On the tail problem in cosmology