Weather the Storm: Analysing the Role of Shock-Responsive Social Protection in Building Flood Resilience using High-Resolution Data

Climate change is increasing the intensity and frequency of floods - and populations living in low-and-middleincome countries (LMICs) will be most affected.

Social protection programmes, such as cash transfers, have the potential to shield societies from the negative effects of climatic shocks - but they require evidence around climate vulnerability to target the right populations.

My first research question will use poverty predictions from machine learning (ML) algorithms, flood data derived from satellite imagery, and spatial econometrics to estimate the impact of floods on poverty across 89 LMICs at a high spatial resolution.

I will then focus on the case study of Pakistan, one of the most climate vulnerable countries in the world, with frequent floods that disproportionately affect the poor.

In 2008, the Government of Pakistan launched its flagship national social protection programme, Benazir Income Support Programme (BISP), providing quarterly unconditional cash transfers to poor households.

I will investigate whether cash transfers under BISP make households more resilient to flooding - a finding that would be generalisable to other LMICs with social protection systems and climatic shocks. 'Big data', such as remotely sensed or coming from social media, are increasingly being used to target cash transfers in response to natural disasters.

My third research question aims to explore whether satellite imagery of flooding and ML methods could provide insights for the targeting of BISP, including anticipatory action that could trigger cash disbursements based on predictions of the socio-economic impact of future flooding.