ABSTRACT
Simultaneous exposure to natural calamities and conflict shocks is a phenomenon that has been largely understudied. The interplay between natural disasters and conflict shocks can have adverse effects extending beyond the current family members to children in utero. The current paper tries to fill this gap by investigating the impact of floods on pregnancy and birth outcomes across conflict-affected and unaffected districts in Pakistan. Using mother fixed effects strategy, the results suggest that in-utero exposure to violence during flood increases the probability of small birth size by 4.7 percentage points. Moreover, simultaneous exposure to flood and violence increases the probability of miscarriages and stillbirths by 6 and 1.9 percentage points, respectively. Significant heterogeneities are found across income groups and education levels.
Disclosure statement
No potential conflict of interest was reported by the author.
Notes
1. All other insurgent groups came under the umbrella of Tehreek-e-Taliban Pakistan (TTP). These included Punjabi Taliban as well as insurgents from other ethnic groups.
2. Children who are born in non-institutional settings such as homes are less likely to be weighted at birth. On the other hand, those who are born in formal health facilities such as hospitals are always weighted. Therefore, the probability of weighting a child at birth is a reasonable proxy for whether or not the parents used a health facility for delivery.
3. The indicators for miscarriage and stillbirth were constructed against the live birth and, hence, these were analyzed separately.
4. The word Agency is used in a geographical sense and is equivalent to a district in the settled areas. There are seven Agencies in the Federally Administered Tribal Region (FATA) in Pakistan; Khyber, Bajaur, Mohmand, Kurram, Orakzai, South Waziristan, North Waziristan.
5. In case a district is affected in both years, it is also given the value 1.
6. 70 and 23 districts were affected by the flood in 2010 and 2011 respectively. There were eight districts that experienced floods in both years.
7. I have also empirically examined the link between birthweight and birth size for the sample at hand and found significant correlation between the two measures.
8. In order to split the districts in conflict-affected and unaffected regions, I added all the attacks during the relevant period of 2007–2012 and used the 75th percentile (or 19 terrorist attacks) as the threshold. All districts that have experienced more than 19 attacks were considered violent districts. The rest were declared as non-violent districts.