Policy Research Working Paper 10697 Does Global Warming Worsen Poverty and Inequality? An Updated Review Hai-Anh H. Dang Stephane Hallegatte Trong-Anh Trinh Development Data Group & Sustainable Development Practice Group February 2024 Policy Research Working Paper 10697 Abstract This paper offers an updated and comprehensive review of a recently constructed global database at the subnational recent studies on the impact of climate change, particularly unit level derived from official national household income global warming, on poverty and inequality, paying special and consumption surveys shows that temperature change attention to data sources as well as empirical methods. has larger impacts in the short term and more impacts While studies consistently find negative impacts of higher on chronic poverty than transient poverty. The results temperature on poverty across different geographical are robust to different model specifications and measures regions, with higher vulnerability especially in poorer of chronic poverty and are more pronounced for poorer Sub-Saharan Africa, there is inconclusive evidence on countries. The findings offer relevant inputs into current climate change impacts on inequality. Further analysis of efforts to fight climate change. This paper is a product of the Development Data Group, Development Economics and the Sustainable Development Practice Group. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http:// www.worldbank.org/prwp. The authors may be contacted at hdang@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Does Global Warming Worsen Poverty and Inequality? An Updated Review Hai-Anh H. Dang, Stephane Hallegatte, and Trong-Anh Trinh* JEL Classification: Q54; I32; O1 Key words: Climate change; temperature; poverty; inequality; subnational data * Dang (hdang@worldbank.org; corresponding author) is a senior economist with the Living Standards Measurement Unit, Development Data Group, World Bank and is also affiliated with GLO, IZA, Indiana University, and London School of Economics and Political Science; Hallegatte (shallegatte@worldbank.org) is a senior climate change advisor with the Sustainable Development Global Practice, World Bank; Trinh (trong- anh.trinh@monash.edu) is a consultant with the Living Standards Measurement Unit, Development Data Group, World Bank, and a research fellow at the Centre for Health Economics, Monash University, Australia. We would like to thank Edward Barbier, Somik Lall, Daniel Mahler, and Jevgenijs Steinbuks for comments on earlier versions of the paper. We would also like to thank the UK Foreign Commonwealth and Development Office (FCDO) for funding assistance through the Data and Evidence for Tackling Extreme Poverty (DEEP) Research Program. 1. Introduction Climate change is a topic of global concern, where a fast-growing literature exists in economics that inform policies and policy makers (Hsiang, 2010; Jones and Olken, 2010; Dell et al., 2012; Kotz et al., 2021). However, despite considerable research on how climate change affects economic factors like GDP or economic growth, there has been less exploration of its distributional impacts, particularly on poverty and inequality (IPCC, 2022). Indeed, concerns have been raised that poorer countries, or vulnerable population groups in the same countries, might be more negatively affected by climate change (Hallegatte et al., 2017; Barbier and Hochard, 2019; Hsiang et al., 2019; Chancel et al., 2023). We make several new contributions in this paper. First, we offer an updated and comprehensive review of the empirical assessments of (past or future) consequences of hotter temperature on poverty and inequality. The literature that we review focuses on the longer- term impacts of gradual changes in temperature, but also includes some of the recent papers estimating the impacts of other climate change phenomena such as changing precipitation (rainfall) patterns, sea level rises, and natural disasters on poverty or inequality. We pay special attention to studies’ data sources as well as empirical methods, which have received relatively less attention in the literature. While studies are difficult to compare and aggregate due to different poverty metrics, assumptions, and country contexts, they consistently indicate a negative impact of temperature changes (and most of the other climate change phenomena) on income, expenditure, and poverty levels across different geographical regions, with higher vulnerability in poorer countries and especially in Sub-Saharan Africa. In contrast with these consistent results for poverty, there appears to be no agreement on the impacts of climate change on inequality, with different studies reaching different conclusions on the directions of the effects. 2 The challenge with comparing results from various studies covering different countries or regions calls for the use of standardized datasets that are comparable across countries. Our second contribution is to offer new, global estimates for the impacts of hotter temperature on inequality and poverty dynamics (including chronic and transient poverty) at the subnational level. We analyze the Subnational Poverty and Inequality Database (SPID), a recently constructed global database derived from official national household income and consumption surveys, incorporating over 1,594 subnational units from 134 countries, and spanning the period from 2003 to 2019 (Nguyen et al., 2023). Using this database, we obtain new global headcount poverty rates at the subnational unit level, using different poverty lines (from the extreme poverty line at $2.15/day to the higher poverty line at $6.85/day). 2 We further analyze richer poverty outcomes by classifying each subnational unit as being in chronic poverty (i.e., poverty extending over multiple consecutive periods), or transient poverty (i.e., temporary poverty lasting over non-consecutive periods). To our knowledge, these estimates on chronic and transient poverty at a global scale were not available before. The policies to address transient poverty are distinctly different from those for chronic poverty. Specifically, strong social protection programs would most effectively address transient poverty (e.g., as they help prevent the non-poor but vulnerable households from falling into poverty), but longer-term investments in human capital and infrastructure would better tackle chronic poverty. 3 Our analysis suggests that change in temperature has a large impact on poverty at the subnational unit level, with larger impacts over the short-term than over the long-term, 2 Unless otherwise noted, all our estimates in this paper are at the subnational unit level. Our poverty measures are adjusted to real values using 2017 PPP dollars. 3 See, e.g., Barret (2005) and Ravallion (2016) for further discussion on different policy interventions regarding chronic poverty versus transitory poverty. Furthermore, distinguishing between chronic and transient poverty has much policy relevance, given the debate over whether and how much shocks that trigger transient poverty can generate chronic poverty, especially in places with low income and limited access to financial tools, through poverty traps (Barrett and Carter, 2013). 3 suggesting different impact channels at different timescales and a potential role of longer-term adaptation. We also find that climate change has a more pronounced impact on chronic (subnational unit) poverty compared to transient poverty. Specifically, a 1°C increase in temperature can result in around 6% increases for different measures of chronic poverty. This disparity suggests that climate change affects long-term trends and structural poverty, with harmful consequences extending beyond short-term losses that typically follow unexpected shocks. The paper has five sections. We present in the next section a simple analytical framework linking climate change and disasters to poverty and inequality (Section 2.1) before summarizing the methodologies and data commonly employed in the literature (Section 2.2). In Section 3, we subsequently review existing studies on the impacts of temperature change on poverty (Section 3.1), inequality (Section 3.2), and other related outcomes (Section 3.3). In Section 4, we discuss the data and our empirical framework (Section 4.1) before presenting our new estimates on general poverty (Section 4.2), poverty dynamics including chronic poverty and transient poverty (Section 4.3), as well as robustness checks (Section 4.4). We finally conclude in Section 5. 2. Analytical Framework 2.1. Theoretical Motivations In order to provide a comprehensive overview of the effects of temperature change on poverty and inequality, this section outlines a simple conceptual model of the pathways through which higher temperature might influence these outcomes. Recognizing the intricate nature of this relationship, Figure 1 unfolds several interconnected mechanisms including economic growth, agriculture, labor productivity and human capital, migration, and other channels (such as rising sea levels). While the channels could carry a mix of both positive and negative impacts from 4 temperature change, the impacts might vary from location to location and are highly context specific. Furthermore, these impacts can also differ in magnitude. As a result, Figure 1 represents our attempt to broadly sign these effects using evidence from either the majority of existing studies or most recent studies. Clearly, understanding the context-specific complexities is essential for developing effective mitigation and adaptation strategies. Economic growth stands as the key channel through which temperature change affects poverty and inequality. Rising temperatures can dampen economic growth rates, negatively affecting income levels and job opportunities, which can lead to increased poverty and inequality (Hsiang, 2010; Jones and Olken, 2010; Dell et al., 2012; Kotz et al., 2021).4 However, the impacts of these changes are unevenly distributed, with research showing differing effects between richer and poorer countries (Moore and Diaz, 2015; Newell et al., 2021). Furthermore, the relationship between temperature and economic growth can be non- linear, yielding both positive and negative impacts depending on the temperature ranges and geographical locations. For example, warmer temperatures can foster economic production in colder countries up to a certain threshold, beyond which they have negative impacts (Burke et al., 2015). Given the complex interplay between climate change and economic growth, the influence of temperature change on poverty and inequality can also vary dynamically. This underscores the necessity for nuanced and context-specific strategies in responding to these multifaceted challenges. Specifically, global warming can impact the agriculture sector, which plays a critical role in poverty and inequality in developing economies that heavily depend on this sector for livelihoods and GDP (Mendelsohn, 2009; Schlenker and Lobell, 2010; Aragón et al., 2021). For instance, higher temperatures and changing rainfall patterns can alter the conditions 4 There is not a straightforward relationship between economic growth and poverty and inequality. Economic growth is generally beneficial for poverty reduction, but this relationship can change depending on inequality levels (Cerra et al., 2022; Ferreira et al., 2022). 