Near-real-time welfare and livelihood impacts of an active war: Evidence from Ethiopia.

• This paper provides fresh evidence on the ex durante impacts of an armed conflict on households' food security in Ethiopia. • We also assess potential mechanisms and hence evaluate impacts on livelihood activities and access to food markets. • Seven months into the conflict, we find that the war increased the probability of moderate to severe food insecurity by 37 percentage points. • We show that exposure to one additional battle leads to 1 percentage point increase in the probability of moderate to severe food insecurity. • The conflict has significantly reduced access to food through supply chain disruptions and by curtailing non-farm livelihood activities. • Non-farm and wage related activities were affected the most whereas farming activities were relatively more resilient. Ethiopia recently experienced a large-scale war that lasted for more than two years. Using unique High-Frequency Phone Survey (HFPS) data, which span several months before and after the outbreak of the war, this paper provides evidence on the immediate impacts of the conflict on households' food security. We also assess potential mechanisms and evaluate impacts on proximate outcomes, including on livelihood activities and access to food markets. We use difference-in-differences and two-way fixed effects estimation to compare trends across affected and unaffected regions (households) and before and after the outbreak of the war. Seven months into the conflict, we find that the war was associated with a 37 percentage points increase in the probability of moderate to severe food insecurity. Using the Armed Conflict Location and Event Data (ACLED), we show that exposure to an additional battle leads to a 1 percentage point increase in the probability of moderate or severe food insecurity. The conflict was associated with significant reduction in access to food through supply chain disruptions and by curtailing non-farm livelihood activities. Non-farm and wage related activities were affected the most, whereas farming activities were relatively more resilient. Our estimates, which likely underestimate the true average effects on the population, constitute novel evidence on the near-real-time impacts of large-scale conflict. Our work highlights the potential of HFPS to monitor active and large-scale conflicts, especially in contexts where conventional data sources are not immediately available. [ABSTRACT FROM AUTHOR]