Droughts and domestic violence: Measuring the gender-climate nexus

Aguilar-Gómez, Sandra; Salazar-Díaz, Andrea Working Paper Droughts and domestic violence: Measuring the genderclimate nexus IDB Working Paper Series, No. IDB-WP-1653 Provided in Cooperation with: Inter-American Development Bank (IDB), Washington, DC Suggested Citation: Aguilar-Gómez, Sandra; Salazar-Díaz, Andrea (2025) : Droughts and domestic violence: Measuring the gender-climate nexus, IDB Working Paper Series, No. IDB-WP-1653, InterAmerican Development Bank (IDB), Washington, DC, https://doi.org/10.18235/0013368 This Version is available at: https://hdl.handle.net/10419/309193 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/3.0/igo/ Droughts and Domestic V iolence: Measuring the Gender-Climate Nexus Sandra Aguilar-Gómez Andrea Salazar-Díaz WORKING PAPER No IDB-WP-1653 InterA merican Development Bank Gender and Diversity Division January 2025 * Universidad de los Andes ** Kennedy School of Government, Harvard University Droughts and Domestic V iolence: Measuring the Gender-Climate Nexus Sandra Aguilar-Gómez* Andrea Salazar-Díaz** InterA merican Development Bank Gender and Diversity Division January 2025 Cataloging-in-Publication data provided by the Inter-American Development Bank Felipe Herrera Library Aguilar Gómez, Sandra. Droughts and domestic violence: measuring the gender-climate nexus / Sandra Aguilar-Gomez, Andrea Salazar-Diaz. p. cm. — (IDB Working Paper Series ; 1653) Includes bibliographic references. 1. Intimate partner violence-Mexico. 2. Women-Violence against-Mexico. 3. Climatic changes-Social aspects-Mexico. 4. Gender mainstreaming-Mexico. I. Salazar-Diaz, Andrea. II. Inter-American Development Bank. Gender and Diversity Division. III. Title. IV. Series. IDB-WP-1653 http://www.iadb.org Copyright © 2025 Inter-American Development Bank ("IDB"). This work is subject to a Creative Commons license CC BY 3.0 IGO (https://creativecommons.org/licenses/by/3.0/igo/legalcode). The terms and conditions indicated in the URL link must be met and the respective recognition must be granted to the IDB. Further to section 8 of the above license, any mediation relating to disputes arising under such license shall be conducted in accordance with the WIPO Mediation Rules. Any dispute related to the use of the works of the IDB that cannot be settled amicably shall be submitted to arbitration pursuant to the United Nations Commission on International Trade Law (UNCITRAL) rules. The use of the IDB's name for any purpose other than for attribution, and the use of IDB's logo shall be subject to a separate written license agreement between the IDB and the user and is not authorized as part of this license. Note that the URL link includes terms and conditions that are an integral part of this license. The opinions expressed in this work are those of the authors and do not necessarily reflect the views of the Inter-American Development Bank, its Board of Directors, or the countries they represent. Abstract* Every year, 245 million women are victims of intimate partner violence (IPV). Climate change is hypothesized to exacerbate this figure through its disruptive impact on household livelihoods, among other channels. However, little causal evidence exists on this aspect of the climate-gender nexus, partly due to measurement challenges that have contributed to gaps in the literature. In this paper, we use three different IPV data sources to examine the effect of drought in Mexico and the role of agricultural vulnerability in intensifying these effects. We find robust evidence of increases in all measures of IPV in response to local precipitation deficits: as unanticipated exposure to days without rain in the previous month rises, more injuries linked to IPV are recorded in the public health system, police reports increase, and more 911 calls related to IPV are made. The effects are stronger in regions highly dependent on agriculture, particularly when the shock occurs during the growing season. In a country where most agricultural income and land are controlled by men, our results align with theoretical predictions from male-backlash IPV models and extractive violence models. We also find that the impact of drought on IPV is more pronounced in municipalities with low state capacity, though potential differences in reporting behavior between IPV measures complicate comparisons. Our findings underscore the need to design gender-sensitive disaster relief policies, strengthen trust in reporting mechanisms and helplines, and reduce the social acceptability of IPV. JEL classifications: I15, J16, Q54, D13 Keywords: Drought, Domestic violence, Intimate partner violence, Agriculture, Mexico ∗This project was partially funded by the Inter-American Development Banks (IDB) Gender and Diversity Lab. We thank the seminar participants of this program for their helpful questions and feedback. We are also grateful to Andrés Moya and Amber Peterman for their comments and suggestions and to the IDB GDLab team for their continued support: Karen Martínez, Monserrat Bustelo, and Claudia Martínez. We thank David Chaparro-Álvarez, Angel Ramírez and Valentina Castilla-Gutiérrez for their research assistance. 1 Introduction Intimate partner violence (IPV) is one of the most pressing challenges to women’s health worldwide: approximately 245 million women—equating to 10% of those aged 15 and above—are subjected to either psychological, sexual, or physical violence by their partners every year. In Mexico, the focus of our study, the prevalence of IPV reflects global patterns. According to the World Bank Gender Portal, 24% of Mexican women have experienced IPV, closely aligning with the global average of 26% (World Bank,2022b). Despite this high prevalence, reporting rates are low: only one in 10 women reported experiencing violence in the past 12 months (ENDIREH,2021). Thus, efforts to improve policies aimed at protecting women are hampered by a lack of high-quality data. Furthermore systematic data collection is recent in many regions with high prevalence (Cullen,2023). The effects of IPV reach beyond immediate physical or emotional harm, significantly influencing long-term health outcomes (Campbell,2002), reducing labor market participation (Bhattacharya,2015;Sabia et al.,2013), and diminishing overall well-being (Oduro et al.,2015;Adams et al.,2012). IPV has been poised as a key dimension of the climate-gender nexus. Climate change, which is anticipated to escalate the frequency and severity of extreme environmental events (Arias et al., 2021), interacts with existing inequalities (Keerthiratne and Tol,2018;Warr and Aung,2019; Chowdhury et al.,2021)), including gender (Gevers et al.,2020;UN Women,2022), disproportionately affecting women’s mortality (Neumayer and Plümper,2007), morbidity (Fatema et al.,2019), and housing security (Logie et al.,2017). 1An environmental hazard of particular concern are droughts, which are among the most expensive natural disasters in the world, with significant costs to ecosystems, agriculture, and human livelihoods (Cook et al.,2018). Climate change is expected to increase drought frequency and severity over much of the globe (Dai,2013). By 2050, an additional 1 billion people are expected to live with extremely high water stress2(Kuzma et al.,2023). This paper seeks to understand the causal impact of drought on IPV in Mexico, a country with both tropical and semi-arid environments, where 40-80% of the population are exposed annually to water stress (Kuzma et al.,2023). Agriculture is a particularly climate-sensitive sector, and in Mexico, it is also predominantly male-biased in terms of labor composition, as well as land ownership and governance. Women constitute only 12.7% of the country’s agricultural laborers. Furthermore, while half of the land is under a communal property and management system known as ejidos, women hold only 26% of ejido rights, and out of the 14.6 thousand ejidos and communities with representative bodies, only 1Some ways in which gender norms reduce women’s resilience include their lower control over income and assets to face shocks (Van Aelst and Holvoet,2016;Roncoli et al.,2001) and facing mobility restrictions that limit their access to aid and healthcare (Parmar et al.,2019;Boyer et al.,2020) 2Water stress, the ratio of water demand to renewable supply, measures the competition over local water resources. The smaller the gap between supply and demand, the more vulnerable a place is to water shortages. 