Modeling of budget time series

Democratic (more or less) countries' annual budgets show specific fluctuations beside the general growing tendency. Autocorrelation functions show weakly the periodicity of elections, the local changes can be well approximated by a random process. Our goal is to specify the underlying random process.

In general, behind the random process concept there are several assumptions: the decision makers got a lot of information from several independent source, they bring decisions more or less independently. The decision maker has to solve problems, react to public claims. Also "keeping the status quo" is a very robust feature of politicians (if something is working then do not change it). The decision maker processes these informations and transforms them according to his/her own values into allocated budget to the "problem". It is important to note that this concept can be valid to democratic systems where the decision makers are (thought to be) driven by public claims. In a dictatorship, the direction of drive is inverse, the dictator defines the rules of living and spendings.

People working on this project

László Zalányi, Péter Érdi