The available evidence on the effects of aid on growth is notoriously mixed. We use a novel empirical methodology, a heterogeneous panel vector-autoregression model identified through factor analysis, to study the dynamic response of exports, imports, and per capita GDP growth to a global aid shock (the common component of individual country aid-to-GDP ratios). We find that the estimated cumulative resposive of exports and per capita GDP growth to a global aid shock are strongly positively correlated, and both responses are inversely related to exchange rate overvaluation measures. We interpret this evidence as consistent with the Dutch disease hypothesis. However, we also find that, in countries with less overvalued real exchange rates, exports and per capita GDP growth respond positively to a global aid shock. This evidence suggests that preventing exchange rate overvaluations may allow aid-receiving countries to avoid the Dutch disease.