1 7.4% are presided over by a woman (Registro Agrario Nacional,2019). Under these conditions, an average decrease in crop yields resulting from droughts directly reduces the relative income of men compared to women. Economic theories of IPV provide differing predictions on the impact of disasters on IPV. Bargaining power models suggest that a decrease in men’s relative income should increase women’s bargaining power within the household and consequently decrease domestic violence (Aizer,2010;Anderberg et al.,2016). On the other hand, extractive violence models argue that reduced income may increase male partners’ incentives to use violence as a tool to regain control over household decision-making or extract financial resources, to compel women to work longer hours, or as a response to stress over income loss (Eswaran and Malhotra,2011;Bobonis et al.,2013;Erten and Keskin,2018,2021,2024). In our study, we investigate the impacts of drought exposure on IPV in Mexican municipalities. We combine gridded precipitation, and temperature estimates data from NASA Daymet repository (Thornton et al.,2022) to generate disaster exposure measures with high temporal and spatial granularity with administrative data from various sources. The identifying assumption is that these events are unexpected to households once controlling for a battery of geographical and seasonal fixed effects and thus plausibly exogenous. To overcome the measurement challenge, we employ three measures of IPV with distinct strengths and weaknesses. First, we utilize hospital discharge records reported by healthcare institutions from Mexico’s largest public health subsystem. We aggregate these data at the weekly level to produce municipality-week measures of the incidence of injuries that can be disaggregated into IPV and nondomestic violence. Second, we incorporate police files from the National Security System (SNSP) for domestic violence data. Finally, through hundreds of information requests to local law enforcement agencies, we collected data from local government hotlines that receive emergency calls about IPV. Overall, two out of three of our gender-based violence (GBV) measures directly capture IPV. Although we can only measure domestic violence in police files without differentiating by sex, public hospital records show that 82% of domestic violence is IPV, and 95% of IPV victims in Mexico are women. Therefore, throughout the document, we will primarily refer to our results as IPV. With the three mentioned measures, we conduct our analyses at the weekly and monthly levels and quantify the effect on violence of each additional extreme day in a given week. Our findings reveal a robust causal effect of drought on IPV and non-domestic violence towards women and non-domestic violence towards men. The rate of female hospitalizations for non-domestic violence increases by 7.4%, and for male hospitalizations, the increase is 14%. IPV rates display a gendered pattern; female hospitalizations for IPV increase by up to 9% during severe drought periods, while there is no statistical impact on male IPV hospitalizations. Using police reports and 911 hotline calls as alternative dependent variables, we consistently find marked increases in IPV during droughts. Regions with lower state capacity exhibit higher increases in police reports and hospitalizations, suggesting that state capacity influences the deterrence effect 2 of IPV. In a deterioration of the rule of law, under-reporting biases might increase. With this mechanism operating against our ability to capture increases in crime, detecting such increases in low-state capacity counties—ceteris paribus—suggests that the rise in incidence outweighs the possible bias in reporting. The setting of our study is relevant for this theoretical debate because droughts do not usually destroy infrastructure or property, thereby not affecting some common triggers of violence such as the deterioration of state capacity, destruction of roads, which weakens support systems for women, and emergency displacement (Thurston et al.,2021;Gevers et al.,2020;UN Women,2022). This unique characteristic of droughts allows us to isolate the economic mechanisms, particularly income loss—or stress due to anticipated income loss, a mechanism especially underscored in the public health literature (Thurston et al.,2021)—to test existing theories of IPV. We examine the mechanisms behind our main findings, highlighting the role of agricultural downturns in a country where the bulk of the workforce and land ownership in this sector is male. We find that droughts significantly increase IPV in regions with high agricultural reliance. Our results align more robustly with the "male backlash" model, where male status inconsistency heightens IPV as men seek to assert control within the household or extract resources to compensate for income losses. We find that effects are more pronounced when the drought hits during the municipalityand crop-specific growing seasons, underscoring the role of immediate economic stress once farmers observe low precipitation levels. Our work contributes to the literature on IPV measurement, which is an important challenge in identifying its causes. Previous works have utilized survey data (e.g., in Díaz and Saldarriaga (2023); Sviatschi and Trako (2024)), police reports (Miller et al.,2024) hospital records (Aizer, 2010;Sviatschi and Trako,2024), and recently, emergency calls (Leslie and Wilson,2020;Agüero, 2021;Miller et al.,2024). While recent papers have highlighted the advantages of hotline calls as a measure of violence against women, in this paper we discuss their advantages and drawbacks by comparing this measure to several years of police reports and hospital records at the national level. We confirm that hospitalizations and police reports experience higher increases in municipalities with low local state capacity. However, further research is required to draw conclusions about 911 calls. In any case, our findings highlight the importance of creating accessible support services for women to seek help and report instances of IPV. This includes the significance of governments gathering and making public measures of IPV. Our collection of datasets enables us to analyze predominantly public space violence and compare it with IPV, which mainly occurs at home. This lets us show that, while generalized violence responds to droughts, women face an additional burden at home. With these insights, our study contributes to a small but growing body of quantitative analyses that investigate the relationship between droughts and gender violence. This literature has focused on low-income 3 countries, frequently finding significant impacts (Hirvonen et al.,2023;Epstein et al.,2020;Munala et al.,2023a;Abiona and Koppensteiner,2018;Cools and Kotsadam,2017;Cooper et al.,2021). However, middle-income countries, home to 75% of the world’s population, remain unexplored. Our findings of the strong effects of droughts on violence in Mexico, the twelfth-largest economy in the world, suggest that women’s vulnerability to climate change will not solve itself through economic development. As climate change unravels, more policies addressing the gendered