This paper is the first attempt to directly explore the long-run nonlinear relationship between the shadow economy and level of development. Using a dataset of 158 countries over the period from 1996 to 2015, our results reveal a robust U-shaped relationship between the shadow economy size and GDP per capita. Our results imply that the shadow economy tends to increase when economic development surpasses a given threshold or at least does not disappear. Our findings suggest that special attention should be given to the country’s level of development when designing policies to tackle issues related to the shadow economy.