5 necessary for plant growth, resulting in decreased crop yields and increased crop failure. This is particularly devastating for farmers with small land holdings or those engaged in subsistence farming, often leading to a cycle of diminished income, and ultimately, poverty (Morton, 2007). Similar to its impacts on economic growth, temperature can also exhibit non-linear effects on crop yields. A certain amount of warming might improve crop yields initially, but beyond an optimal point, any additional temperature increase can drastically reduce productivity (Schlenker and Roberts, 2009; Lobell et al., 2011). Furthermore, the negative impacts on agriculture do not solely affect those directly involved in farming but can also reverberate through the broader economy. For example, reduced agricultural output can lead to increased food prices and food insecurity (Wheeler and Von Braun, 2013), which in turn, can have negative consequences on the whole economy and exacerbate poverty and inequality. Another key channel that global warming can affect poverty and inequality is through labor productivity and human capital. While outdoor sectors such as agriculture, construction, and tourism are most vulnerable under hotter temperatures, indoor work activities without adequate cooling can suffer. Previous studies show that heat stress can negatively affect physical labor capacity and cognitive function (Graff Zivin and Neidell, 2014; Zhang et al., 2018; Somanathan et al., 2021; LoPalo, 2023). As temperatures rise, workers may be unable to work effectively, leading to declined labor productivity. This would negatively impact earnings and employment, pushing more people into poverty and widening income gaps (Deryugina and Hsiang, 2014). The global warming-induced negative impacts on human capital can be further demonstrated with health and education outcomes. Indeed, negative health effects of temperature change can be observed even before birth. Previous studies have found that high temperatures are negatively associated with birth indicators, leading to premature birth and low birth weight (Deschênes et al., 2009; Deschênes and Greenstone, 2011; Molina and 6 Saldarriaga, 2017). These early health issues could result in harmful, lifelong consequences for health and other outcomes (Graff Zivin et al., 2018). Similarly, exposure to extreme temperatures over the life course can also contribute to worsening physical and mental health outcomes (Barreca, 2012; Barreca et al., 2016; Mullins and White, 2019), which limit an individual’s capacity to work and their long-term income-earning potential (Fishman et al., 2019). Regarding education, prolonged exposure to heat has been found to impact learning ability, with studies showing that high temperatures cause lower school attendance and worse student performance on standardized tests (Randell and Gray, 2019; Park et al., 2020; Park et al., 2021; Park, 2022). Additionally, children may be required to stay at home to assist their parents or supplement family income to help cope with weather-related income shocks (Mottaleb et al., 2015; Colmer, 2021), which can further reduce their learning opportunities. As a result, the negative effects of temperature increases on health and education erode human capital, leading to decreased future employment outcomes and subsequently more poverty and inequality. Migration is another crucial pathway through which temperature increases can contribute to poverty and inequality. Changing climate conditions, particularly rising temperatures, can make certain areas less habitable or economically viable. Even worse, changing climate conditions can worsen conditions in poverty "hotspots" that are already vulnerable to climate change such as agricultural lands constrained by difficult terrain, poor soil quality, limited rainfall, or with limited access to markets or rural low-elevation coastal zones (Barbier and Hochard, 2018a and 2018b). For example, extreme heat can reduce agricultural productivity or increase prevalence of heat-related illnesses, making it difficult for individuals to sustain their livelihoods. Consequently, households facing these adverse conditions might be prompted to migrate, often from rural to urban areas within the same country, or even across national borders in search of better living conditions and economic opportunities (Marchiori et al., 2012; 7 Kubik and Maurel, 2016; Thiede et al., 2016; Jessoe et al., 2018). While this could lead to spatial inequality as people move from areas of lower to higher economic productivity, the results on poverty could be mixed. On one hand, migration could potentially decrease poverty if migrants send home remittances. On the other hand, these same climate-induced economic hardships can also reduce individuals’ ability to migrate due to liquidity constraints (Cattaneo and Peri, 2016; Hirvonen, 2016; Letta et al., 2023). It is possible that the worse the economic conditions caused by the climate shock, the less likely individuals have the financial means necessary to migrate. In such cases, households are forced to remain in “poverty traps”, leading to increased poverty rates (Barrett and Carter, 2013; Kraay and McKenzie, 2014). The simple theoretical framework discussed above focuses on temperature change as the most important factor of climate change affecting poverty and inequality. However, in the broader framework of climate change impacts, there are multiple temperature-related manifestations, which can profoundly influence the channels mentioned. For example, sea level rise, a significant aspect of climate change, can force many coastal populations to relocate, putting added pressure on urban resources and livelihoods through increased migration (Chen and Mueller, 2018; Hauer et al., 2020). Additionally, sea level rise can also impact freshwater resources, leading to potential shortage, which can directly reduce agricultural productivity and consequently the earnings of those who rely on agricultural production (Dasgupta et al., 2009; Chen et al., 2012). Parallel to this, another dimension of climate change – unusual shifts in rainfall patterns – has its own implications. For many economies, especially those reliant on agriculture, variations in rainfall can negatively affect productivity levels, harvest yields, and consequently, overall GDP (Sarsons, 2015; Barrios et al., 2010; Damania et al., 2020; Kotz et al., 2022). Such reduced income caused by rainfall shocks can have adverse effects on human capital (Maccini and Yang, 2009). For example, families with limited resources may struggle to access proper 8 health care, increasing their vulnerability to diseases and other health issues (Rosales-Rueda, 2018). Furthermore, children’s education could be affected if parents find it challenging to afford school fees or other educational expenses. In some extreme cases, children might be pulled out of school to support the family financially (Colmer, 2021). In addition to these effects, studies have shown that rainfall variability induced by climate change is also associated with higher likelihoods of conflict and increased patterns of migration (Miguel et al., 2004; Kleemans and Magruder, 2018). Given these various climate-induced factors, addressing climate change effects on poverty and inequality requires a comprehensive understanding of various manifestations and their intertwined impacts. 5 Our proposed framework briefly describes household responses in the short term (i.e., coping) and in the longer term (i.e., adaptation). For example, households could adopt coping strategies to natural disasters by not sending children to school (Fuller et al., 2018), or respond to climatic shifts by switching to more drought-resistant crops or altering their planting seasons (Roesch-McNally et al., 2018; Ponce, 2020). While coping/ adaptation to climate change could lessen the impacts of global warming (say, through technological advancements (Hsiang et al., 2017)), the intertwined linkages between short-term coping and longer-term adaption are complicated and could either reduce or exacerbate poverty and inequality depending on the time period under consideration. For example, a household’s decision to withdraw their children from school might help cut expenses in the short term, but it could negatively affect their human capital achievement and could lead to worse poverty outcomes in the future. Or if the government subsidizes housing protection with more resources going to larger houses (i.e., richer households), this policy response could increase inequality. As another example, a recent study points to lower global impacts of temperature change on poverty in the longer term— 5 For more focused discussion, in Figure 1 we present one-directional arrows indicating the directions of impacts. But the impacts can be multi-directional. For example, poverty reduction is observed to bring beneficial impacts on economic growth in the longer term (Thorbecke and Ouyang, 2022). 9 suggesting household adaptation to gradual warmer temperature over time (Dang, Nguyen, and Trinh, 2023). But this study also finds some evidence for more impacts on inequality in the longer term. But more importantly, this framework helps highlight the fact that not all climate change impacts on poverty and inequality are expected to be mediated by economic growth (Hallegatte et al., 2017), or any single sector alone. It would be important to investigate the final impacts of climate change on poverty and inequality, which are generated as a sum total through all these different channels. Furthermore, the existing literature has mostly focused on the impacts of temperature change on short-term poverty. Indeed, existing studies do not distinguish between various forms of poverty dynamics, such as chronic and transient poverty. We attempt to fill these gaps with new estimates on the global impacts of temperature change on both chronic and transient poverty. 2.2. Data and Empirical Models We review recent studies and pay special attention to data and empirical models in this section. We focus on questions whether studies use household consumption (income) surveys as the main sources of data, whether they use panel data, whether recent studies use big data (e.g., satellite data) more often, and whether studies analyze data that are more disaggregated than the country level. Regarding empirical models, we review studies that employ cross sectional models, panel data models, as well as non-parametric and simulation techniques. 2.2.1. Data  Are household consumption (income) surveys the main sources of data? While household surveys appear to be the standard data sources for climate change studies, it is useful to review this question and limitations with household surveys before we discuss the new data sources employed in recent studies. Indeed, existing studies examining the effects of 10 climate change and natural disasters on poverty often rely on household consumption surveys as a dominant source of data. This is true both for single-country (e.g., Akter and Mallick, 2013; Arouri et al., 2015; Bangalore et al., 2018) and cross-country analyses (Azzarri and Signorelli, 2020; Carter et al., 2007). A notable advantage of most living standards measurement study (LSMS)-type household surveys is that they provide comprehensive data on household consumption levels, enabling researchers to construct poverty indicators that directly align with the study’s objectives (Wong and Brown, 2011; Jakobsen, 2012). In single- country analyses, the scope of the survey can be more closely tailored to specific regional or national contexts, providing nuanced understanding of the poverty situation. These studies often employ the headcount poverty rate and the Gini coefficient to measure inequality, which can be estimated using household surveys. 6 While single-country studies often focus on specific events or circumstances within a country, their findings are typically limited in generalizability. In contrast, cross-country studies attempt to overcome this limitation by comparing and contrasting experiences across a range of countries. This broader perspective can shed light on patterns and relationships that might not be apparent within a single-country context, thereby offering findings with greater generalizability. However, assembling comparable data across multiple countries presents a significant challenge due to variations in data collection methods, definitions of key variables, and data quality among different nations. As a result, many studies often rely on standardized databases on poverty and inequality such as the World Bank’s PovcalNet database, later transitioned into the Poverty and Inequality Platform (PIP), which offer estimates for inequality and poverty under various poverty lines or the United Nations University’s World Income 6 While most studies primarily rely on the Gini coefficient (Cappelli et al., 2011; Tol, 2021; Paglialunga et al., 2022; Pleninger, 2022), other inequality indexes are also estimated based on household surveys such as income ratios, Theil index, or Palma ratio to analyze inequality in more depth (Keerthiratne and Tol, 2018; Diffenbaugh and Burke, 2019). Beyond income poverty and inequality, recent studies also follow a lack of basic needs approach and examine other welfare aspects including food, education, health care, and social protection (e.g., Paavola, 2017; González et al., 2022; Rana et al., 2023). 11 Inequality Database (WIID) (Azzarri and Signorelli, 2020; Cappelli et al., 2021; Paglialunga et al., 2022). 7 Still, a major limitation with household surveys is that such surveys are often not frequently collected, particularly in poorer countries due to resources and logistical constraints. A survey by Beegle et al. (2016) indicates that just slightly more than half (i.e., 27) of the 48 countries in Sub-Saharan Africa had two or more comparable household surveys for the period between 1990 and 2012. Dang et al. (2019) find a 10 percent increase in a country’s household consumption is associated with almost one-third (i.e., 0.3) more surveys, indicating the paradoxical fact that poorer countries with a stronger need for poverty reduction also face a more demanding challenge of poverty measurement given their smaller numbers of surveys. Furthermore, many studies exploring the impacts of natural disasters on poverty and inequality traditionally rely on household surveys that collected data on extreme events (Datt and Hoogeveen, 2003; Khandker, 2007; Bui et al., 2014; Arouri et al., 2015; Karim, 2018). While these surveys, usually based on respondents’ recollections, provide direct insights into how households are affected by and respond to disasters, the accuracy of such data can be subject to recall bias and other inherent errors associated with self-reported information. In addition, many (older) surveys do not provide GIS information that accurately geo-locates households. Consequently, recent studies have increasingly turned to other sources of data that offer the geographical and temporal granularity needed to provide better and timely analysis of climate change and natural disasters that we discuss below.  