Investigations into relationships between climate, crop production and migration in Malawi

Abstract

Across the world, climate change is affecting migration patterns through changes to natural resources and people’s livelihoods. In Malawi, the developing agrarian economy is particularly vulnerable to climate change through changing crop production and food security, which may ultimately drive people to migrate in search of alternative income sources. Critical examination of literature yielded a conceptual model for ‘climate migration’. This model was then applied to the case study of Malawi and used to inform a set of regression models for a number of system compartments. Models utilised data from Malawi’s periodic Integrated Household Surveys (IHS) to examine associations between various predictor variables and crop production, food security and migration outcomes. By isolating the climate change signal, a negative relationship was also found between crop production and maximum temperature anomaly. Significant relationships were identified between maximum temperature and a small decrease in odds of prolonged hunger. However, studying the years preceding a household survey did not identify any relationship with crop production during survey years, nor did a multi-level crop production model identify any association with a range of climate variables. Sociodemographic predictors were found to be highly significant, including, gender, household wealth and farm size. Modest relationships were identified between migration status and food security metrics. In general, non-migrants were found to experiencing higher food insecurity than migrants though high district variation was uncovered. Prolonged hunger was associated with increased migration odds. These modest findings also challenge popular narratives which may over-assert the current known role of climate change as a driver of migration. Further research is recommended particularly focusing on a range of rainfall data, extreme events and a wider range of agro-climatic variables. More detailed migration data is also called for to better understand these relationships. Variables should be explored as both drivers of migration as well as impeding migration through a ‘trapping effect’.