Seems to me you need to isolate roads from other factors to establish cause and effect. It also seems to me that's just about impossible to do in practice. Therefore, in most cases, one can't know if an increase or decrease in local stability is directly attributable to road or some other factor.