Conference Agenda

There is still uncertainty on which areas will be more exposed to climate-related conflict risks in the long-term, and what is the magnitude of these risks. We contribute to this debate by proposing a forecasting analysis of armed conflicts in Africa along different SSPs and RCPs scenarios to account for alternative economic and climate pathways and assess the role of geographical spillovers in the diffusion of violence. We propose a model to forecast conflicts on a 1° grid of 2,653 cells covering the entire African continent up to 2050. Among different forecasting methods, we use that able to better predict conflict events occurred in the past (1990-2020). This occurs when dynamic recursive is preferred to simpler plug-in prediction methods and spatial relations are accounted for. We use our findings to reveal the exponential effects on conflict magnitude played by geographical diffusion processes when climate mitigation ambitions are low and the development pattern is based on unequal economic growth processes.

Climate-related migration depends on local contexts and on the interaction of climatic events with other migration drivers. The climate-migration association is highly context dependent and mediated through various channels. So far, we have a limited understanding of these contextual effects and mechanisms at play. Yet, such information is crucial to reduce welfare losses from migration in a changing climate. This paper contributes to closing these research gaps by addressing the questions: i) Who are the rural climate migrants in low- and middle-income countries? and ii) Why do they move?. We combine ERA5 weather data with socio-economic panel data from four low- and middle-income countries (i.e., South Africa, Tanzania, Malawi and China) and estimate a series of two-way fixed effects linear regression models. We apply the Roy-Borjas framework to explain the self-selection of climate migrants based on skills, where changing income distributions and levels are the mechanisms at play. Our preliminary results indicate that, on aggregate, positive temperatures anomalies drive migration by reducing financial constraints of low income households. Negative precipitation anomalies, on the contrary, serve as a negative income shock. While education generally drives rural-urban moves, we find an indication that migrants from households affected by weather anomalies are likely to be less educated.

Sub-Saharan Africa has an immense potential for city growth, which is increasingly seen as an outcome of rural-to-urban migration. Recent case studies show that the climate and subsequent weather realizations can act as a push factor causing people to move from rural to urban areas. In this paper, we uncover how climate-driven agricultural productivity shocks in rural areas affect economic activity in nearby cities and towns in Sub-Saharan Africa. We assemble a rich set of high-resolution geospatial information to identify the size and location of cities, their rural hinterlands, and changes in water availability for agricultural production. Our results show that the effect of changes in agricultural productivity on cities' growth is non-linear. Cities grow when rural areas experience droughts, shrink in times with weather conditions favourable for agricultural production, and again start growing with wetter than optimal conditions. Additionally, we show that this effect is heterogeneous conditional on the initial size of cities. When the surrounding rural area faces dry conditions, primary cities and cities with a population size above 1,000,000 people attract the most human activity. While wet conditions lead small cities to shrink, large cities grow. Our findings have important implications for urban planning, especially for evaluating future climate change effects on rural-to-urban migration. Understanding when agricultural productivity shocks due to climate variability and change push economic activity to cities is key to uncover where population and economic growth will be geographically allocated.

We use data from 2003-2019 for 2.47 million test-takers of a national high stakes university entrance exam in Ethiopia to study the impacts of temperature on learning outcomes. We find that high temperatures in the school year leading up to the exam reduce test scores, controlling for temperatures when the exam is taken. Our results suggest that the scores of female students are less impacted by higher temperatures as compared to their male counterparts. Additionally, we find that the scores of students from schools located in hotter regions are less impacted by higher temperatures as compared to their counterparts from cooler regions. Our evidence suggests the adverse effects of temperature are driven by impacts from within-classroom temperatures, rather than from indirect impacts on agriculture.