Aided Unequally: A New Method to Measure Local Disaster Response in the Global South
About this Session
Thu. 11.04.'24 15:05
In light of the increasing frequency of extreme weather events driven by the ongoing climate crisis, Global South nations are poised to bear the brunt of these impacts. This global disparity in climate-induced vulnerabilities is mirrored at the local level. Furthermore, local governments’ disaster responses are accused of often exacerbating economic and political vulnerabilities, ultimately perpetuating inequality. Unfortunately, the scarcity of data on disaster responses in lower and middle-income countries has hindered the identification of these critical patterns. This paper introduces an innovative methodology that leverages remote sensing and search data to unveil the responsiveness of Global South governments to local disasters such as drought, flooding, and heat waves. The proposed method integrates multiple data sources, including remote sensing data of extreme weather events and information on infrastructure quality, utilizing machine learning techniques. Additionally, the measure incorporates internet search data to assess resource concerns, where concern is expected to have an inverse correlation with the extent of disaster response by local authorities. The method produces a global grid at the neighborhood level, wherein pre-disaster vulnerability is computed with image processing techniques that evaluate the condition of infrastructure. Subsequently, event analysis is conducted to incorporate the timing and intensity of extreme weather occurrences. Outcome measures are then computed, capturing resource concerns based on internet search data from Google Trends, employing a two-stage regression process with the pre-disaster vulnerability metric. The output is a granular, neighborhood-level perspective on disaster response efforts undertaken by local governments. The analysis reveals extreme inequality in disaster response within municipalities across the Global South, and structural factors such as neighborhood income have some explanatory power in understanding the distribution of outcomes. However, I also show how electoral considerations play a pivotal role in explaining disaster responses, particularly in scenarios where reelection incentives are significant and democracy is of minimal quality. The paper also acknowledges certain limitations, notably that the measure’s validity is contingent on factors such as local internet penetration and the availability of subnational trend analysis through Google Search. The methodology presented in this paper could significantly enhance social scientists’ ability to understand inequality in the face of climate change in lower-income countries.