impacts of climate extremes will be required.3 We also contribute to understanding the climate-gender nexus through disasters. Thurston et al. (2021) finds in a literature review that quantitative studies with a greater geographical scope and use of spatial data are required to understand the magnitude of the problem and the mechanisms behind it. They identify that some authors of quantitative papers present hypothesized pathways from disaster exposure to increased violence against women and girls (VAWG) but fail to test it empirically. While some work has looked at the effects of drought on self-reported measures of IPV using repeated cross-section survey data (Díaz and Saldarriaga,2023;Cools et al.,2020), most of the current literature comprises qualitative investigations (Ylipaa et al.,2019) and casespecific analyses of individual extreme events (van Daalen et al.,2022). Our findings are consistent with a model in which men react to income loss and stress by exerting more violence toward their partners. More broadly, we contribute to our understanding of the climate-violence nexus. We find that the effect of drought on violence against women is larger than the climate-conflict estimates reported in many previous studies4. We estimate an increase between 9% and 26%—depending on the variable employed to measure IPV—from the sample mean when a full month of unexpected drought hits a municipality. The magnitude of these effects in a middle-income country like Mexico suggests that the effects of gendered violence are a crucial, previously unexplored dimension of the climate-violence literature. The remainder of this paper is organized as follows. Section 2provides a theoretical background and context, Section 3describes the data used in our analysis and presents descriptive statistics. Section 4lays out our empirical strategies. Section 5presents our baseline results and analyzes the mechanisms, heterogeneity, and robustness checks. Section 6concludes. 3Figure A.1 shows that Mexico’s GDP per capita is between 1.5 and 6 larger than other countries studied in the quantitative drought-IPV literature: 2.8 times the GDP per capita of India, 5.3 times the Sub-Saharan Africa average, and 1.6 times the GDP per capita of Peru, the only other Latin American country studied in this literature. 4In a review of this literature, Hsiang et al. (2013) find that the effect of temperature is generally larger than the standardized effect of rainfall, and the effect on intergroup violence (e.g., civil war) is larger than the effect on interpersonal violence (e.g., assault). Particularly, for each standard deviation change in climate toward warmer temperatures or more extreme rainfall, median estimates indicate that the frequency of interpersonal violence rises 4% and the frequency of intergroup conflict rises 14%. 4 a) 911 call data availability by state b) Coverage period State Availability Aguascalientes 2005-2021 Bajacalifornia, Sonora 2006-2021 Chiapas 2007-2021 Baja California Sur 2015-2021 Nuevo Leon, Veracruz 2016-2021 Distrito Federal, Puebla, Sinaloa, San Luis de Potosi, Campeche, Quintana Roo 2017-2021 Mexico, Guanajuato, Michoacan, Guerrero, Hidalgo, Tabasco, Yucatan, Nayarit 2018-2021 Jalisco, Chihuahua 2019-2021 Oaxaca 2010-2021 Zacatecas 2011-2021 Figure 1: 911 data availability Note: Panel a) shows the availability of hotline data by state. States are sorted by population size according to the 2010 Census. Panel b) Shows the specific years available for each state that provided information through the transparency platform. 3.1.4 Descriptive statistics While neither police records nor hotline data register the sex of the victim, we know from public hospital records that 82% of domestic violence is IPV and that 95% of IPV victims in Mexico are women. Therefore, throughout the results analyses, we will primarily refer to our results as IPV. Intimate partner violence is a highly feminized crime and the most common form of violence 11 against women, as shown in Table 1. The table is also informative in terms of illustrating the “iceberg" of domestic violence. Every month in Mexican municipalities, 208 emergency calls per 100,000 inhabitants are made on average in response to IPV and 67 incidents are reported to the police. Every week, 23 women per 100,000 inhabitants seek healthcare due to violence perpetrated by their partner. Table 1: Violence data and descriptive statistics Panel A: Hospital weekly records rate per 100,000 inhabitants N Mean SD Min Max Female victims of assault 1,003,808 2.55 11.5 0 1149 Male victims of assault 1,003,808 3.0 12.74 0 658 Female victims of IPV 1,003,808 5.71 29.2 0 3349 Male victims of IPV 1,003,808 0.11 1.85 0 554 Female victims of child maltreatment 1,003,808 0.42 4.01 0 707 Male victims of child maltreatment 1,003,808 0.09 1.70 0 455 Female victims of sexual violence 1,003,808 0.28 2.82 0 233 Male victims of sexual violence 1,003,808 0.025 0.78 0 220 Panel B: Monthly crime reports rate per 100,000 inhabitants N Mean SD Min Max Domestic Violence 193,380 66.77 213 0 8940 Panel C: Monthly hotline data rate per 100,000 inhabitants N Mean SD Min Max IPV 110,174 208 1225 0 185913 Notes: Panel A) Hospital Discharge Subsystem (SAEH) (2012-2019), weekly data from 75% of Mexican municipalities. Panel B) National Security System (SNSP) (2015-2021), monthly data from all Mexican municipalities. Panel C), Mexico’s 911 System (2016-2021) Monthly data from municipalities with available data as described in Figure 1, covering the place of residence of 76% of the national population. 3.2 Drought We construct a measure of exposure that counts the anomalous days without rain in various reference periods (e.g., past week, past month). Anomalous days are unexpected days without rain in a specific municipality and week of the year. We use precipitation data from Daymet (Thornton et al.,2022), a data product provided by NASA’s Oak Ridge National Laboratory (ORNL). This dataset provides data on a 1 km x 1 km gridded surface over continental North America since 1980. The estimates are generated through the interpolation and extrapolation of daily meteorological observations, with model inputs encompassing a digital elevation model, derived horizon files, and a land-water mask. The resulting output provides gridded estimates of total daily precipitation. We selected this source of weather information not only because it has a high spatial resolution but 12 also because it has the best performance for daily precipitation in Mexico (Rincón-Avalos et al., 2022). We also process the average daily maximum temperature to control for this factor. We spatially average Daymet pixels using municipality boundaries to estimate total precipitation levels and maximum temperature for each municipality’s surface from 2012 to 2021. Appendix Figure A.2 illustrates the Daymet resolution compared to one of the smallest municipalities in Mexico. The amount of precipitation and temperature varies throughout the year and across the Mexican territory (Figures 2and A.3), supporting the external validity of our analyses. The dry season spans approximately from November to April and is characterized by lower precipitation levels across most of the country, with January (left) exemplifying this period’s conditions. Conversely, most of the country’s rainy season extends from late May to October, when rainfall significantly increases. Figure 2(right) shows that monthly total precipitation can reach 300 milliliters per month. Figure 2: Precipitation data Source: ORNL DAACDaymet Version 4 Revision 1; authors’ formulation. 