I would apply a state budget expenditure structure in dynamic series as a proxy to evaluate the general government policy. I would use a data on comparable countries (peer actors) to get more insight. You still need to remember that there could be a qualitative and not quantitative change in government policies that is not well described by aggregate data.