We analyze the secular decline in gross interstate migration in the United States from 1991 to 2011. We argue that migration fell because of a decline in the geographic specificity of returns to occupations, together with an increase in workers' ability to learn about other locations before moving. Micro data on earnings and occupations across space provide evidence for lower geographic specificity. Other explanations do not fit the data. A calibrated model formalizes the geographic specificity and information mechanisms and is consistent with cross‐sectional and time‐series evidence. Our mechanisms can explain at least half of the decline in migration.