Are panel data often used? The analysis of climate change and natural disasters’ effects on poverty and inequality has evolved in tandem with advancements in data collection and processing. Early studies often 7 For more details of the PIP database, see: https://pip.worldbank.org/home. 12 rely on cross-sectional survey data due to the constraints of data availability (Adger, 1999; Datt and Hoogeveen, 2003; Carter et al., 2007; Lal et al., 2009). Cross-sectional studies observe a set of units (e.g., households or individuals) at a single point in time, allowing researchers to make comparisons and draw conclusions about prevalence of poverty and inequality. However, cross-sectional data comes with inherent limitations. It only provides a snapshot view of poverty across units at one point in time but does not offer insights into changes over time or the dynamics that drive these changes. This lack of temporal perspective can lead to omitted variable bias and severely limit causal inferences. Recognizing these limitations, many studies have turned to panel data, which observes the same units across multiple time points (Giesbert and Schindler, 2012; Rodriguez-Oreggia et al., 2012; Yamamura, 2015; Keerthiratne and Tol, 2018; Baez et al., 2020; Sedova et al., 2020; Sohnesen, 2020; Cappelli et al., 2021). Panel data allows researchers to track changes over time and explore the temporal dynamics of poverty and inequality; consequently, it also offers a more in-depth understanding of the effects of climate change and natural disasters, which often have a longitudinal nature themselves. However, the use of panel data is subject to several barriers. In particular, tracking the same households (individuals) over a long period is expensive and poses logistical challenges, especially where households migrate or live in remote locations. In particular, attrition—where households drop out of the panel for various reasons—is a well-known issue affecting the quality of panel data and would require considerable resources investment and advanced survey strategies to address (Schoeni et al., 2013). As an example, the attrition rate in the panel Russia Longitudinal Monitoring Survey (RLMS) is already 40 percent in the first 10 years after it was fielded, but sharply increases to more than 70 percent after another 10 years (Kozyreva et al., 2016). These challenges can be amplified in poorer countries (or conflict situations), where both resources and data collection capacity are limited. Given these challenges, alternative data imputation methods have been proposed to construct synthetic panels from repeated cross 13 sections that allow richer analysis of poverty dynamics (Dang et al., 2019; Dang and Lanjouw, 2023). But since there are few applications of these imputation methods for studying the impacts of climate change (see, e.g., Nakamura et al., 2023), more studies in different contexts would be useful. Recent studies have utilized administrative data, which typically offers better census-type coverage and longer-run panel data, especially in wealthier countries where such data is readily available. For instance, Duque et al. (2018) use administrative data from Colombia to examine how early-life exposure to adverse weather conditions and subsequent human capital investments via conditional cash transfers (CCTs) affect children’s long-term outcomes. Focusing on school performance, Park (2022) exploits student-level administrative data from a public school district in the U.S. to examine how hot temperatures impact high-stakes exam performance and subsequent educational attainment. Similarly, Heyes and Saberian (2019) analyze U.S. immigration judges’ decisions, demonstrating a link between outdoor temperature and immigration adjudications. Yet, while administrative data offer multiple advantages over household survey data such as accuracy, representativeness, and longer panel data, it is still rarely used in poorer countries due to weaker statistical capacity (Dang et al., 2023b), which could range from fewer resources to less established data infrastructure. This explains the limited analysis of administrative data for poorer countries, particularly in the context of poverty and inequality.  Do recent studies use big data (e.g., satellite data) more often than previous studies? The rise of big data, together with progress in geographic information system (GIS) technology, has greatly expanded the range of data sources available for studies on climate change, particularly natural disasters. An alternative to household surveys is to use global disaster databases such as the Emergency Events Database (EM-DAT), which provides comprehensive 14 data on the occurrence and impacts of over 22,000 mass disasters worldwide since 1900 (Yamamura, 2015; Cappelli et al., 2021). Another source of disaster data is DesInventar, which provides records of disaster at the country level from 1980 (Rodriguez-Oreggia et al., 2012; Keerthiratne and Tol, 2018). The advantage of such databases is their wide coverage and the standardization of disaster data, enabling cross-country comparisons. However, they often lack detailed local-level data, which can be crucial for understanding variations in disaster impacts within countries. Moreover, these databases predominantly focus on large-scale disasters, potentially excluding smaller, localized disasters that can have a substantial cumulative impact on poverty and inequality. 8 The reliance on official reports or media coverage for disaster validation may also underrepresent disasters in areas with less robust reporting systems. These limitations highlight the importance of incorporating additional data sources to provide a more comprehensive understanding of disaster impacts. Another nuanced approach in the study of disaster impacts combines traditional survey techniques with geospatial and localized risk information (i.e., footprint data). Geo-referencing is employed to further validate self-reported exposure data, thus combining ground truth with computational models. Such methodologies aim to generate a more intricate understanding of localized risks and household experiences, and therefore bring a multidimensional perspective that blends traditional and high-tech methods. Specifically, recent studies use community flood maps and satellite imagery to identify flood risks or use self-reported survey data that were validated with a newly developed flood map, offering a more complex understanding of households’ experiences during flood events (Erman et al., 2019; Erman et al., 2020). A particularly significant development is the use of reanalysis data, which combines satellite observations, ground measurements, and statistical models to offer near-global coverage of 8 To be included in EM-DAT, a disaster must meet at least one of the following criteria: 10 or more people reported killed, 100 or more people reported affected, declaration of a state of emergency, or a call for international assistance. 15 climate variables (Giesbert and Schindler, 2012; Diffenbaugh and Burke, 2019; Azzarri and Signorelli, 2020). In particular, Rentchler et al. (2022) combine different data sources on flood, population, and poverty to provide more granular assessment of exposure to flood risks and poverty. Combining various data sources is especially valuable for providing critical weather information in remote and inaccessible areas where ground-based networks are limited. 9 While satellites can capture a broad spectrum of data simultaneously, satellite data comes with its own set of limitations. For instance, these records typically cover a shorter time span than station- based data, owing to the relatively recent deployment of satellite technologies. Furthermore, the data require complex processing to account for non-climatic factors, such as variations in satellite orbits and the need for instrument calibration.  Do studies analyze data that are more disaggregated than the country level? While most existing studies analyze country-level data, using such data has certain drawbacks. Country-level data can mask significant disparities within countries, such as regional variations in weather conditions, exposure to natural disasters, and levels of poverty and inequality. This issue can be particularly pronounced in developing countries, which are often located in tropical areas with diverse weather conditions and disaster risks. Consequently, important intra- country variations can be obscured when data is aggregated at the national level. In response to this limitation, a growing number of studies have begun to exploit subnational data. These studies observe more pronounced effects of climate change on economic growth and poverty at this more granular level (Damania et al., 2020; Kalkuhl and Wenz, 2020; Dang et al., 2023b). This shift towards more localized analyses underscores the need for more disaggregated, 9 Temperature and rainfall data are primarily sourced from ground-based weather stations, which are valued for their long-term records and high precision (Jones and Olken, 2010; Dell et al., 2012). However, station-based data often has uneven geographic distribution, resulting in coverage gaps particularly in poorer countries or regions where there is limited infrastructure to support these stations (Dell et al., 2014). 16 region-specific data and further highlights the potential of new data sources in shedding light on the intricate relationships between climate change, poverty, and inequality. 2.2.2. Empirical Models Previous studies often rely on three primary types of econometric models to examine the economic effects of climate change and natural disasters: cross-sectional models, panel models, and long differences models. The first two models have been used to evaluate the effects on poverty and inequality (Arouri et al., 2015; Azzarri and Signorelli, 2020; Paglialunga et al., 2022), while the last model has been used recently on other outcomes such as agriculture, health, and labor reallocation (Burke and Emerick, 2016; Obradovich et al., 2018; Liu et al., 2023). Cross-sectional models offer the starting point where the relationship between climate change and poverty (inequality) is commonly formulated as = + + + (1) In such models, the dependent variable ( ) typically represents a socio-economic indicator, such as poverty (inequality), in a specific location (i) during a given period. These models factor in variables like annual temperature ( ) and precipitation ( ) (Azzarri and Signorelli, 2020). 10 A potential omitted variable bias is an inherent challenge in Equation (1), which refers to a situation where unobserved correlations between temperature and other factors like technological changes or labor productivity, could influence the outcome variable. To 10 While rainfall is not the central focus of our review, it is worth noting that a substantial body of literature has applied similar models to explore the impacts of (various metrics of) rainfall on different outcomes. For example, previous studies have used total annual precipitation (Damania et al., 2020), rainfall deviations from the historical means (Corno and Voena, 2023), and extreme rainfall events exceeding specific thresholds (Shah and Steinberg, 2017). Each of these measures provides a distinct lens through which the multifaceted impacts of rainfall can be understood. 17 overcome this limitation, researchers use panel models. Panel models often take the following form , = + , + , + + + , (2) These models employ fixed effects for both location ( ) and year ( ), providing control for unobserved time-invariant factors and temporal changes that may affect the dependent variable (, ) (Paglialunga et al., 2022). Notably, Equation (2) assumes that climatic variables have linear effects. To test the robustness of this assumption, some studies have proposed variations improvements, such as introducing quadratic temperature terms, controlling for temperature change, or further examining interactive effects between temperature levels and changes (Kalkuhl and Wenz, 2020). Yet, to better address potential non-linear effects of climate change, other studies propose a temperature-bin approach that allows for a more flexible function of temperature by dividing it into designated bins (Chen and Gong, 2021; Mullins and White, 2020) , = + ∑ =1 , + , + + + , (3) A criticism of panel models, however, is that while such models can identify short-term effects of climatic changes, they do not capture long-run effects, particularly if these effects are mediated through household adaptation or intensify over time. To address this, long differences models aiming at analyzing gradual changes in the outcome variable and climatic variables between time periods are often employed (Burke and Emerick, 2016; Chen and Gong, 2021; Wing et al., 2021) ∆ = + ∆ + ∆ + , (4) 18 Robustness checks for the long-differences model include alternative specifications such as different period-average definitions, further controlling for geography or resource endowment covariates, and employing additional model specifications similar to those in panel models. 