This paper adds to aid volatility literature in three ways: First it tests the validity of the aid volatility and growth relationship from various aspects: across different time horizons, by sources of aid, and by aid volatility interactions with country characteristics. Second, it investigates the relationship by the level of aid absorption and spending. Third, when examining the relationship between International Development Association aid volatility and growth, it isolates International Development Association aid volatility due to the recipient country's performance from that due to other sources. The findings suggest that, in the long run, on average, aid volatility is negatively correlated with real economic growth. But the relationship is not even. It is stronger for Sub-Saharan African countries than for other regions and it is not present in middle-income countries or countries with strong institutions. For economies where aid is fully absorbed, aid volatility matters for long-run growth; economies with full aid spending also bear a negative impact of aid volatility on long-run growth. Where aid is not fully absorbed, or where it is not fully spent, the aid volatility relationship is not significant. Looking at International Development Association aid separately, the volatility arising from the recipient country's International Development Association performance does not have a causal relationship with growth. In policy terms, the results suggest that low- income countries with weak institutions, especially in Sub-Saharan Africa, could benefit from reduced aid volatility or from being better prepared for the volatility that is there.