The marginal effect of school leaving age on demographic events. A contribution to the discussion on causality