11 Beyond empirical estimation, a simulation-based approach is also popular for forecasting climate change impacts (Hsiang et al., 2013; Burke et al., 2015; Moore and Diaz, 2015). The method typically involves constructing a theoretical model, which is subsequently calibrated with actual data to simulate potential future outcomes under various climate scenarios. The most common practice in the economics literature involves using estimates from historical data and empirical models, as detailed in Equations (1)-(4), to project future climate change impacts (Burke and Emerick, 2016; Kalkuhl and Wenz, 2020; Newell et al., 2021). The estimating model can be simplified as = ( , ) (5) where X include climatic variables such as temperature, rainfall, and other relevant socio- economic indicators that are used to predict the future outcomes like poverty rates, inequality, and economic growth. The simulation approach is forward-looking and allows for predictions into the future under various scenarios. It also offers flexibility in incorporating diverse factors and complex interactions. However, two key limitations of most simulation studies are the assumptions that the impacts of warming due to one degree increase in one place (i.e., what they measure in the past when using mostly weather variations) would be the same everywhere else (i.e., what they try to predict in the future, which is a global increase in temperature), and that future impacts will be similar in magnitude to impacts in the past. When these assumptions do not hold, estimation results are likely biased and miss impacts that are of a systemic nature (e.g., a 11 Variants of the long-differences model include stacking long-term data into intervals (e.g., decades) to examine the long-term effects of disasters on economic outcomes (see, for example, Boustan et al., 2020). Another example is triple-differences models that make use of variation in multiple weather shocks across different channels such as space, time, and cropping cycles (Baez et al., 2020). 19 collapse in major ecosystems or physical systems) or linked to absolute thresholds (e.g., physiological limits). It is useful to note that future impacts of climate change on income, inequality, or poverty does not depend only on the nature and magnitude of climate change impacts, but socioeconomic contexts also matter (Kelly and Adger, 2000; Hallegatte and Rozenberg, 2017). Future socioeconomic contexts (and household vulnerability) are extremely uncertain and driven by socioeconomic and technological trends, as well as policies and collective action. For example, the share of farmers engaged in small-scale agriculture, or the number of households lacking improved water and sanitation, can significantly influence population vulnerability to poverty. Most empirical studies do not adequately capture these socioeconomic factors, and assume unchanged vulnerability over time. This limit underscores the need for future research to integrate these elements for a more comprehensive understanding of the complex relationship between climate change, disasters, and poverty. 3. Impacts on Poverty and Inequality 3.1. Poverty A large body of literature has examined the impacts of climate change—but mostly through disasters—on poverty, both from individual empirical studies and literature reviews. 12 However, due to data limitations, not all studies have been able to quantify these effects on poverty (Hallegatte and Rozenberg, 2017). This section summarizes the key findings from several sample empirical studies that have successfully quantified these impacts. In Viet Nam, Arouri et al. (2015) explore the effects of storms, floods, and droughts on household welfare, finding that all three types of disasters reduce per capita expenditure by around 1.5%, 4.4%, and 3.5%, respectively and increase the probability of poverty for 12 For a recent review of the literature, see Hallegatte et al. (2020). 20 households in flood-affected communes by 1.8 percentage points. Another study by Bui et al. (2014) for the same country estimate that poverty would have decreased by 2.7 percentage points if households had not been exposed to natural disasters. Analyzing data from Mexico during 2000-2005, Rodriguez-Oreggia et al. (2013) highlight significant adverse effects of natural disasters on both human development and poverty, with an increase in poverty of between 1.5% and 3.7%. Notably, floods and droughts are the most devastating. Similar results are found in Argentina by González et al. (2022) where they find that a variety of natural disasters from 1970 to 2010 contributed to more multi-dimensional poverty. On a global scale, combining household surveys from 24 Sub-Saharan African countries, Azzarri and Signorelli (2020) find that flood shocks result in a 35% decrease in total and food per-capita consumption and a 17-percentage point increase in extreme poverty. Azzarri and Signorelli (2020) also predict that increases in rainfall and temperature, and consequently the incidence of disasters across Western and Central Africa, could result in dramatic increases in extreme poverty rates by up to 30 percentage points. Analyzing a new global, subnational database on poverty across 134 countries, Dang et al. (2023a) find strong and statistically significant global effects of both higher and lower temperature on poverty, with a one-degree Celsius (i.e., 1◦C) annual increase in temperature causing poverty increases of 0.9, 1.8, and 2.3 percentage points, respectively, using the daily poverty lines of $1.90, $3.20, and $5.50 (which correspond to 9.1 percent, 9.0 percent, and 6.8 percent increases). But this study also finds smaller impacts in the longer term, suggesting household adaptation to gradual warmer temperature over time. Other studies simulate the future effects of climate change and disasters on poverty. For example, using a CGE model to explore how future climate scenarios would affect economic growth and poverty in Zambia, Thurlow et al. (2012) find that even small changes in climate expected by 2025 can increase the absolute number of poor people by 32,000 over 10 years, 21 while also yielding variable impacts on GDP growth depending on the climate scenario. Meanwhile, Hallegatte and Rozenberg (2017) forecast that the number of people in extreme poverty due to climate change could reach between 35 million and 122 million by 2030, depending on the severity of climate impacts and the socioeconomic context and trends (e.g., pace of demographic transition, changes in economic structure, evolution of the skill premium, or the progress in providing universal access to basic infrastructure services). Overall, previous studies consistently indicate a negative impact of various types of natural disasters on household income, expenditure, and poverty levels across different geographical regions. Furthermore, simulation studies all point to potentially severe impacts of climate change on global poverty, particularly in Africa, further emphasizing the urgency of addressing this pressing issue. However, directly comparing these estimates proves difficult, given the diverse measures of poverty used across the studies. This highlights the necessity for a harmonized database to facilitate more meaningful comparisons and better understanding of climate change impacts. 3.2. Inequality A number of empirical studies have evaluated the global impacts of climate change on inequality, with more recent studies finding negative impacts. In particular, Yamamura (2015) finds that while natural disasters have increased income inequality in the short-term (5 years), this effect disappears in the long term (10 years). A recent study by Song et al. (2023), on the other hand, find catastrophic natural disasters to have negative relationships with inequality, as measured by the Gini index, in both the short and long run. Diffenbaugh and Burke (2019) similarly argue that global warming has significantly increased economic inequality between countries. Their findings reveal that global warming has reduced per capita GDP by 17%-31% at the poorest four deciles, contributing to a wealth gap that is 25% larger than in a world without global warming. Similarly, Paglialunga et al. (2022) find a 1% increase in temperature 22 to correspond to a 0.5-point increase in the Gini index. Cevik and Jalles (2023) focus more on the difference in the climate change effects across countries and find the impacts to be seven times greater in developing countries. Gilli et al. (2023) analyze country-level income deciles data and find rising temperature to have most negative impact on the poorest deciles within countries. Analyzing more disaggregated subnational inequality data across 134 countries, Dang et al. (2023a) similarly find that a 1◦C rise in temperature leads to 0.8 and 1.4 percent increases in the Gini and Theil indices, respectively. Yet, single-country studies often find mixed evidence of climate change impacts on income inequality. For instance, Bui et al. (2014) demonstrate that natural disasters can have a significant impact on income distribution in Viet Nam, particularly in regions with larger rural populations and agricultural workforces. Other within-country studies, however, present a more complex picture, highlighting uneven climate change impacts across different income levels within the same country. Pleninger (2022) finds that in the United States, the short-term effects of natural disasters primarily affect middle income population groups, leaving overall income inequality unchanged. This finding aligns with Otrachshenko and Popova’s (2021) study for the Russian Federation showing that extreme hot temperatures negatively impact regional GDP per capita but do not significantly affect income inequality. However, both studies confirm the variation in vulnerability among income levels and regions, with the poor regions being more susceptible to disasters. Some studies even highlight potential reductions in inequality following climate-related events. For example, Keerthiratne and Tol (2018) document a decrease in the Gini coefficient by 0.01 points in the aftermath of natural disasters in Sri Lanka. This reduction appears to be driven not by an increase in the income of poorer households, but rather by a decrease in the income of richer households. This finding adds another layer of complexity to our 23 understanding, highlighting that climate change impacts on inequality can vary greatly depending on country-specific socioeconomic and environmental contexts. 3.3. Other Outcomes/ Channels Studies that explicitly examine the mechanisms linking climate change and disasters with poverty and inequality are still relatively few, perhaps due to data limitations. The handful of studies that explore these mechanisms often converges on a few key themes, with agriculture standing out as a primary driver (Azzarri and Signorelli, 2020; Paglialunga et al., 2022). Climate change impacts tend to be amplified in countries with large rural populations or a high proportion of the workforce in agriculture. Such demographic characteristics not only exacerbate the vulnerability of these populations to climate change but also enhance the potential for climate change to aggravate inequality. Extreme climatic events may have differing impacts depending on the rural-urban context. While rising temperatures seem to exacerbate inequality more in largely rural countries (or regions), extreme precipitation events are associated with increased disparities in predominantly urban areas. Paglialunga et al. (2022) show that precipitation deviation can trigger flood risks in urban areas, damaging infrastructure and overloading drainage systems. Such events can disrupt food availability and inflate prices, disproportionately affecting the urban poor who rely more on markets for food. Combining household surveys with climatic datasets, our analyses of Chile, Colombia, and Indonesia, Nakamura et al. (2023) find that while households in large metropolitan areas are more likely to escape from poverty, climate shocks such as extreme rainfalls and high flood risks significantly reduce upward mobility, thus offsetting such benefits of urban agglomerations. In terms of adaptation mechanisms, social safety nets such as unemployment insurance can mitigate the negative impacts of natural disasters on income (Pleninger, 2022). However, these 24 measures may fall short in the face of multiple or severe disasters occurring within short timeframes. This indicates that the capacity to adapt and mitigate climate change effects is a critical factor in determining how these events influence poverty and inequality. In summary, understanding the mechanisms that link climate change and disasters with poverty and inequality is a multifaceted endeavor. While we have gained substantial insights from the focus on agriculture, numerous mechanisms remain relatively unexplored. These might include conflict, migration, and changes in labor productivity, which are promising areas for future research. 13 4. New Estimates for Chronic and Transient Poverty 4.1. Data and Analytical Framework Our empirical analysis relies on data from the Subnational Poverty and Inequality Database (SPID), a recently introduced global database mapping poverty and inequality on a subnational scale (Dang et al., 2023a; Nguyen et al., 2023). SPID is compiled from the Global Monitoring Database (GMD), an extensive repository of household survey data managed by the World Bank. The database derives from official national household income and consumption surveys, incorporating over 1,594 subnational units from 134 countries and spanning the period from 2003 to 2019 with most observations starting from 2010. Typically, a subnational unit aligns with a state or province (that is, first-level administrative boundaries - ADM1), but can also represent a cluster of regions, as determined by the specific household surveys’ sampling approach. 