This paper investigates compositional effects on public-debt dynamics. Our contribution is twofold. First, we improve the methodology by taking explicit account of public-debt dynamics, eliminating a bias present in previous studies. Second, we estimate the evolution of debt dynamics induced by changes in the primary expenditure and revenue ratios – looking at all fiscal periods – not only consolidations as in the previous literature. Whilst we confirm that expenditure cuts are more important for debt reductions than are revenue increases, our findings contrast with the literature in that we obtain persistent debt-augmenting effects of tax reductions.