On Feenberg's Perturbation Formula