13 For example, climate change and related disasters could lead to an increase in conflict over increasingly scarce resources (Burke et al., 2009; Koubi, 2019), which in turn could exacerbate poverty and inequality. Migration, both within and across borders, is another possible response to climate change and disasters (Gröger and Zylberberg, 2016; Missirian and Schlenker, 2017), which could either mitigate or exacerbate inequality depending on the socio-economic characteristics of the migrants and the receptiveness of the destination regions. Climate change impacts on labor productivity, especially in sectors that are highly dependent on physical labor and weather conditions could have far-reaching effects on income distribution and poverty (Somanathan et al., 2021). 25 The key measure of SPID includes the (headcount) poverty rates estimated for each subnational unit using the daily poverty lines of $2.15, $3.65, and $6.85 and the relevant household income (consumption) surveys based on the GMD. That is, these poverty rates are calculated at the subnational level using household-level data and the specified poverty lines. The database also provides a multidimensional poverty index for a comprehensive perspective on poverty assessment. For assessing inequality, the SPID offers the Gini and Theil indexes (at the subnational unit level), which are the most common metrics for evaluating income inequality. All measurements are adjusted to real values using 2017 PPP dollars. To examine the effects of warmer temperature on chronic poverty, we integrate the SPID data with the ERA5 satellite reanalysis data from the European Centre for Medium-Range Weather Forecasts Reanalysis 5 (ECMWF). The ERA5 offers hourly assessments of multiple climate-related variables across an approximate 0.25 longitude by 0.25 latitude degree resolution grid, with data accessible from 1979 onwards. In line with previous studies, we compile the gridded data at the regional level by determining area-weighted averages (i.e., averaging all grid cells within a given region) (Kalkuhl and Wenz, 2020). 4.2. Temperature Effects on Poverty Our review so far has considered the effects of temperature on both poverty and inequality due to their inherent connection. In the following empirical analysis, we narrow our focus to explore the link between temperature and poverty. This choice is driven by our interest in examining poverty dynamics, an area that has not been extensively covered in current studies. For a deeper look into the effects of temperature on inequality, readers can refer to the comprehensive study presented in Dang et al. (2023). We start with the analysis of temeprature effects on contemporary poverty in this section and present new estimates on poverty dynamics in the subsequent section. 26 We present the results of the effects of warmer temperature on contemporary poverty in Table 1. Our main focus is the headcount poverty rate at the subnational unit level, using the $3.65/day or $6.85/day poverty thresholds. 14 To further highlight the differences between estimates using country-level data versus using subnational unit level data, we aggregate the subnational unit level data to the country level. 15 We begin with an examination at the country- level (Panel A), followed by the localized analysis at the subnational level (Panel B). We employ two models: the fixed-effects model and the long differences model, as shown in Equations (2) and (4). In each panel, the results of the fixed-effects panel model are presented in Columns (1) and (3), with the results of the long differences model in Columns (2) and (4). 16 Across both panels, the findings reveal strong evidence of the negative effects of rising temperature on poverty, regardless of the poverty measures used. Furthermore, the estimates derived from the subnational-level analysis (Panel B) exhibit stronger magnitudes compared to the country-level analysis (Panel A). Results of the equality tests confirm that these variations between the two levels of analysis are significant, indicating that an analysis based on country-level data may underestimate the effects of increased temperatures. This finding is consistent with that in another study using different poverty data (Dang et al., 2023a) and aligns with earlier research that has identified stronger effects of temperature on economic growth at the subnational level (e.g., Damania et al., 2020; Kalkuhl and Wenz, 2020). We then turn our focus to the subnational analysis to interpret the estimation results. Specifically, Column (1) of Panel B illustrates that a 1◦C increase in temperature corresponds to a 7.2 percentage point increase in poverty, translating to a 61.5% increase in poverty given 14 We also check the robustness of our findings using alternative measures of poverty in the Appendix. 15 Alternatively, we also use country-level data from the World Development Indicators (WDI) database. Our findings remain consistent (see Table A4 in the Appendix). 16 For the long differences model, we measure poverty and weather variables as the differences between 2010 and 2019. 27 an average poverty rate of 11.7%. The results of the long differences model are also significant and consistent with previous findings. However, the coefficient estimates from the long differences model are relatively less pronounced, with a 1◦C increase in temperature being associated with 1.6 percentage points (or 13.7%) increase in poverty. The t-tests at the bottom of Panel B indicate that the differences between the two approaches are statistically significant. The difference between the two specifications may arise from the difference between (i) short-term vulnerability, which links short-term weather fluctuations to poverty, including transient poverty in response to extreme weather, and (ii) a longer-term and more systemic vulnerability, which links changes in climate conditions to poverty, and especially chronic poverty. Part of this difference originates from the mechanisms through which weather and climate affect people and activities, but possibly also from households’ long-term adaptation in response to longer-term temperature change. The role of household adaptation in reducing the negative effects of temperature on various outcomes, such as economic production, agriculture, and human capital, has already been highlighted in empirical studies (e.g., Chen and Gong, 2021; Graff Zivin et al., 2018; Kalkuhl and Wenz, 2020). While our main interest lies on the effects of temperature on poverty, we also find mixed evidence regarding the impacts of rainfall on poverty, which align with findings from previous studies on economic growth (Dell et al., 2012; Damania et al., 2020; Kotz et al., 2022). The varying effects of rainfall might be attributed to its different outcomes on the economy. Adequate rainfall can benefit agricultural regions by enhancing crop yields, potentially alleviating poverty. Conversely, excessive rainfall can result in flooding, damaging infrastructure and agriculture, which could exacerbate poverty conditions. We further examine different functional forms of the rainfall variable in Table A5 (Appendix A). These include the squared term of rainfall (Columns (1) and (2)), deviations from long-term averages of rainfall (Columns (3) and (4)), and particularly high rainfall amounts, defined as being above certain 28 percentiles (Columns (5) to (8)). Again, our results on the effects of temperature remain consistent while the effects of rainfall appear inconclusive. 4.3. Temperature Effects on Poverty Dynamics Chronic Poverty We turn next to investigate climate change impacts on chronic poverty. Chronic poverty, characterized by prolonged and persistent poverty conditions, can have severe and lasting consequences on individuals and communities, affecting multiple generations and hindering overall socio-economic development. Addressing chronic poverty is therefore essential for sustainable poverty reduction, as transient poverty alleviation measures might not effectively target the root causes of long-term poverty. We present our analysis of the effects of warmer temperatures on chronic poverty using three measures: (i) the number of times each subnational unit is in poverty (i.e., the subnational unit’s mean consumption per individual in 2017 falls beneath the $3.65/day or $6.85/day poverty thresholds) from when it was first observed until it was last observed in the SPID database; (ii) the share of time that a subnational unit is in poverty during the same period. The measures are then regressed on temperature and rainfall variables using the same period of measurement. 17 Alternatively, we also employ a third measure, which is the average poverty rate using a window of 5-year period, conditional on each subnational unit having at least two data points during a particular period. 18 We then apply a fixed-effects model for the last measure. These three measures offer different but complementary aspects (i.e., absolute and 17 We also include the number of data points we have for each subnational unit in the database to control for potential differences due to data availability for each subnational unit. 18 For example, a subnational unit with data available from 2010 to 2019 was divided into two 5-year periods, with poverty and weather variables being averaged during each period. 29 relative measures) of the intensity of poverty depth that a subnational unit can fall into during the specified period. We discuss another measure of chronic poverty below. 19 Our analysis, presented in Table 2, consistently reveals negative effects of hotter temperatures on the three measures of chronic poverty, irrespective of the poverty threshold or the measure of chronic poverty employed. For example, using the first measure, a 1°C increase in temperature corresponds to a 2.3 percentage point (or 6.4%) increase in the number of poverty incidences at the $3.65/day threshold. Similarly, the second measure reveals an increase of 1.1 percentage points (or 6.3%) in the proportion of times a subnational unit is in poverty for the same temperature increase. The third measure, which looks at the average poverty rate over 5-year periods, shows a robust effect of warmer temperatures on chronic poverty. Notably, the effects of rainfall on chronic poverty are mixed. 20 These results underline the severity of the adverse effects of rising temperatures on chronic poverty and underscore the need for targeted interventions to mitigate these effects. Finally, we delve into a more detailed examination of the effects of hotter temperature across regions. This part of our analysis leverages the temperature bin approach (Equation 3), offering a more flexible functional form of temperature, as shown in Figure 2. We specify temperature as a series of indicator variables corresponding to 3◦C bins, where coefficients can be interpreted as the relative effects of falling within a particular bin compared to a reference “comfortable” bin (that is, 15-18◦C). We classify extreme hot weather as temperature being above 27◦C, and extremely cold weather as temperature being below 6◦C. Figure 2 highlights that rising temperature tends to exacerbate chronic poverty, particularly in poorer regions such as Sub-Saharan Africa. (For this exercise, we measure chronic poverty as the number of times each subnational unit is in poverty). We also observe adverse effects of colder temperature, 19 See Fusco and Van Kerm (2023) for a recent review on chronic poverty measures. 20 It could be the case that non-linear effects of rainfall, or variations in rainfall over time, explain the mixed effects on the outcomes. However, other studies that analyze disaggregated data find negative impacts of rainfall on economic growth (Damania et al., 2020; Kotz et al., 2022). 30 predominantly in Europe and Central Asia. 21 These insights correspond with several studies that find negative effects of cold weather on aspects such as productivity, health, and economic growth (Cook and Heyes, 2020; Dell et al., 2012; Deschênes and Moretti, 2009). Transient Poverty Unlike chronic poverty, transient poverty is characterized by short-term fluctuations in poverty status, often due to temporary shocks or income volatility. Understanding transient poverty is therefore crucial for assessing how individuals and communities cope with short-term disruptions caused by climate change and weather shocks, which could have long-term implications if not effectively managed. For better comparison with earlier results, we employ three related measures of poverty: (1) chronically poor, (2) transiently poor, and (3) generally poor. Specifically, we classify a subnational unit as “chronically poor” if it falls below the poverty threshold in both periods within a 5-year window, conditional on having at least two data points during that period. Conversely, a subnational unit is considered “transiently poor” if it falls below the poverty threshold in either (but not both) of the two periods within the same 5-year window. Finally, a subnational unit is classified as “generally poor” if it is either chronically poor or transiently poor—that is, it falls below the poverty threshold in either of the two periods. While this new definition of chronic poverty is somewhat simpler than those employed for Table 2 (and it is most similar to the measure for the share of time that a subnational unit is in poverty during a specified period), it allows for better comparison with transient poverty and general poverty. 