This article investigates the empirical linkages between crop and livelihood diversification strategies, extreme weather events, and household welfare using a unique dataset that integrates harmonized, national representative household surveys and geo-referenced climatic information collected in Malawi, Niger and Zambia. In doing so, the paper addresses the potential endogeneity arising from the selection bias and the heterogeneity of the effect across the quantiles of the income distribution. Results show that exposure to extreme rainfall events is positively associated with either crop or livelihood diversification in all the countries analyzed, suggesting that climate-related shocks are key push factors for diversification. Moreover, results show that the effect of diversification on household income are varied across countries and diversification strategies. However, most of this heterogeneity disappears when estimating the quantile treatment effects. In particular, the results show that the impact of both crop and income diversification on households' welfare is generally higher for the poorest (people located at the bottom tail of the distribution) while it decreases, and in some cases turns to be negative, moving toward the upper end of the income distribution in all three countries. The findings, therefore, highlight the pro-poor impact of diversification strategies in multiple rural African contexts, as well as the need to tailor diversification interventions toward specific socio-economic segments of the rural population. (C) 2019 Elsevier Ltd. All rights reserved.