Page 76 - IJPS-11-4

International Journal of
Population Studies Droughts and intimate partner violence

exposure to the earthquake’s devastation increased the Second, only a few studies have analyzed the effect of
likelihood of both physical and sexual IPV. In India, Sekhri drought on IPV within the Indian context. Rai et al. (2021)
and Storeygard (2014) found that deviations in annual examined the effects of drought and cyclones on IPV
rainfall by one standard deviation from the local average among Indian women and found a positive association
led to a 7.8% increase in dowry deaths, whereas wet shocks between exposure to cyclones and all forms of IPV, but no
had no noticeable effect. Sahni and Sinha (2018) reported such relationship was observed between drought exposure
that communities experiencing higher-than-average and IPV. However, the data collection period in their study
annual precipitation were less likely to report IPV. Rao coincided with the drought period; thus, the result may not
(2020) examined IPV in Andhra Pradesh, Tamil Nadu, fully reflect the effects of drought. Our study overcomes
Karnataka, and Kerala before and after the 2004 Indian this limitation. Third, we used subsample analysis to
Ocean tsunami and found a correlation between being understand the heterogeneous effects of drought on rural
affected by the tsunami and IPV. and urban sectors. Fourth, we benefited from conducting a
DHS just before the NEM drought and a second DHS after
1.1.3. Drought and IPV the drought. The question about IPV refers to incidences
Research on the relationship between drought and IPV within the last 12 months, aligning with the data collection
has offered mixed results. Epstein et al. (2020) analyzed time of the second survey and the NEM drought. Unlike
cross-sectional surveys of 83,990 women from 19 sub- nationwide droughts, exposure to the NEM drought was
Saharan African countries between 2011 and 2018, localized; thus, we did not account for spatial correlation.
finding that several droughts were associated with a In addition, we used drought as a binary variable rather
higher risk of reporting PV, SV, or emotional violence than employing actual rainfall deficiency data across
(EV). However, they did not find a connection between districts. Districts are smaller units compared to states
drought and partners’ controlling behavior. Cools et al. or countries, and rainfall deficiency is more uniform in
(2020) used DHSs from 17 sub-Saharan countries between smaller geographical areas. We also used district-level
2003 and 2013 to study the effects of rainfall shocks on fixed effects to control for unobserved heterogeneities and
IPV, concluding that there was no significant association. clustered the standard errors at the district level.
Unlike Epstein et al. (2020), Cools et al. (2020) controlled Moreover, this study improves on previous studies by
for spatial correlations. Cooper et al. (2021) attempted to incorporating an extensive set of control variables, which
reconcile these contradictory findings by combining the allows for more precise estimates of the drought’s effect.
methodologies of both studies. Although their sample Thus, the objective of this study is to isolate the impact of
size was larger, including DHSs from sub-Saharan Africa, drought on different forms of IPV.
Latin America, and Asia, they found insignificant relation
between drought and IPV on any of the three continents. South India experienced an extreme NEM drought
However, Rai et al. (2021) examined the effect of cyclones from 2016 to 2018 (Mishra et al., 2021). This drought
and drought on IPV among Indian women and found no was the worst in the past 150 years, with a precipitation
statistically significant association between drought and deficit of 45%. As a result, four of Chennai’s reservoirs
IPV, although a positive correlation was noted between dried up, leading to the declaration of a “Day Zero” in the
drought and PV. summer of 2019 (Jain, 2021). As the drought occurred
and affected lives before the NFHS-5 data collection and
1.2. Research purpose the recall period for IPV questions coincides with the
After reviewing the existing literature on IPV, we realized drought period, this situation creates a natural experiment.
that the debate surrounding the association between Therefore, the data should capture the effect of drought on
droughts and IPV remains open, with mixed evidence IPV. Although the data are not longitudinal, which makes
at best. Several attempts have been made to establish a it technically difficult to establish a causal relationship
statistically significant association between rainfall shocks between drought and IPV, one thing is certain: the relation
and IPV using large sample sizes (Cools et al., 2020; is not bidirectional; incidences of IPV cannot cause
Cooper et al., 2021; Epstein et al., 2020) while controlling drought.
for known individual- and household-level covariates 2. Data and methods
of IPV (Cooper et al., 2021; Epstein et al., 2020), as well
as spatial correlations (Cools et al., 2020; Cooper et al., DHSs are nationally representative household surveys
2021). We employed a slightly different approach. First, we that collect population, health, and nutrition data across
focused on data only from Indian states to avoid cultural regions, including sub-Saharan Africa, North Africa/West
and contextual differences that might affect the results. Asia/Europe, Central Asia, South and Southeast Asia,

Volume 11 Issue 4 (2025) 70 https://doi.org/10.36922/ijps.3065