21 We provide additional heterogeneity analyses in Appendix A. For Figure A2, we use the World Bank’s income classifications and divide our sample into two groups: firstly, the low income and lower-middle income countries, and secondly, the upper-middle income and high-income nations. The figure consistently reveals the negative effects of hotter temperatures on chronic poverty. In addition, the effects of colder temperatures are also discernible, but predominantly within the high-income bracket. In Figure A3, we present the estimates of hotter temperature on chronic poverty, adjusted for real GDP per capita on a country-by-country basis. This underscores poor countries’ strong exposure to warming temperatures, with African countries such as Togo, Malawi, and Uganda being particularly vulnerable. 31 Our findings, summarized in Table 3, demonstrate negative effects of hotter temperatures on poverty, and these effects are robust across various measures of poverty. These results are reassuringly consistent with those based on different definitions of chronic poverty shown in Table 2. For example, a 1°C increase in temperature is associated with a 1.6 percentage point increase in chronic poverty at the $3.65/day threshold. The coefficient estimate is stronger when using the poverty line of $6.85/day. We also find negative effects of temperature change on transient poverty, and that the effects on transient poverty are much less pronounced than (i.e., around one-fifth) those on chronic poverty, as confirmed by the equality test for the higher poverty line. Similar to previous analyses, the impacts of rainfall are mixed, generally showing a negative correlation with chronic poverty and a positive but weaker correlation with transient poverty. Overall, these results underscore the need to better distinguish chronic and transient poverty for more effective policy advice. 4.4. Robustness Tests We assess the robustness of our results by examining them through different measures and specifications. We start with the results of contemporary poverty, using different model specifications for the panel and long difference models. The results are presented in Table A1 (Appendix). First, we show that our estimates remain stable regardless of the inclusion of country specific time trends (Column 1). Another issue to address is the potential misspecification of the functional form of temperature. Consequently, in Columns (2) and (3), we employ alternative functional forms of temperature, incorporating quadratic and cubic temperature terms. Results of these exercises provide further support to our main findings. Additionally, we explore different variants of the long differences model using alternative choices for the window length, specifically 4-year and 5-year period. Columns (4) and (5) confirm that our findings are not affected by these alternative window lengths. 32 Subsequently, we perform further tests in Table A2 (Appendix) to ensure the robustness of our results against different temperature measures. We conduct these tests by employing (i) log of temperature (Column 1); (ii) temperature data at 0.5◦ resolution from the Climate Research Unit of the University of East Anglia (CRU) (Column 2); (iii) the number of days when temperature exceeds 28oC (Column 3); (iv) excluding regions where temperature exceeds this level (Column 4); and (v) employing temperature shock, defined as the difference between actual temperature and long-term temperature being greater (less) than 2 (-2) standard deviations (Column 5). The outcomes of these tests show little deviation from our baseline specification. We also conduct alternative measurements of temperature bins for Figure 2 by dividing the temperature distribution within each region into deciles, with the 60th percentile serving as the baseline group. The results presented in Figure A1 (Appendix A) indicate that the effects of extreme weather remain consistent with those observed in Figure 2. Finally, we examine alternative chronic poverty measures. Specifically, we define chronic poverty using the poverty line of $2.15/day (both number of incidences and proportion of incidence), and we also measure the average poverty rate using various window lengths (4- year and 5-year windows). For this test, we employ the fixed-effects model as outlined in Equation (2). The results shown in Table A3 (Appendix A) reaffirm our previous findings. 5. Conclusion We offer an updated and comprehensive review of recent studies on the effects of climate change on poverty and inequality, paying special attention to studies’ data sources as well as empirical methods. Our findings suggest that while studies generally find negative impacts of climate change on poverty, especially for poorer countries, there is less agreement on its impacts on inequality. We further analyze a recent global database at the subnational unit level that is constructed from official national household income and consumption surveys for more 33 than 1,594 subnational units from 134 countries and spanning the period from 2003 to 2019. Our results suggest that temperature change has larger impacts over the short-term than over the long-term and more impacts on chronic poverty than transient poverty. These results are robust to various measures of chronic poverty and are more pronounced for poorer countries. These findings are relevant for policy advice and highlight the importance of tackling chronic poverty to more effectively reduce poverty. Our findings also help emphasize the need of collecting better-quality data, which include panel household survey data (e.g., to better monitor poverty mobility patterns), more disaggregated data (e.g., to more accurately estimate climate change impacts), more use of administrative data in poorer countries, and better integration of different data sources to address their respective limitations. Data imputation methods may also offer promising alternatives in contexts where actual panel data are unavailable. Our analysis suggests that promising future research can further identify the specific channels through which climate change impacts poverty and inequality. It is also useful to better quantify the long-term impacts of climate change on various forms of poverty and inequality, including multi-dimensional aspects. 34 References Adger, W. N. (1999). Social vulnerability to climate change and extremes in coastal Vietnam. World Development, 27(2), 249-269. Akter, S., & Mallick, B. (2013). The poverty–vulnerability–resilience nexus: Evidence from Bangladesh. Ecological Economics, 96, 114-124. Aragón, F. M., Oteiza, F., & Rud, J. P. (2021). Climate change and agriculture: Subsistence farmers’ response to extreme heat. American Economic Journal: Economic Policy, 13(1), 1-35. Arouri, M., Nguyen, C., & Youssef, A. B. (2015). Natural disasters, household welfare, and resilience: evidence from rural Vietnam. World Development, 70, 59-77. Azzarri, C., & Signorelli, S. (2020). Climate and poverty in Africa South of the Sahara. World Development, 125, 104691. Baez, J. E., Caruso, G., & Niu, C. (2020). Extreme weather and poverty risk: evidence from multiple shocks in Mozambique. Economics of Disasters and Climate Change, 4, 103-127. Bangalore, M., Smith, A., & Veldkamp, T. (2019). Exposure to floods, climate change, and poverty in Vietnam. Economics of Disasters and Climate Change, 3, 79-99. Barbier, E. B., & Hochard, J. P. (2018a). The impacts of climate change on the poor in disadvantaged regions. Review of Environmental Economics and Policy, 12(1):26-47. ---. (2018b). Poverty, rural population distribution and climate change. Environment and Development Economics, 23(3), 234-256. ---. (2019). Poverty-environment traps. Environmental and Resource Economics, 74, 1239- 1271. Barreca, A. I. (2012). Climate change, humidity, and mortality in the United States. Journal of Environmental Economics and Management, 63(1), 19-34. Barreca, A., Clay, K., Deschenes, O., Greenstone, M., & Shapiro, J. S. (2016). Adapting to climate change: The remarkable decline in the US temperature-mortality relationship over the twentieth century. Journal of Political Economy, 124(1), 105-159. Barrett, C. B. (2005). Rural poverty dynamics: development policy implications. Agricultural Economics, 32, 45-60. Barrett, C. B., & Carter, M. R. (2013). The economics of poverty traps and persistent poverty: empirical and policy implications. Journal of Development Studies, 49(7), 976-990. Barrios, S., Bertinelli, L., & Strobl, E. (2010). Trends in rainfall and economic growth in Africa: A neglected cause of the African growth tragedy. Review of Economics and Statistics, 92(2), 350-366. 35 Beegle, K., Christiaensen, L., Dabalen, A., & Gaddis, I. (2016). Poverty in a rising Africa. Washington, DC: The World Bank. Boustan, L. P., Kahn, M. E., Rhode, P. W., & Yanguas, M. L. (2020). The effect of natural disasters on economic activity in US counties: A century of data. Journal of Urban Economics, 118, 103257. Bui, A. T., Dungey, M., Nguyen, C. V., & Pham, T. P. (2014). The impact of natural disasters on household income, expenditure, poverty and inequality: evidence from Vietnam. Applied Economics, 46(15), 1751-1766. Burke, M. B., Miguel, E., Satyanath, S., Dykema, J. A., & Lobell, D. B. (2009). Warming increases the risk of civil war in Africa. Proceedings of the National Academy of Sciences, 106(49), 20670-20674. Burke, M., & Emerick, K. (2016). Adaptation to climate change: Evidence from US agriculture. American Economic Journal: Economic Policy, 8(3), 106-140. Cappelli, F., Costantini, V., & Consoli, D. (2021). The trap of climate change-induced “natural” disasters and inequality. Global Environmental Change, 70, 102329. Carter, M. R., Little, P. D., Mogues, T., & Negatu, W. (2007). Poverty traps and natural disasters in Ethiopia and Honduras. World Development, 35(5), 835-856. Cattaneo, C., & Peri, G. (2016). The migration response to increasing temperatures. Journal of Development Economics, 122, 127-146. Cerra, M. V., Lama, M. R., & Loayza, N. (2022). “Links between growth, inequality, and poverty: a survey”. In Cerra et al. (2022). (Eds.). How to Achieve Inclusive Growth. United Kingdom: Oxford University Press. Cevik, S., & Jalles, J. T. (2023). For whom the bell tolls: Climate change and income inequality. Energy Policy, 174, 113475. Chancel, L., Bothe, P., & Voituriez, T. (2023). Climate inequality report 2023, Fair taxes for a sustainable future in the global South. World Inequality Lab (WIL). Chen, C. C., McCarl, B., & Chang, C. C. (2012). Climate change, sea level rise and rice: global market implications. Climatic Change, 110(3-4), 543-560. Chen, J., & Mueller, V. (2018). Coastal climate change, soil salinity and human migration in Bangladesh. Nature climate change, 8(11), 981-985. Chen, S., & Gong, B. (2021). Response and adaptation of agriculture to climate change: Evidence from China. Journal of Development Economics, 148, 102557. Colmer, J. (2021). Rainfall variability, child labor, and human capital accumulation in rural Ethiopia. American Journal of Agricultural Economics, 103(3), 858-877. 36 Cook, N., & Heyes, A. (2020). Brain freeze: outdoor cold and indoor cognitive performance. Journal of Environmental Economics and Management, 101, 102318. Corno, L., & Voena, A. (2023). Child marriage as informal insurance: Empirical evidence and policy simulations. Journal of Development Economics, 162, 103047. Damania, R., Desbureaux, S., & Zaveri, E. (2020). Does rainfall matter for economic growth? Evidence from global sub-national data (1990–2014). Journal of Environmental Economics and Management, 102, 102335. Dang, H. A. H., & Lanjouw, P. (2023). “Regression-based Imputation for Poverty Measurement in Data Scarce Settings”. In Jacques Silber. (Eds.). Handbook of Research on Measuring Poverty and Deprivation. Edward Elgar Press. Dang, H. A. H., Jolliffe, D., & Carletto, C. (2019). Data gaps, data incomparability, and data imputation: A review of poverty measurement methods for data‐scarce environments. Journal of Economic Surveys, 33(3), 757-797. Dang, H. A. H., Nguyen, M. C., & Trinh, T. A. (2023). Does Hotter Temperature Increase Poverty and Inequality? Global Evidence from Subnational Data Analysis. London School of Economics and Political Science, International Inequities Institute Working Paper 104. Dang, H. A. H., Pullinger, J., Serajuddin, U., & Stacy, B. (2023). Statistical performance indicators and index—a new tool to measure country statistical capacity. Scientific Data, 10(1), 146. Dasgupta, S., Laplante, B., Meisner, C., Wheeler, D., & Yan, J. (2009). The impact of sea level rise on developing countries: a comparative analysis. Climatic Change, 93(3-4), 379-388. Datt, G., & Hoogeveen, H. (2003). El Niño or El Peso? Crisis, poverty and income distribution in the Philippines. World Development, 31(7), 1103-1124. Dell, M., Jones, B. F., & Olken, B. A. (2014). What do we learn from the weather? The new climate-economy literature. Journal of Economic literature, 52(3), 740-798. Deryugina, T., & Hsiang, S. M. (2014). Does the environment still matter? Daily temperature and income in the United States (No. w20750). National Bureau of Economic Research. Deryugina, T., & Hsiang, S. (2017). The marginal product of climate (No. w24072). National Bureau of Economic Research. Deschênes, O., Greenstone, M., & Guryan, J. (2009). Climate change and birth weight. American Economic Review, 99(2), 211-217. Deschênes, O., & Moretti, E. (2009). Extreme weather events, mortality, and migration. Review of Economics and Statistics, 91(4), 659-681. 