On the universality of the leading, power corrections in QCD