3.3 Local characteristics Municipality boundaries are obtained from the 2010 Geostatistical Framework version 5.0 (INEGI, 2010)11. We use individual-level occupation data from the Mexican Census conducted in 2010 (INEGI,2010) to distinguish between agricultural and non-agricultural municipalities. From this Census, we define agricultural municipalities as those where the percentage of the population working in agriculture is above the national median (39 percent). The distribution of this variable is shown in Appendix Figure A.4. The municipality-level population used to calculate violence rates is also obtained from this source. 11The 2010 map is divided into 2,546 municipalities. Since some municipalities were created after this framework, all databases were adjusted to maintain these 2,456 municipalities. 13 3.3.1 Municipality-specific growing seasons To examine whether potential income loss in agriculture is a source of increased domestic violence, we construct a municipality-specific indicator of whether the shock coincides with the growing season of the municipality’s main crops in terms of total revenue. To identify the main crop, we employ data from the yearly statistical reports published by the Agricultural and Fishery Information Service SIAP (2024). This report contains municipality-by-crop-by-cycle yields and revenue. We identified the municipality’s main crop-cycle12 based on the aggregate revenue between 2010 and 2018. Next, we identify the relevant growing season for each crop using the manuals for growing each crop published by the National Institute of Forestry, Agricultural and Livestock Research (INIFAP). We complemented this information with technical documents from the USDA and the FAO. A detailed description of the data sources employed to identify the growing season of each crop is displayed in Table A.1. 3.3.2 State capacity We construct a state capacity index with municipal-level variation. Based on previous works seeking to capture subnational differences in state capacity, our index is based on two core dimensions: administration and extraction (Soifer,2012). We employ sewerage and tap water coverage and local state employees per capita as proxies for administration capabilities. Data for these variables come from the Censo de Población y Vivienda 2010 reported by INEGI. Extraction capacity is proxied with fiscal autonomy, measuring municipality taxes collected as a share of total resources. This information is obtained from Finanzas Públicas Estatales y Municipales (EFIPEM). All variables are standardized and then used to create an additive index. This index then ranges between -3.13 and 1.65, where lower numbers indicate worse state capacity. 4 Empirical framework The first step in our empirical strategy is to quantify the causal impact of drought on IPV. We follow the standard climate econometrics literature (Dell et al.,2014;Hsiang,2016) and create a municipality-level drought exposure metric that counts unexpected days without rain within specific reference periods. We estimate the following equation: DV mt = 5 X b=1 χbDb,mt + Λmt +γs(t) + δmy +θt+εmt (1) 12Different regions might grow the same crop at different cycles due to agro-climatic differences. For instance, as shown in Table A.1, the majority of counties’ most important non-perennial crop is corn grain, but not all of them grow between April and October 14 In equation 1, the dependent variable, DVmt—the IPV hospitalization, calls or police reports per 100,000 inhabitants in municipality m, and time (week or month) t—is modeled as a flexible semi-parametric function of drought Dmt. Specifically, Db,mt are dummy variables indicating the number of abnormal days without precipitation in the past month/four weeks, which allows for potential non-linearities. For the monthly variables (police records and hotline calls), Dmt represents the days without rain within the current calendar month. For hospital records, which are on a weekly timescale and matched to weekly weather indicators, we use the sum of days without rain over the past four weeks (weeks w∈[0,−3]), so Dmt =P3 w=0 Daysnorainm,t−w.13 We define these unexpected days based on deviations from historical precipitation norms for each municipality and time of the year, incorporating municipality-by-year and date-fixed effects. The methodology is designed to leverage exposure to precipitation levels below those to which regions have historically adapted. Λmt includes controls for other extreme weather events in time t(days above/below the 5th and 95th percentile of temperature and days above the 95 percentile of precipitation). We also include state-specific time trends γs(t)to capture regional development patterns that could correlate with precipitation trends. δmy are municipality-year fixed effects and θtare date fixed effects. Municipality fixed effects capture time-invariant characteristics specific to each municipality, such as local institutions, socio-economic conditions, and cultural factors that might influence IPV rates. These effects account for any persistent differences across municipalities, including baseline climate and prevalence of drought. Interacted with the year, we allow more flexibility for these different conditions to change yearly. Date fixed effects (θt) control for standard shocks that affect all municipalities simultaneously, such as national holidays, economic downturns, or significant political events. By including these, we adjust for temporal variations that could confound the relationship between drought and IPV. Lastly, state-specific time trends (γs(t)) address regional patterns that evolve, such as infrastructure improvements, economic development, demographic changes, and even region-specific global warming, ensuring that our results are not driven by long-term trends within states that could be correlated with both drought and IPV. Our identification assumption is that we are capturing anomalous drought days—–deviations from expected precipitation (i.e., precipitation that could be predicted by the fixed effects described above)—–which are therefore orthogonal to other relevant drivers of IPV. 5 Results We begin our analysis with the results from equation 1, which flexibly estimates the relationship between the intensity of anomalous drought conditions over the past month and the rate of IPV. 13Since these two measures (days in the past four weeks vs. days in the current month) have different upper limits (28 vs. 31), we use different bins to accurately capture their variability. 15 The dependent variable in this analysis is operationalized through three distinct measures, each representing a different type of interaction between public services and IPV victims: hospital records, police reports, and hotline calls. As discussed above, for hospital records and hotline calls, we can obtain IPV rates, but for police reports, the data aggregates all forms of domestic violence. Only in the case of hospital records are we able to separate incidents by the sex of the victim. 5.1 Hospital records Figure 3shows the impact of anomalous drought conditions in the preceding four weeks on the municipality-level rate of hospital admissions for assault cases involving female victims (a) and male victims (b) in a given week. The y-axis represents the total anomalous days without rain in that week and the three previous ones, grouped in bins. The omitted category includes weeks with exposure to less than a total of 10 anomalous days of drought during the previous weeks t−w, w ∈0,−3. The figure shows a monotonic relationship between drought and violence against men, while nondomestic violence against women plateaus after 19 days of drought. The rate of female hospitalizations for non-domestic violence increases by 0.2 when the number of days of drought in the past four weeks exceeds 26. This amounts to a 7.4% increase from the sample mean of 2.5 weekly cases per 100,000 inhabitants. The male non-domestic violence rate displays a rise of 0.43 in the driest bin, which amounts to a 14% increase from a sample mean of 3.0 weekly cases per 100,000 inhabitants. Overall, drought impacts nondomestic violence against men and women, with a stronger increase for men but significantly affecting all the population. Intimate partner violence tells a different story, displaying a markedly gendered relation to drought. The number of victims hospitalized every week in the municipality increases as the intensity of the unexpected lack of rain experienced in the preceding four weeks increases (Figure 4(a)). There is an average increase in hospitalizations of female victims of IPV, up to 9% of the sample mean, which represents 5.2 weekly cases per 100,000 inhabitants. On the other hand, there is no statistically significant impact on male hospitalizations for IPV (Figure 4(b)). We find no discernible impacts on hospitalizations for child maltreatment, as shown in Appendix Figure A.5. 