37 Deschênes, O., & Greenstone, M. (2011). Climate change, mortality, and adaptation: Evidence from annual fluctuations in weather in the US. American Economic Journal: Applied Economics, 3(4), 152-185. Diffenbaugh, N. S., & Burke, M. (2019). Global warming has increased global economic inequality. Proceedings of the National Academy of Sciences, 116(20), 9808-9813. Duque, V., Rosales-Rueda, M., & Sanchez, F. (2018). How do early-life shocks interact with subsequent human-capital investments? Evidence from administrative data. In IZA world of labor conference. Erman, A. E., Tariverdi, M., Obolensky, M. A. B., Chen, X., Vincent, R. C., Malgioglio, S., ... & Yoshida, N. (2019). Wading out the storm: The role of poverty in exposure, vulnerability and resilience to floods in Dar Es Salaam. World Bank Policy Research Working Paper, (8976). Erman, A., Motte, E., Goyal, R., Asare, A., Takamatsu, S., Chen, X., ... & Hallegatte, S. (2020). The road to recovery the role of poverty in the exposure, vulnerability and resilience to floods in Accra. Economics of Disasters and Climate Change, 4, 171-193. Ferreira, I. A., Salvucci, V., & Tarp, F. (2022). Poverty, inequality, and growth: trends, policies, and controversies (No. wp-2022-43). World Institute for Development Economic Research (UNU-WIDER). Fishman, R., Carrillo, P., & Russ, J. (2019). Long-term impacts of exposure to high temperatures on human capital and economic productivity. Journal of Environmental Economics and Management, 93, 221-238. Fuller, T. L., Clee, P. R. S., Njabo, K. Y., Tróchez, A., Morgan, K., Meñe, D. B., ... & Smith, T. B. (2018). Climate warming causes declines in crop yields and lowers school attendance rates in Central Africa. Science of the Total Environment, 610, 503-510. Fusco, A., & Van Kerm, P. (2023). “Measuring poverty persistence”. In Jacques Silber. (Eds.). Handbook of Research on Measuring Poverty and Deprivation. Edward Elgar Press. Gilli, M., Emmerling, J., Calcaterra, M., & Granella, F. (2023). Climate Change Impacts on the Within-Country Income Distributions. Available at SSRN 4520461: https://ssrn.com/abstract=4520461. González, F. A. I., Santos, M. E., & London, S. (2022). Multidimensional poverty and natural disasters in Argentina (1970–2010). Journal of Human Development and Capabilities, 23(2), 206-227. Graff Zivin, J., & Neidell, M. (2014). Temperature and the allocation of time: Implications for climate change. Journal of Labor Economics, 32(1), 1-26. 38 Graff Zivin, J., Hsiang, S. M., & Neidell, M. (2018). Temperature and human capital in the short and long run. Journal of the Association of Environmental and Resource Economists, 5(1), 77-105. Graff Zivin, J., Song, Y., Tang, Q., & Zhang, P. (2020). Temperature and high-stakes cognitive performance: Evidence from the national college entrance examination in China. Journal of Environmental Economics and Management, 104, 102365. Gröger, A., & Zylberberg, Y. (2016). Internal labor migration as a shock coping strategy: Evidence from a typhoon. American Economic Journal: Applied Economics, 8(2), 123- 153. Hallegatte, S., & Rozenberg, J. (2017). Climate change through a poverty lens. Nature Climate Change, 7(4), 250-256. Hallegatte, S., Bangalore, M., Bonzanigo, L., Fay, M., Narloch, U., Rozenberg, J., ... & Vogt- Schilb, A. (2017). Shock waves: managing the impacts of climate change on poverty. World Bank. Hallegatte, S., Vogt-Schilb, A., Rozenberg, J., Bangalore, M., & Beaudet, C. (2020). From poverty to disaster and back: A review of the literature. Economics of Disasters and Climate Change, 4, 223-247. Hauer, M. E., Fussell, E., Mueller, V., Burkett, M., Call, M., Abel, K., ... & Wrathall, D. (2020). Sea-level rise and human migration. Nature Reviews Earth & Environment, 1(1), 28-39. Hertel, T. W., & Rosch, S. D. (2010). Climate change, agriculture, and poverty. Applied Economic Perspectives and Policy, 32(3), 355-385. Heyes, A., & Saberian, S. (2019). Temperature and decisions: evidence from 207,000 court cases. American Economic Journal: Applied Economics, 11(2), 238-265. Hirvonen, K. (2016). Temperature change, household consumption, and internal migration: evidence from Tanzania. American Journal of Agricultural Economics, 98(4), 1230-1249. Hsiang, S. M. (2010). Temperatures and cyclones strongly associated with economic production in the Caribbean and Central America. Proceedings of the National Academy of sciences, 107(35), 15367-15372. Hsiang, S., Oliva, P., & Walker, R. (2019). The distribution of environmental damages. Review of Environmental Economics and Policy, 13(1), 83–103. Hsiang, S. M., Burke, M., & Miguel, E. (2013). Quantifying the influence of climate on human conflict. Science, 341(6151), 1235367. 39 Hsiang, S., Kopp, R., Jina, A., Rising, J., Delgado, M., Mohan, S., ... & Houser, T. (2017). Estimating economic damage from climate change in the United States. Science, 356(6345), 1362-1369. Jessoe, K., Manning, D. T., & Taylor, J. E. (2018). Climate change and labour allocation in rural Mexico: Evidence from annual fluctuations in weather. Economic Journal, 128(608), 230-261. Jones, B. F., & Olken, B. A. (2010). Climate shocks and exports. American Economic Review, 100(2), 454-459. Kalkuhl, M., & Wenz, L. (2020). The impact of climate conditions on economic production. Evidence from a global panel of regions. Journal of Environmental Economics and Management, 103, 102360. Karim, A. (2018). The household response to persistent natural disasters: Evidence from Bangladesh. World Development, 103, 40-59. Keerthiratne, S., & Tol, R. S. (2018). Impact of natural disasters on income inequality in Sri Lanka. World Development, 105, 217-230. Kelly, P. M., & Adger, W. N. (2000). Theory and practice in assessing vulnerability to climate change and facilitating adaptation. Climatic change, 47(4), 325-352. Khandker, S. R. (2007). Coping with flood: role of institutions in Bangladesh. Agricultural Economics, 36(2), 169-180. Kleemans, M., & Magruder, J. (2018). Labour market responses to immigration: Evidence from internal migration driven by weather shocks. Economic Journal, 128(613), 2032- 2065. Kotz, M., Levermann, A., & Wenz, L. (2022). The effect of rainfall changes on economic production. Nature, 601(7892), 223–227. Kotz, M., Levermann, A., & Wenz, L. (2022). The effect of rainfall changes on economic production. Nature, 601(7892), 223-227. Koubi, V. (2019). Climate change and conflict. Annual Review of Political Science, 22, 343- 360. Kraay, A., & McKenzie, D. (2014). Do poverty traps exist? Assessing the evidence. Journal of Economic Perspectives, 28(3), 127-148. Kubik, Z., & Maurel, M. (2016). Weather shocks, agricultural production and migration: Evidence from Tanzania. Journal of Development Studies, 52(5), 665-680. Lal, P. N., Lal, P. N., Singh, R., & Holland, P. (2009). Relationship between natural disasters and poverty: a Fiji case study. SOPAC. 40 Letta, M., Adriana, P., & Montalbano, P. (2023). Understanding the Climate Change-Migration Nexus through the Lens of Household Surveys: An Empirical Review to Assess Data Gaps. Journal of Economic Surveys. Doi: https://doi.org/10.1111/joes.12584 Liu, M., Shamdasani, Y., & Taraz, V. (2023). Climate change and labor reallocation: Evidence from six decades of the Indian Census. American Economic Journal: Economic Policy, 15(2), 395-423. Lobell, D. B., Bänziger, M., Magorokosho, C., & Vivek, B. (2011). Nonlinear heat effects on African maize as evidenced by historical yield trials. Nature Climate Change, 1(1), 42-45. LoPalo, M. (2023). Temperature, worker productivity, and adaptation: evidence from survey data production. American Economic Journal: Applied Economics, 15(1), 192-229. Maccini, S., & Yang, D. (2009). Under the weather: Health, schooling, and economic consequences of early-life rainfall. American Economic Review, 99(3), 1006-1026. Marchiori, L., Maystadt, J. F., & Schumacher, I. (2012). The impact of weather anomalies on migration in sub-Saharan Africa. Journal of Environmental Economics and Management, 63(3), 355-374. Mendelsohn, R. (2009). The impact of climate change on agriculture in developing countries. Journal of Natural Resources Policy Research, 1(1), 5-19. Miguel, E., Satyanath, S., & Sergenti, E. (2004). Economic shocks and civil conflict: An instrumental variables approach. Journal of Political Economy, 112(4), 725-753. Missirian, A., & Schlenker, W. (2017). Asylum applications respond to temperature fluctuations. Science, 358(6370), 1610-1614. Molina, O., & Saldarriaga, V. (2017). The perils of climate change: In utero exposure to temperature variability and birth outcomes in the Andean region. Economics & Human Biology, 24, 111-124. Moore, F. C., & Diaz, D. B. (2015). Temperature impacts on economic growth warrant stringent mitigation policy. Nature Climate Change, 5(2), 127-131. Mottaleb, K. A., Mohanty, S., & Mishra, A. K. (2015). Intra-household resource allocation under negative income shock: A natural experiment. World Development, 66, 557-571. Morton, J. F. (2007). The impact of climate change on smallholder and subsistence agriculture. Proceedings of the National Academy of Sciences, 104(50), 19680-19685. Mullins, J. T., & White, C. (2019). Temperature and mental health: Evidence from the spectrum of mental health outcomes. Journal of Health Economics, 68, 102240. Mullins, J. T., & White, C. (2020). Can access to health care mitigate the effects of temperature on mortality? Journal of Public Economics, 191, 104259. 41 Nakamura, S., Abanokova, K., Dang, H. A. H., Takamatsu, S., Pei, C., & Prospere, D. (2023). Is Climate Change Slowing the Urban Escalator Out of Poverty? Evidence from Chile, Colombia, and Indonesia. International Journal of Environmental Research and Public Health, 20(6), 4865. Newell, R. G., Prest, B. C., & Sexton, S. E. (2021). The GDP-temperature relationship: implications for climate change damages. Journal of Environmental Economics and Management, 108, 102445. Nguyen, M. C., Yang, J., Dang, H-A., Sabatino, C. (2023). On the Construction of the World Bank’s Subnational Poverty and Inequality Databases: Documentation. Global Poverty Monitoring Technical Note, 31. World Bank, Washington, DC. Obradovich, N., Migliorini, R., Paulus, M. P., & Rahwan, I. (2018). Empirical evidence of mental health risks posed by climate change. Proceedings of the National Academy of Sciences, 115(43), 10953-10958. Paavola, J. (2017). Health impacts of climate change and health and social inequalities in the UK. Environmental Health, 16(1), 61-68. Paglialunga, E., Coveri, A., & Zanfei, A. (2022). Climate change and within-country inequality: New evidence from a global perspective. World Development, 159, 106030. Park, R. J., Goodman, J., Hurwitz, M., & Smith, J. (2020). Heat and learning. American Economic Journal: Economic Policy, 12(2), 306-339. Park, R. J., Behrer, A. P., & Goodman, J. (2021). Learning is inhibited by heat exposure, both internationally and within the United States. Nature Human Behaviour, 5(1), 19-27. Park, R. J. (2022). Hot temperature and high-stakes performance. Journal of Human Resources, 57(2), 400-434. Pleninger, R. (2022). Impact of natural disasters on the income distribution. World Development, 157, 105936. Ponce, C. (2020). Intra-seasonal climate variability and crop diversification strategies in the Peruvian Andes: A word of caution on the sustainability of adaptation to climate change. World Development, 127, 104740. Jakobsen, K. T. (2012). In the eye of the storm—The welfare impacts of a hurricane. World Development, 40(12), 2578-2589. Jones, B. F., & Olken, B. A. (2010). Climate shocks and exports. American Economic Review, 100(2), 454-459. Rana, I. A., Khan, M. M., Lodhi, R. H., Altaf, S., Nawaz, A., & Najam, F. A. (2023). Multidimensional poverty vis-à-vis climate change vulnerability: Empirical evidence from 42 flood-prone rural communities of Charsadda and Nowshera districts in Pakistan. World Development Sustainability, 2, 100064. Randell, H., & Gray, C. (2019). Climate change and educational attainment in the global tropics. Proceedings of the National Academy of Sciences, 116(18), 8840-8845. Ravallion, M. (2016). The economics of poverty: History, measurement, and policy. Oxford University Press. Rodriguez-Oreggia, E., De La Fuente, A., De La Torre, R., & Moreno, H. A. (2013). Natural disasters, human development and poverty at the municipal level in Mexico. Journal of Development Studies, 49(3), 442-455. Roesch-McNally, G. E., Arbuckle, J. G., & Tyndall, J. C. (2018). Barriers to implementing climate resilient agricultural strategies: The case of crop diversification in the US Corn Belt. Global Environmental Change, 48, 206-215. Rosales-Rueda, M. (2018). The impact of early life shocks on human capital formation: Evidence from El Niño floods in Ecuador. Journal of Health Economics, 62, 13-44. Sarsons, H. (2015). Rainfall and conflict: A cautionary tale. Journal of development Economics, 115, 62-72. Schlenker, W., & Roberts, M. J. (2009). Nonlinear temperature effects indicate severe damages to US crop yields under climate change. Proceedings of the National Academy of sciences, 106(37), 15594-15598. Schlenker, W., & Lobell, D. B. (2010). Robust negative impacts of climate change on African agriculture. Environmental Research Letters, 5(1), 014010. Schoeni, R. F., Stafford, F., McGonagle, K. A., & Andreski, P. (2013). Response rates in national panel surveys. Annals of the American Academy of Political and Social Science, 645(1), 60-87. Sedova, B., Kalkuhl, M., & Mendelsohn, R. (2020). Distributional impacts of weather and climate in rural India. Economics of Disasters and Climate Change, 4, 5-44. Shah, M., & Steinberg, B. M. (2017). Drought of opportunities: Contemporaneous and long- term impacts of rainfall shocks on human capital. Journal of Political Economy, 125(2), 527-561. Sohnesen, T. P. (2020). Two sides to same drought: measurement and impact of Ethiopia’s 2015 historical drought. Economics of Disasters and Climate Change, 4(1), 83-101. Somanathan, E., Somanathan, R., Sudarshan, A., & Tewari, M. (2021). The impact of temperature on productivity and labor supply: Evidence from Indian manufacturing. Journal of Political Economy, 129(6), 1797-1827. 