Can we use seasonally adjusted variables in dynamic factor models?