Predicting Impacts of Climate Change on Conflict in Kenya Using National Surveys

• In Event: Climate Change and Conflict

Abstract

To understand the mechanisms linking environmental change and conflict in Kenya, we adopt a novel research design and make predictions across two national surveys in Kenya. From the results of a 2014 survey (reported in earlier articles), we predict the rates of violence victimization in an identical 2018 sample drawn from the same 175 enumeration areas and with the same questionnaire. Extrapolating from surveys into future possibilities appears viable if our previous estimates of the role that drought, food availability, resources management institutions, and mobility have on conflict also hold for victimization reported in 2018. If Kenyans who reported moving due to drought in 2014 were twice as likely to be victims of violence, we expect that those who have relocated in the 2018 sample report victimization twice as often as others in the sample. Such a validation is not inevitable, however. In more sophisticated comparisons of multivariate regression model results, we investigate the changing relationship between covariates and conflict over the four-year interval between surveys. Among country-level institutional factors, including the implementation of a new constitution with increased autonomy for county governments, we study the impacts of the changing environmental conditions that respondents have experienced. Such a validation of independent cross-sectional models provide insight into the stability of predictor estimates and enable us to apply the models to future scenarios of Kenya with a better understanding of their uncertainty.

Authors

• Andrew Martin Linke, University of Utah

• Frank D.W. Witmer, University of Alaska Anchorage

• John V. O'Loughlin, University of Colorado