16 (a) Female victims (b) Male victims Figure 3: Anomalous days of drought and non-domestic violence Notes: The dependent variables are weekly hospital admissions within the Mexican public health system from 2012 to 2019, categorized according to ICD-10 codes. The regression model adjusts for other extreme weather events, including days with temperatures above or below the 5thand 95thpercentiles and days exceeding the 95thpercentile of precipitation. The model incorporates municipality-year and date fixed effects, state-specific time trends, and controls for municipality population sizes. Error bars represent 95% confidence intervals. (a) Female victims (b) Male victims Figure 4: Anomalous days of drought and IPV Notes: The dependent variables are weekly hospital admissions within the Mexican public health system from 2012 to 2019, categorized according to ICD-10 codes. The regression model adjusts for other extreme weather events, including days with temperatures above or below the 5thand 95thpercentiles and days exceeding the 95thpercentile of precipitation. The model incorporates municipality-year and date fixed effects, state-specific time trends, and controls for municipality population sizes. Error bars represent 95% confidence intervals. The differential impact of drought on non-domestic violence can be understood through the lens of economic and psychological theories. Economic distress induced by drought may heighten household stress and conflict due to financial insecurity, as discussed in the economic literature on 17 resource scarcity and violence in Section 2.2. Men may exhibit increased aggression as a coping mechanism for stress and financial pressures, behavior consistent with the social stress theory, which posits that environmental stress can exacerbate aggressive behaviors (Jia,2014). This type of violence may manifest against both male and female victims. The more marked increase for men could be due to the fact that they spend more time in public spaces, leading to more interactions that can potentially end in violence. Finally, in a model of rational criminal behavior, the relative returns to criminal activity increase when income from legal activities decreases. If some of these additional crimes committed involve violence as a means of coercion, there is an expected increase in aggressions against men and women, with men once again being disproportionately impacted as they occupy more public spaces. 5.2 Police files and hotline calls Hospital records of IPV only capture cases severe enough to warrant an encounter with the health system and thus represent “the tip of the iceberg” of the problem. Hence, in this Section, we employ two additional data sources to show that our results using hospital records are robust to other measures of IPV. Alternative measures have their drawbacks, too. First, in Mexico, IPV is rarely reported to the police. The impunity rate for this crime is 98.6% nationally, and it surpasses 90% in every state (ENVIPE,2022). Second, as noted before, from police reports and hotline calls, we cannot distinguish the sex of the victim. There is no reliable data on the percentage of IPV victims or witnesses who call 911 during or after an episode. However, in principle, this data might suffer from fewer biases than the police reports since it connects women with more services and can be made by a witness, such as a neighbor or relative. The results of implementing equation 1using monthly data from the sources discussed above are shown in Figure 5. Police files and hotline calls exhibit consistent patterns in response to the number of unexpected days without rain in the current month. In the highest bin, indicating more than 28 days without rain, police reports and hotline calls increase by 18 files per 100,000 inhabitants and 44 calls per 100,00 inhabitants per month. These marked increases diverge 26% and 21%, respectively, from the sample mean. 18 (a) Police files (b) Hotline calls Figure 5: Anomalous days of drought and domestic violence Notes:The dependent variables are monthly police files (a) and IPV hotline calls (b), grouped based on their date of occurrence. Police records cover 2015-2021, and hotline calls include 2016-2021. The regression model adjusts for other extreme weather events, including days with temperatures above or below the 5thand 95thpercentiles and days exceeding the 95thpercentile of precipitation. The model incorporates municipality-year and date fixed effects, state-specific time trends, and controls for municipality population sizes. Error bars represent 95% confidence intervals. Hotline calls face a significant challenge due to insufficient geographical coverage, limiting the ability to conduct representative regional heterogeneity analysis. As an emerging data source, only eight out of 32 states provided data over a sufficiently long period (2016-2021) to enable a fixed effects estimation that captures the impact of extreme events. However, this dataset offers the advantage of including reports on various forms of violence, which, due to their prevalence, significantly affect women’s everyday lives. It reveals a larger portion of the proverbial iceberg. Figure 6examines the incidence of some of these crimes. While reports of rape and sexual harassment do not appear to be influenced by drought conditions, reports of sexual abuse increase monotonically, with coefficients becoming statistically significant in months with more than 25 days of drought. The rise in sexual abuse reports is particularly concerning: according to Mexican legislation, the distinction between sexual abuse and rape hinges on whether there was vaginal, anal, or oral penetration14, but both acts involve non-consensual sexual contact. Needless to say, both crimes can inflict significant trauma on women. 14Articles 178 and 179 of the penal code. 19 (a) Rape (b) Sexual Abuse (c) Sexual harrassment Figure 6: Anomalous days of drought and other forms of GBV Notes: The dependent variables are expressed as a rate per 100,000 inhabitants. Hospital records cover 2012-2019, police records cover 2015-2021, and hotline calls include 2016-2021. The regression model adjusts for other extreme weather events, including days with temperatures above or below the 5thand 95thpercentiles and days exceeding the 95thpercentile of precipitation. The model incorporates municipality-year and date-fixed effects and state-specific time trends. Error bars represent 95% confidence intervals. Altogether, we find a robust relationship between drought and IPV. Analogously to the temperature-violence link identified in Mexico (Baysan et al.,2019), the contemporary and linear response of crime to drought identified in this paper suggests that psychological factors have potential explanatory power.15. However, as discussed in Section 2.2, mechanisms driving violence against women may differ from typical inter/intragroup violence. In the following sections, we provide suggestive evidence that potential agricultural income loss is a potential mechanism driving 15(Baysan et al.,2019) find that the response of suicide to temperature in Mexico strongly matches the responses of interpersonal and intergroup violence to temperature: the response is linear, contemporary, standard across regions, not mediated by observable economic factors or access to a conditional cash transfer program, and only barely affected by growing season temperatures. The authors do not examine the role of drought. 