We propose a cross-nested logit (CNL) approach to investigate how individuals adjust their migration decisions in response to changes in the global landscape. In contrast to the widely used logit model, the CNL enables more intricate substitution patterns among destinations. Leveraging migration aspiration data from India, we demonstrate that the CNL approach outperforms competing approaches in terms of model fit and predictive accuracy. It reveals greater heterogeneity in responses to shocks, and uncovers intricate and intuitive substitution patterns. Our analysis underscores the limited substitutability between the home and foreign alternatives, as well as within specific subgroups of destination countries.