis of USACE data, Bureau of Transportation Statistics.

[27] U.S. Energy Information Administration (2024), Electric Power Monthly, EIA.

[28] U.S. Energy Information Administration (2022), Drought effects on hydroelectricity generation in western U.S. differed by region in 2021, Electri Power Monthly.

[20] UNCCD (n.d.), UNCCD Data Dashboard, https://data.unccd.int/land-degradation (accessed on 13 December 2024).

[36] UNCTAD (2024), UN Trade and Development chief visits the Panama Canal ahead of first Global Supply Chain Forum, https://unctad.org/news/un-trade-and-development-chief-visits-panama-canal-ahead-first-global-supply-chain-forum (accessed on 31 July 2024).

[41] Wheaton, E. et al. (2008), “Dry times: hard lessons from the Canadian drought of 2001 and 2002”, Canadian Geographies / Géographies canadiennes, Vol. 52/2, pp. 241-262, https://doi.org/10.1111/j.1541-0064.2008.00211.x.

[19] White, J. et al. (2023), “Drought impacts on river water temperature: A process‐based understanding from temperate climates”, Hydrological Processes, Vol. 37/10, https://doi.org/10.1002/hyp.14958.

[23] Wilcox, J. and D. Makowski (2014), “A meta-analysis of the predicted effects of climate change on wheat yields using simulation studies”, Field Crops Research, Vol. 156, pp. 180-190, https://doi.org/10.1016/j.fcr.2013.11.008.

[3] Wilschut, R. et al. (2022), “Combined effects of warming and drought on plant biomass depend on plant woodiness and community type: a meta-analysis”, Proceedings of the Royal Society B: Biological Sciences, Vol. 289/1984, https://doi.org/10.1098/rspb.2022.1178.

[14] Wolff, E. and M. van Vliet (2021), “Impact of the 2018 drought on pharmaceutical concentrations and general water quality of the Rhine and Meuse rivers”, Science of The Total Environment, Vol. 778, p. 146182, https://doi.org/10.1016/j.scitotenv.2021.146182.

[96] World Economic Forum (2023), Droughts are creating new supply chain problems. This is what you need to know.

[51] Zaveri, E., R. Damania and N. Engle (2023), “Droughts and Deficits: Summary Evidence of the Global Impact on Economic Growth.”, Water Global Practice Working Paper, World Bank, Washington, DC.

[9] Zhang, J. et al. (2018), “Effect of Drought on Agronomic Traits of Rice and Wheat: A Meta-Analysis”, International Journal of Environmental Research and Public Health, Vol. 15/5, p. 839, https://doi.org/10.3390/ijerph15050839.

[93] Zheng, X. et al. (2016), “The vulnerability of thermoelectric power generation to water scarcity in China: Current status and future scenarios for power planning and climate change”, Applied Energy, Vol. 171, pp. 444-455, https://doi.org/10.1016/j.apenergy.2016.03.040.

[42] Ziolkowska, J. (2016), “Socio-Economic Implications of Drought in the Agricultural Sector and the State Economy”, Economies, Vol. 4/4, p. 19, https://doi.org/10.3390/economies4030019.

Notes

Copy link to Notes

← 1. All three channels (plant survival rates, plant size, ratio of above-to-below biomass) are relevant for agricultural production. While the first may appear as trivial, the third is less straightforward. Stems and leaves are crucial for photosynthesis and the development of fruits, grains and other harvestable parts. With less biomass going to these structures, the overall growth and yield of the crop may decrease.

← 2. For instance, the survival probability of Salix trees (e.g. willows and osiers) exposed to a 30-day drought episode is half of those not exposed to it. In sharp contrast, tamarisc trees exposed to 30 days of a drought have almost 90% probability of survival, relative to non-exposed tamarisc trees.

← 3. The key numerical findings are as follows. Reducing soil moisture by 1.0 standard deviation reduces cropland vegetation during the same year by approximately 0.4 standard deviations, as most of the used models predict an effect between 0.30 and 0.55. For forests and wetlands, the central estimates are 0.25 and 0.20 respectively. Econometric model specifications and research details are provided in (Tikoudis, Gabriel and Oueslati (2025[4]).

← 4. The central numerical finding is that 400 millimetres less rain per year may have an effect that is equivalent to a 1.0 standard deviation reduction in soil moisture.

← 5. Here, the reported estimates originate from back-of-the envelope calculations using the summary statistics in Table 2 and the econometric estimates for dryland US counties that do not control for temperature or precipitation (Table 4, column 1) in Kuwayama (2019[25]). The average exposure was 8.5 weeks in mild droughts (D0), 5.7 weeks in moderate droughts (D1), 3.9 weeks in severe droughts (D2), 2.3 weeks in extreme droughts (D3), and 0.8 weeks in exceptional droughts (D4).

← 6. The authors report the standard deviation to be 2.7 units of the PDSI.

← 7. However, droughts and water stress are not the only reasons behind vulnerabilities in the supply of energy from thermoelectric power stations. A branch of literature stresses the general problem of water scarcity and water allocation across residential consumption, industrial use, energy and food systems. See for example, Zheng et al. (2016[93]) for China, Hejazi et al. (2023[94]) for the Middle East and North Africa region.

← 8. This number was originally reported by AccuWeather (2022[95]) and has been cited by World Economic Forum (2023[96]).

← 9. A related reference empirically examining the effect of rainfall on GDP is the paper by Zaveri, Damania and Engle (2023[51]).

← 10. Šedová, Čizmaziová and Cook (2021[97]) and Minx et al. (2017[98]).

← 11. Cipollina, De Benedictis and Scibè (2024[99]). For other bibliometric reviews of the field see Maretti, Tontodimamma and Biermann (2019[102]); Milán-García et al. (2021[100]); Priovashini and Mallick (2022[101]).

← 12. Therefore, each observation in the meta-analysis represents the (reported) effect of a 1.0 standard deviation change in a climate variable on migration, with that reported effect expressed also in standard deviations.

← 13. The contribution and limitations of qualitative approaches in the general environmental security literature are investigated by Bernauer, Böhmelt and Koubi (2012[103]).

← 14. Agder et al. (2014[85]) attributes this to Theisen, Gleditsch and Buhaug (2013[104]).

Featured topics

Agriculture and fisheries

Climate change

Development

Digital

Economy

Education and skills

Employment

Environment

Finance and investment

Governance

Health

Industry, business and entrepreneurship

Regional, rural and urban development

Science, technology and innovation

Society

Taxation

Trade

Energy

Nuclear energy

Transport

Featured topics

Agriculture and fisheries

Climate change

Development

Digital

Economy

Education and skills

Employment

Environment

Finance and investment

Governance

Health

Industry, business and entrepreneurship

Regional, rural and urban development

Science, technology and innovation

Society

Taxation

Trade

Energy

Nuclear energy

Transport

Countries A - C

Countries D - I

Countries J - M

Countries N - R

Countries S - T

Countries U - Z

Regional and global engagement

Countries

Countries A - C

Countries D - I

Countries J - M

Countries N - R

Countries S - T

Countries U - Z

Regional and global engagement

Publications

Publications

Featured publications

Data

Data

Featured data

News & events

News & events

Featured events

About OECD

About

Engage with us

Work with us

Featured topics

Agriculture and fisheries

Climate change

Development

Digital

Economy

Education and skills

Employment

Environment

Finance and investment