43 Song, Z., Hochman, G., & Timilsina, G. (2023). Natural Disaster, Infrastructure, and Income Distribution. World Bank Policy Research Working Paper 10504. Thiede, B., Gray, C., & Mueller, V. (2016). Climate variability and inter-provincial migration in South America, 1970–2011. Global Environmental Change, 41, 228-240. Thorbecke, E., & Ouyang, Y. (2022). Towards a virtuous spiral between poverty reduction and growth: Comparing sub Saharan Africa with the developing world. World Development, 152, 105776. Thurlow, J., Zhu, T., & Diao, X. (2012). Current climate variability and future climate change: estimated growth and poverty impacts for Zambia. Review of Development Economics, 16(3), 394-411. Tol, R. S. (2021). The distributional impact of climate change. Annals of the New York Academy of Sciences, 1504(1), 63-75. Wheeler, T., & Von Braun, J. (2013). Climate change impacts on global food security. Science, 341(6145), 508-513. Wing, I. S., De Cian, E., & Mistry, M. N. (2021). Global vulnerability of crop yields to climate change. Journal of Environmental Economics and Management, 109, 102462. Wong, P. Y., & Brown, P. H. (2011). Natural disasters and vulnerability: Evidence from the 1997 forest fires in Indonesia. BE Journal of Economic Analysis & Policy, 11(1). Yamamura, E. (2015). The impact of natural disasters on income inequality: analysis using panel data during the period 1970 to 2004. International Economic Journal, 29(3), 359- 374. Zhang, P., Deschenes, O., Meng, K., & Zhang, J. (2018). Temperature effects on productivity and factor reallocation: Evidence from a half million Chinese manufacturing plants. Journal of Environmental Economics and Management, 88, 1-17. 44 Figure 1: Temperature Change Impacts on Poverty Temperature change (+/-) (+/-) (+/-) (-) (-) Econ. Labor Other factors Migration (e.g., rising growth Agriculture productivity, human capital sea levels) coping (+/-) (+) (+/-) (+/-) adaptation Short-term Short-term (+/-) (+) (+/-) (+/-) Poverty inequality Longer-term Longer-term poverty inequality Notes: The positive and negative impacts are shown respectively using the plus (+) and minus (-) signs. Both signs (+/-) indicate mixed impacts. 45 Figure 2: Effects of temperature on chronic poverty by region Notes: The figures show the point estimates and their 95% confidence intervals of temperature bins using regression with rainfall and subnational fixed effects. Robust standard errors are clustered at the subnational level. Chronic poverty is measured by the number of times each subnational unit is in poverty. The reference temperature bin is [15,18). Hotter temperature and colder temperature are defined as temperature being in the top decile (i.e., greater than 27◦C) and bottom decile (i.e., less than 6◦C) of the temperature range, respectively. 46 Table 1: Effects of temperature on poverty Dependent variable: Poverty $3.65/day Poverty $6.85/day Long Fixed-effects Long differences Fixed-effects differences model model model model (1) (2) (3) (4) Panel A: Country-level analysis Temperature 0.007*** 0.004*** 0.022*** 0.015*** (0.002) (0.001) (0.004) (0.002) Rainfall -0.0002 0.0002 -0.0002 -0.0001 (0.0002) (0.0004) (0.0005) (0.0006) Country & Year FEs Yes No Yes No Mean poverty 0.059 0.059 0.210 0.210 Observations 461 134 461 134 Equality test (Panel vs. p = 0.150 p = 0.030 long differences) Panel B: Subnation-level analysis Temperature 0.072*** 0.016*** 0.051** 0.028*** (0.023) (0.001) (0.023) (0.001) Rainfall -0.006** -0.013*** 0.014*** -0.005 (0.003) (0.002) (0.004) (0.004) Subnational & Year FEs Yes No Yes No Mean poverty 0.117 0.117 0.267 0.267 Observations 4,958 1,594 4,958 1,594 Equality test (Panel vs. p = 0.000 p = 0.000 long differences) Equality test (country p = 0.000 p = 0.000 p = 0.000 p = 0.000 vs. subnational) Number of countries 134 134 134 134 Number of regions 1,594 1,594 1,594 1,594 Notes: Robust standard errors in parentheses. Standard errors are clustered at the country level (Panel A) and subnational level (Panel B). Poverty and weather variables in the long-differences model are measured by the difference between 2010 and 2019. The long differences estimation is based on cross-sectional data with a smaller sample size compared with panel data. The equality test p-values show the t-test between the panel FE results vs. the long differences results, and the country analysis results vs. the subnational analysis results. *** p<0.01, ** p<0.05, * p<0.1. 47 Table 2: Effects of temperature on chronic poverty Poverty Poverty Poverty $3.65/day Poverty $6.85/day $3.65/day $6.85/day Number of Proportion of Number of Proportion of Average poverty rate incidences incidence incidences incidence (1) (2) (3) (4) (5) (6) Temperature 0.023*** 0.011*** 0.067*** 0.019*** 0.027*** 0.027*** (0.003) (0.001) (0.005) (0.002) (0.008) (0.008) Rainfall 0.092*** 0.048*** -0.332*** 0.011 0.075*** 0.046*** (0.020) (0.009) (0.041) (0.012) (0.012) (0.012) Subnational & Year No No No No Yes Yes FEs Mean dependent var. 0.361 0.176 0.894 0.351 0.132 0.290 Observations 1,262 1,262 1,262 1,262 278 278 Notes: Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1. 48 Table 3: Effects of temperature on transient vs. chronic poverty Poverty $3.65/day Poverty $6.85/day Chronically Transiently Generally Chronically Transiently Generally poor poor poor poor poor poor (1) (2) (3) (4) (5) (6) Temperature 0.016*** 0.012*** 0.028*** 0.036*** 0.007** 0.043*** (0.003) (0.002) (0.003) (0.005) (0.003) (0.005) Rainfall -0.007*** -0.004** -0.011*** 0.002 0.004* 0.007 (0.002) (0.002) (0.003) (0.004) (0.002) (0.004) Equality test (Chronically poor p = 0.224 p = 0.000 vs. transiently poor) Subnational & Year FEs Yes Yes Yes Yes Yes Yes Mean dependent var. 0.041 0.021 0.063 0.139 0.039 0.178 Observations 935 935 935 935 935 935 Notes: Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1. 49 Appendix A: Additional Tables and Figures Figure A1: Effect of temperature on chronic poverty by region – Alternative bin approach Notes: The figure shows the point estimates and their 95 percent confidence intervals of temperature bins using regression with rainfall and subnational fixed effects. Robust standard errors are clustered at the subnational level. Chronic poverty is measured by the number of times each subnational unit is in poverty. Temperature bins are identified by dividing regional average temperature into deciles with the temperature bin in the 6th decide being the reference group. Hotter temperature and colder temperature are defined as temperature being in the top decile (i.e., bin 10) and bottom decile (i.e., bin 1) of the temperature range, respectively. 50 Figure A2: Effect of temperature on chronic poverty by income group Notes: The figure shows the point estimates and their 95 percent confidence intervals of temperature bins using regression with rainfall and subnational fixed effects. Robust standard errors are clustered at the subnational level. Chronic poverty is measured by the number of times each subnational unit is in poverty. The reference temperature bin is [15,18). Hotter temperature and colder temperature are defined as temperature being in the top decile (i.e., greater than 27◦C) and bottom decile (i.e., less than 6◦C) of the temperature range, respectively. 51 Figure A3: Effect of temperature on chronic poverty across countries adjusted by real GDP Notes: Chronic poverty is measured by the number of times each subnational unit is in poverty. The figure shows the point estimates of temperature and the country dummies using regression with control variable and subnational fixed effects. Each country’s marker is proportional to its real GDP per capital using the World Development Indicator (WDI) database (i.e., a larger size indicates a higher GDP per capita level). 52 Table A1: Robustness tests – Alternative specifications Fixed-effects model Long differences model Adding country- Adding Adding specific linear temperature temperature 4-year average 5-year average time trend squared term cubic term (1) (2) (3) (4) (5) Temperature 0.072*** 0.082*** 0.091*** 0.017*** 0.016*** (0.024) (0.030) (0.029) (0.001) (0.001) Temperature squared 0.000 0.001** (0.000) (0.000) Temperature cubic -0.000 (0.000) Control for rainfall Yes Yes Yes Yes Yes Subnational & Year FEs Yes Yes Yes No No Mean poverty 0.117 0.117 0.117 0.227 0.224 Observations 4,604 4,958 4,958 2,310 1,726 Notes: Poverty is measured using the poverty line of $3.65/day. Robust standard errors in parentheses. Standard errors are clustered ats the subnational level. The long differences estimation is based on cross-sectional data with a smaller sample size compared with panel data. *** p<0.01, ** p<0.05, * p<0.1. 53 Table A2: Robustness tests – Alternative measures of temperature Poverty $3.65/day Number of days Dropping subregions Log Temperature Temperature temperature with temperature temperature from CRU shock above 28 above 28 (1) (2) (3) (4) (5) Temperature 0.155*** 0.090*** 0.001** 0.072*** 0.019*** (0.041) (0.027) (0.000) (0.023) (0.004) Control for rainfall Yes Yes Yes Yes Yes Subnational & Year FEs Yes Yes Yes Yes Yes Mean poverty 0.117 0.117 0.117 0.117 0.117 Observations 4,958 4,927 4,105 4,724 4,958 Notes: Results of fixed-effects model. Robust standard errors in parentheses. Standard errors are clustered at the subnational level. The long differences estimation is based on cross-sectional data with a smaller sample size compared with panel data. *** p<0.01, ** p<0.05, * p<0.1. 54 Table A3: Robustness tests – Alternative measures of chronic poverty Poverty Poverty Average Average $2.15/day $2.15/day poverty rate – poverty rate – (Number of (Proportion of 4-year 5-year incidences) incidence) (1) (2) (3) (4) Temperature 0.004*** 0.002*** 0.015*** 0.021*** (0.001) (0.001) (0.004) (0.003) Control for rainfall Yes Yes Yes Yes Subnational & Year FEs Yes Yes Yes Yes Mean dependent var. 0.079 0.039 0.119 0.091 Observations 1,262 1,262 642 1,008 Notes: Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1. 55 Table A4: Country-level analysis using alternative data from WDI Dependent variable: Poverty $3.65/day Poverty $6.85/day Long Fixed-effects Long differences Fixed-effects differences model model model model (1) (2) (3) (4) Temperature 0.016*** 0.011*** 0.023*** 0.014*** (0.002) (0.001) (0.003) (0.001) Rainfall -0.000 -0.000 -0.000 -0.000 (0.000) (0.000) (0.000) (0.000) Country & Year FEs Yes No Yes No Observations 1,847 420 1,846 420 Equality test (Panel vs. p = 0.000 p = 0.000 long differences) Notes: Robust standard errors in parentheses. Standard errors are clustered at the country level. Poverty and weather variables in the long-differences model are measured by the difference between 2010 and 2019. The long differences estimation is based on cross-sectional data with a smaller sample size compared with panel data. The equality test p-values show the t-test between the panel FE results vs. the long differences results. *** p<0.01, ** p<0.05, * p<0.1. 56 Table A5: Alternative measures of rainfall Poverty: $3.65/day $6.85/day $3.65/day $6.85/day $3.65/day $6.85/day $3.65/day $6.85/day (1) (2) (3) (4) (5) (6) (7) (8) Temperature 0.074*** 0.048** 0.078*** 0.049** 0.070*** 0.054** 0.069*** 0.055** (0.024) (0.023) (0.024) (0.023) (0.023) (0.023) (0.023) (0.023) Rainfall -0.009* 0.020*** (0.005) (0.006) Rainfall squared 0.000 -0.000*** (0.000) (0.000) Rainfall deviation -0.006** 0.006** (0.002) (0.002) Rainfall > 90th percentile -0.001 0.007 (0.006) (0.007) Rainfall > 95th percentile 0.006 0.010 (0.010) (0.011) Subnational and Year FEs 0.057 0.054 0.058 0.051 0.056 0.049 0.056 0.050 Observations 4,958 4,958 4,958 4,958 4,958 4,958 4,958 4,958 Notes: Results of fixed-effects model using subnational data. Robust standard errors in parentheses. Standard errors are clustered at the subnational level. *** p<0.01, ** p<0.05, * p<0.1. 57