20 5.4 Robustness checks To validate the consistency of our main findings, we employ a binned specification linking the average precipitation in a given week and the incidence of our dependent variables of interest. We use the same fixed effects as in our baseline specification and control for temperature. Figure 10 displays the results for non-domestic violence, with panels (a) and (b) showing the effect of anomalous drought days on weekly hospital admissions for assault cases involving female and male victims, respectively. The results indicate a positive relationship between the number of anomalous drought days and the number of hospital admissions for assault cases, with a somewhat stronger pattern for male victims, consistent with our main findings. Figure 11 focuses on IPV, with panels (a) and (b) showing the relationship between weekly average precipitation and IPV hospital admissions for female and male victims, respectively. The results show that lower precipitation levels are associated with higher IPV hospital admissions for female victims, while no significant relationship is observed for male victims, as in our main specification. (a) Assault rates - female victims (b) Assault rates - male victims Figure 10: Anomalous days of drought and non-domestic violence Notes: The dependent variables are weekly hospital admissions within the Mexican public health system from 2012 to 2019, categorized according to ICD-10 codes. The regression model adjusts for average temperature. The model incorporates municipality-year and date fixed effects, state-specific time trends, and controls for municipality population sizes. Error bars represent 95% confidence intervals. 27 (a) IPV rates - female victims (b) IPV rates - male victims Figure 11: Weekly average precipitation and IPV Notes: The dependent variables are weekly hospital admissions within the Mexican public health system from 2012 to 2019, categorized according to ICD-10 codes. The regression model adjusts for average temperature. The model incorporates municipality-year and date fixed effects, state-specific time trends, and controls for municipality population sizes. Error bars represent 95% confidence intervals. 6 Conclusion We find robust evidence of an increase in domestic violence in Mexico when households face drought conditions. Our findings are robust to utilizing three data sources: hospital records, police files, and 911 calls. Our results suggest that a lower municipal state capacity level is related to a greater IPV drought impact due to a weaker deterrence effect. While accounting for 8% of national GDP, agriculture is an essential source of livelihood for millions of people across the country. It is also a highly climate-sensitive industry and a maledominated sector. Our results indicate that drought-induced disruptions intensify IPV more in agriculturally intensive areas and when it hits during the growing season. We hypothesize that adverse weather events may challenge the conventional role of men as primary earners, potentially leading to increased violence as a response mechanism. Overall, our evidence aligns more with stress-related or extractive violence in response to drought during the growing season in agricultural regions than with economic models of expressive violence. Mexico is a middle-income country, to the best of our knowledge, wealthier than every other country where the link between drought and IPV has been studied. This fact suggests that economic growth itself will not shield women from the exacerbated climate impacts they face. While there might be a human rights basis for policy intervention to hasten this convergence, in a world with limited resources, there might be more scope for government intervention in dimensions of gender inequality that are unlikely to diminish in the medium term under current economic growth trends. 28 Thus, our findings have several policy implications with external validity and relevance for a high share of the global population. First of all, disaster response programs should consider the disparate impacts that women face, incorporating IPV prevention and attention protocols into emergency response systems, especially for drought. Our findings also underscore the need to strengthen trust and use of women’s emergency resources. Better reporting and use of helplines and other resources will generate better data that can contribute to diagnosing vulnerability and target interventions. Finally, unlike other forms of violence that may be climate-sensitive, more policing is unlikely to mitigate the effects of weather extremes on violence against women since this type of violence usually does not take place in public spaces. In Mexico, more than 9 out of 10 cases of IPV against women happen at home (ENDIREH,2021). Changing social roles regarding the acceptability of domestic violence and the male role as primary earners could be a significant step in mitigating the climate-gender nexus. 29 References Abiona, O. and Koppensteiner, M. F. (2018). The impact of household shocks on domestic violence: Evidence from tanzania. IZA discussion paper. Adams, A. E.,Tolman, R. M.,Bybee, D.,Sullivan, C. M. and Kennedy, A. C. (2012). The impact of intimate partner violence on low-income women’s economic well-being: The mediating role of job stability. Violence Against Women,18 (12), 1345–1367. Agüero, J. M. (2021). Covid-19 and the rise of intimate partner violence. World development, 137, 105217. Aguilar-Gomez, S.,Arceo-Gomez, E. and De la Cruz Toledo, E. (2019). Inside the black box of child penalties: Unpaid work and household structure. Available at SSRN 3497089. Aizer, A. (2010). The gender wage gap and domestic violence. American Economic Review, 100 (4), 1847–1859. Anderberg, D.,Rainer, H.,Wadsworth, J. and Wilson, T. (2016). Unemployment and domestic violence: Theory and evidence. The Economic Journal,126 (597), 1947–1979. Angelucci, M. (2008). Love on the rocks: Domestic violence and alcohol abuse in rural mexico. The BE Journal of Economic Analysis & Policy,8(1). —and Heath, R. (2020). Women empowerment programs and intimate partner violence. In AEA Papers and Proceedings, American Economic Association 2014 Broadway, Suite 305, Nashville, TN 37203, vol. 110, pp. 610–614. Arias, P.,Bellouin, N.,Coppola, E.,Jones, R.,Krinner, G.,Marotzke, J.,Naik, V.,Palmer, M.,Plattner, G.-K.,Rogelj, J.,Rojas, M.,Sillmann, J.,Storelvmo, T.,Thorne, P.,Trewin, B.,Achuta Rao, K.,Adhikary, B.,Allan, R.,Armour, K., Bala, G.,Barimalala, R.,Berger, S.,Canadell, J.,Cassou, C.,Cherchi, A.,Collins, W.,Collins, W.,Connors, S.,Corti, S.,Cruz, F.,Dentener, F.,Dereczynski, C., Di Luca, A.,Diongue Niang, A.,Doblas-Reyes, F.,Dosio, A.,Douville, H.,Engelbrecht, F.,Eyring, V.,Fischer, E.,Forster, P.,Fox-Kemper, B.,Fuglestvedt, J., Fyfe, J.,Gillett, N.,Goldfarb, L.,Gorodetskaya, I.,Gutierrez, J.,Hamdi, R., Hawkins, E.,Hewitt, H.,Hope, P.,Islam, A.,Jones, C.,Kaufman, D.,Kopp, R., Kosaka, Y.,Kossin, J.,Krakovska, S.,Lee, J.-Y.,Li, J.,Mauritsen, T.,Maycock, T.,Meinshausen, M.,Min, S.-K.,Monteiro, P.,Ngo-Duc, T.,Otto, F.,Pinto, I., Pirani, A.,Raghavan, K.,Ranasinghe, R.,Ruane, A.,Ruiz, L.,Sallée, J.-B.,Samset, B.,Sathyendranath, S.,Seneviratne, S.,Sörensson, A.,Szopa, S.,Takayabu, I.,Tréguier, A.-M.,van den Hurk, B.,Vautard, R.,von Schuckmann, K.,Zaehle, S.,Zhang, X. and Zickfeld, K. (2021). Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change: Technical summary, Cambridge, United Kingdom and New York, NY, USA: Cambridge University Press, pp. 33–144. Arora-Jonsson, S. (2011). Virtue and vulnerability: Discourses on women, gender and climate change. Global environmental change,21 (2), 744–751. Banxico (2022). Sequía en méxico y su potencial impacto en la actividad económica. Banco Central México-Informe Trimestral Abril–Junio. Baysan, C.,Burke, M.,González, F.,Hsiang, S. and Miguel, E. (2019). Non-economic factors in violence: Evidence from organized crime, suicides and climate in mexico. Journal of Economic Behavior & Organization,168, 434–452. Bednar-Friedl, B.,Biesbroek, R.,Schmidt, D.,Alexander, P.,Børsheim, K.,Car30 nicer, J.,Georgopoulou, E.,Haasnoot, M.,Cozannet, G. L.,Lionello, P.,Lipka, O.,Möllmann, C.,Muccione, V.,Mustonen, T.,Piepenburg, D. and Whitmarsh, L. (2022). Climate Change 2022: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge, UK and New York, USA: Cambridge University Press, pp. 1817–1927. Bhalotra, S.,Kambhampati, U.,Rawlings, S. and Siddique, Z. (2021). Intimate partner violence: The influence of job opportunities for men and women. The World Bank Economic Review,35 (2), 461–479. Bhattacharya, H. (2015). Spousal violence and women’s employment in india. Feminist Economics,21 (2), 30–52. Blakeslee, D.,Chaurey, R.,Fishman, R.,Malghan, D. and Malik, S. (2021). In the heat of the moment: Economic and non-economic drivers of the weather-crime relationship. Journal of Economic Behavior & Organization,192, 832–856. Bobonis, G. J.,González-Brenes, M. and Castro, R. (2013). Public transfers and domestic violence: The roles of private information and spousal control. American Economic Journal: Economic Policy,5(1), 179–205. Boesten, J. (2012). The state and violence against women in peru: Intersecting inequalities and patriarchal rule. Social Politics,19 (3), 361–382. Bosland, P. and Walker, S. (2014). Growing chiles in new mexico. https://pubs.nmsu.edu/ _h/H230/e. Boyer, A.,Meijer, S. and Gilligan, M. (2020). Advancing gender in the environment: Exploring the triple nexus of gender inequality, state fragility, and climate vulnerability. Washington, DC: IUCN & USAID. Burke, M.,Hsiang, S. M. and Miguel, E. (2015). Climate and conflict. Annu. Rev. Econ., 7(1), 577–617. Campbell, J. C. (2002). Health consequences of intimate partner violence. The lancet,359 (9314), 1331–1336. Caruso, R.,Petrarca, I. and Ricciuti, R. (2016). Climate change, rice crops, and violence: Evidence from indonesia. Journal of Peace Research,53 (1), 66–83. Chowdhury, J. R.,Parida, Y. and Goel, P. A. (2021). Does inequality-adjusted human development reduce the impact of natural disasters? a gendered perspective. World Development, 141, 105394. CONAGUA (2022). Estadísticas del agua en méxico 2021. Ciudad de México: Secretaría del Medio Ambiente y Recursos Naturales. Obtenido de https://sina. conagua. gob. mx/publicaciones. Cook, B. I.,Mankin, J. S. and Anchukaitis, K. J. (2018). Climate change and drought: From past to future. Current Climate Change Reports,4, 164–179. Cools, S.,Flatø, M. and Kotsadam, A. (2020). Rainfall shocks and intimate partner violence in sub-saharan africa. Journal of Peace Research,57 (3), 377–390. —and Kotsadam, A. (2017). Resources and intimate partner violence in sub-saharan africa. World Development,95, 211–230. Cooper, M.,Sandler, A.,Vitellozzi, S.,Lee, Y.,Seymour, G.,Haile, B. and Azzari, C. (2021). Re-examining the effects of drought on intimate-partner violence. PLoS one,16 (7), e0254346. Croppenstedt, A.,Goldstein, M. and Rosas, N. (2013). Gender and agriculture: Inefficiencies, segregation, and low productivity traps. The World Bank Research Observer,28 (1), 31 79–109. Cullen, C. (2023). Method matters: The underreporting of intimate partner violence. The World Bank Economic Review,37 (1), 49–73. Dai, A. (2013). Increasing drought under global warming in observations and models. Nature climate change,3(1), 52–58. Dell, M.,Jones, B. F. and Olken, B. A. (2014). What do we learn from the weather? the new climate-economy literature. Journal of Economic literature,52 (3), 740–798. Díaz, J.-J. and Saldarriaga, V. (2023). A drop of love? rainfall shocks and spousal abuse: evidence from rural peru. Journal of health economics,89, 102739. Ellsberg, M.,Heise, L.,Pena, R.,Agurto, S. and Winkvist, A. (2001). Researching domestic violence against women: methodological and ethical considerations. Studies in family planning,32 (1), 1–16. ENDIREH (2021). Encuesta Nacional sobre la Dinámica de las Relaciones en los Hogares 2021 ENDIREH Principales Resultados. Inegi. ENVIPE (2022). Encuesta Nacional de Victimización y Percepción sobre Seguridad Pública. Inegi. Epstein, A.,Bendavid, E.,Nash, D.,Charlebois, E. D. and Weiser, S. D. (2020). Drought and intimate partner violence towards women in 19 countries in sub-saharan africa during 20112018: A population-based study. PLoS Medicine,17 (3), e1003064. Erten, B. and Keskin, P. (2018). For better or for worse?: Education and the prevalence of domestic violence in turkey. American Economic Journal: Applied Economics,10 (1), 64–105. —and —(2021). Female employment and intimate partner violence: evidence from syrian refugee inflows to turkey. Journal of Development Economics,150, 102607. —and —(2024). Trade-offs? the impact of wto accession on intimate partner violence in cambodia. Review of Economics and Statistics, pp. 1–12. Eswaran, M. (2018). State Complicity in the Sexual Assault of Women: The Fate of Cassandra? Tech. rep., UBC Working Paper, Vancouver School of Economics, University of British .... —and Malhotra, N. (2011). Domestic violence and women’s autonomy in developing countries: theory and evidence. Canadian Journal of Economics/Revue canadienne d’économique,44 (4), 1222–1263. FAO (2019). México se prepara para mitigar mejor las sequías. https://www.fao.org/mexico/ noticias/detail-events/es/c/1252317/. —(2020). Preassure in water resources. https://data.apps.fao.org/aquastat/?lang=en. —(2021). The impact of disasters and crises on agriculture and food security: 2021. Italy: Food and Agriculture Organization of the United Nations Rome, Italy. —(2022). Giews country brief mexico. https://www.fao.org/giews/countrybrief/country/ MEX/pdf_archive/MEX_Archive.pdf. Fatema, S. R.,Islam, M. S.,East, L. and Usher, K. (2019). Women’s health-related vulnerabilities in natural disasters: a systematic review protocol. BMJ open,9(12), e032079. Gevers, A.,Musuya, T. and Bukuluki, P. (2020). Why climate change fuels violence against women: United nations development programme. Heilmann, K.,Kahn, M. E. and Tang, C. K. (2021). The urban crime and heat gradient in high and low poverty areas. Journal of Public Economics,197, 104408. Hewage, S.,Bandara, M.,Rathnayake, D.,Research, H. K. A. and Institute, T. (2022). Local Hybrid Chili Cultivation and Dry Chili Production: Towards Spicing Up the Process. Occasional publication, Hector Kobbekaduwa Agrarian Research and Training Institute. 32 Hirvonen, K.,Gilligan, D. O.,Leight, J.,Tambet, H. and Villa, V. (2023). Do ultra-poor graduation programs build resilience against droughts? Evidence from rural Ethiopia. Intl Food Policy Res Inst. Hsiang, S. (2016). Climate econometrics. Annual Review of Resource Economics,8, 43–75. Hsiang, S. M.,Burke, M. and Miguel, E. (2013). Quantifying the influence of climate on human conflict. Science,341 (6151), 1235367. ICAMEX (2023a). Cultivo de jitomate. https://icamex.edomex.gob.mx/jitomate. —(2023b). Cultivo de papa. https://icamex.edomex.gob.mx/papa#:~:text=La%20siembra% 20se%20debe%20de,octubre%20al%2015%20de%20diciembre. ILO (2012). El empleo de las mujeres rurales en américa latina. International Labour Organization. IMCO (2024). La gestiÓn del agua no es una prioridad en mÉxico. https://imco.org.mx/ la-gestion-del-agua-no-es-una-prioridad-en-mexico/. INEGI (2010). Censo de población y vivienda 2010. Instituto Nacional de Estadística y Geografía. INIFAP (2012). Tomate. http://www.inifapcirne.gob.mx/Biblioteca/Paquetes2012/149. pdf. —(2013a). Frijol. http://www.inifapcirne.gob.mx/Biblioteca/Paquetes2012/137.pdf. —(2013b). TecnologÍa para la producciÓn de avena forrajera en el altiplano de tamaulipas. http: //www.inifapcirne.gob.mx/Biblioteca/Publicaciones/956.pdf. —(2015a). Cultive pasto rye grass para la alimentaciÓn del ganado en la Época invernal en el norte y centro de tamaulipas. http://www.inifapcirne.gob.mx/Eventos/2015/Siembra+Rye+ Grass.pdf. —(2015b). Paquete tecnológico para maíz forrajero. https://www.producechihuahua.org/paqs/ PT-0007Maiz2.pdf. —(2024a). Biblioteca digital del instituto nacional de investigaciones forestales, agrícolas y pecuarias. https://vun.inifap.gob.mx/BibliotecaWeb/_Content. —(2024b). Calabacita. http://oiapes.sagarhpa.sonora.gob.mx/paq-tec/paq-calabacita. pdf. —(2024c). Sandía. http://oiapes.sagarhpa.sonora.gob.mx/paq-tec/paq-sandia.pdf. —(2024d). Sorgo forrajero. http://oiapes.sagarhpa.sonora.gob.mx/paq-tec/ paq-sorgo-forrajero.pdf. Jia, R. (2014). Weather shocks, sweet potatoes and peasant revolts in historical china. The Economic Journal,124 (575), 92–118. Keerthiratne, S. and Tol, R. S. (2018). Impact of natural disasters on income inequality in sri lanka. World Development,105, 217–230. Kuzma, S.,Saccoccia, L. and Chertock, M. (2023). 25 countries, housing one-quarter of the population, face extremely high water stress. World Resources Institute. https://www. wri. org. Leslie, E. and Wilson, R. (2020). Sheltering in place and domestic violence: Evidence from calls for service during covid-19. Journal of public economics,189, 104241. Logie, C. H.,Daniel, C.,Ahmed, U. and Lash, R. (2017). ‘life under the tent is not safe, especially for young women’: understanding intersectional violence among internally displaced youth in leogane, haiti. Global health action,10 (sup2), 1270816. Miedzian, M. (2002). Boys will be boys: Breaking the link between masculinity and violence. Lantern Books. Miller, A. R.,Segal, C. and Spencer, M. K. (2024). Effects of the covid-19 pandemic on domestic violence in los angeles. Economica,91 (361), 163–187. 33 Munala, L.,Allen, E. M.,Frederick, A. J. and Ng˜ unjiri, A. (2023a). Climate change, extreme weather, and intimate partner violence in east african agrarian-based economies. International journal of environmental research and public health,20 (23), 7124. —,—,—and Ng˜ unjiri, A. (2023b). Climate change, extreme weather, and intimate partner violence in east african Agrarian-Based economies. Int J Environ Res Public Health,20 (23). Neumayer, E. and Plümper, T. (2007). The gendered nature of natural disasters: The impact of catastrophic events on the gender gap in life expectancy, 1981–2002. Annals of the association of American Geographers,97 (3), 551–566. Nguyen, H. T. (2019). Gendered vulnerabilities in times of natural disasters: male-to-female violence in the philippines in the aftermath of super typhoon haiyan. Violence Against Women, 25 (4), 421–440. Oduro, A. D.,Deere, C. D. and Catanzarite, Z. B. (2015). Women’s wealth and intimate partner violence: Insights from ecuador and ghana. Feminist Economics,21 (2), 1–29. ONU (). Jornaleras agrícolas mexicanas luchan por sus derechos y la protección del medio ambiente. ONU Mujeres. América Latina y El Caribe. Parmar, P. K.,Jin, R. O.,Walsh, M. and Scott, J. (2019). Mortality in rohingya refugee camps in bangladesh: historical, social, and political context. Sexual and reproductive health matters,27 (2), 39–49. Quisumbing, A. R. and Doss, C. R. (2021). Gender in agriculture and food systems. Handbook of agricultural economics,5, 4481–4549. Registro Agrario Nacional (2019). Estadística con perspectiva de género, tabulados. http://www.ran.gob.mx/ran/index.php/sistemas-de-consulta/estadistica-agraria/ estadistica-con-perspectiva-de-genero. Rincón-Avalos, P.,Khouakhi, A.,Mendoza-Cano, O.,Cruz, J. L.-D. l. and ParedesBonilla, K. M. (2022). Evaluation of satellite precipitation products over mexico using google earth engine. Journal of Hydroinformatics,24 (4), 711–729. Roncoli, C.,Ingram, K. and Kirshen, P. (2001). The costs and risks of coping with drought: livelihood impacts and farmers1responses in burkina faso. Climate research,19 (2), 119–132. Sabia, J. J.,Dills, A. K. and DeSimone, J. (2013). Sexual violence against women and labor market outcomes. American Economic Review,103 (3), 274–278. Sekhri, S. and Storeygard, A. (2014). Dowry deaths: Response to weather variability in india. Journal of development economics,111, 212–223. SIAP (2024). Anuario estadístico de la producción agrícola. Servicio de Información Agroalimentaria y Pesquera. Soifer, H. D. (2012). Measuring state capacity in contemporary latin america. Revista de Ciencia Política,32 (3), 585–598. S&P (2023). Más estados mexicanos podrían verse afectados por estrés hídrico en 2050. https://www.spglobal.com/ratings/es/pdf-articles/2023/ 2023-04-04-mas-estados-mexicanos-podrian-verse-afectados-por-estres-hidrico-en-2050. Sviatschi, M. M. and Trako, I. (2024). Gender violence, enforcement, and human capital: Evidence from women’s justice centers in peru. Journal of Development Economics, p. 103262. Terrazas-Carrillo, E. C. and McWhirter, P. T. (2015). Employment status and intimate partner violence among mexican women. Journal of interpersonal violence,30 (7), 1128–1152. Thornton, M.,Shrestha, R.,Wei, Y.,Thornton, P.,Kao, S.-C. and Wilson, B. (2022). Daymet: Daily surface weather data on a 1-km grid for north america, version 4 r1. 34 Thurston, A. M.,Stöckl, H. and Ranganathan, M. (2021). Natural hazards, disasters and violence against women and girls: a global mixed-methods systematic review. BMJ global health, 6(4), e004377. Udry, C. (1996). Gender, agricultural production, and the theory of the household. Journal of political Economy,104 (5), 1010–1046. UN Women (2022). Gender, climate and conflict analysis in somalia and assessment of opportunities for climate smart agriculture and livelihood opportunities for crisis-affected and at-risk women in somalia. USDA (2024). Crop calendars for mexico, central america and the caribbean. United States Department of Agriculture,https://ipad.fas.usda.gov/rssiws/al/crop_calendar/ca.aspx. Van Aelst, K. and Holvoet, N. (2016). Intersections of gender and marital status in accessing climate change adaptation: Evidence from rural tanzania. World Development,79, 40–50. van Daalen, K. R.,Kallesøe, S. S.,Davey, F.,Dada, S.,Jung, L.,Singh, L.,Issa, R., Emilian, C. A.,Kuhn, I.,Keygnaert, I. et al. (2022). Extreme events and gender-based violence: a mixed-methods systematic review. The Lancet Planetary Health,6(6), e504–e523. Warr, P. and Aung, L. L. (2019). Poverty and inequality impact of a natural disaster: Myanmar’s 2008 cyclone nargis. World Development,122, 446–461. World Bank (2022a). Agriculture, forestry, and fishing, value added (% of gdp). Accessed: 202405-19. World Bank (2022b). Violence against women and girls – what the data tell us. https: //genderdata.worldbank.org/en/data-stories/overview-of-gender-based-violence. World Bank (2023). Water. https://www.bancomundial.org/es/topic/water/overview. Ylipaa, J.,Gabrielsson, S. and Jerneck, A. (2019). Climate change adaptation and gender inequality: insights from rural vietnam. Sustainability,11 (10), 2805. Zaveri, E. D.,Damania, R. and Engle, N. (2023). Droughts and deficits. World Bank, Washington, DC. 35 A Online Appendix Figure A.1: Level of economic development of countries with quantitative studies of the causal impact of droughts on IPV Notes: The published referenced works are include Sekhri and Storeygard (2014) (India), Cools and Kotsadam (2017) (Sub-Saharan Africa) and Díaz and Saldarriaga (2023) (Peru). Working papers referenced include Abiona and Koppensteiner (2018) (Tanzania) and Hirvonen et al. (2023) (Ethiopia). 36