The present study is, in particular, an attempt to test the relationship between tax level and political stability by using some economic control variables and to see the relationship among government effectiveness, corruption, and GDP. For the purpose, we used the Vector Autoregression (VAR) approach in the panel framework, using a country-level panel data from 59 countries for the period 2002 to 2008. The salient features of this model are: (a) simplicity is based on a limited number of variables (five) are categorical or continuous and not dependent on complex interactions or nonlinear effects. (b) accuracy: a low level of errors, the model achieves a high percentage of accuracy in distinguishing countries with inclination to political instability, compared to countries with political stability, (c) generality: the model allows to distinguish types of political instability, both resulting from acts of violence and failure of democracies to show, and (d) novelty: the model incorporates a tool that helps evaluate and exclude many variables used by the conventional literature. This approach is mainly based on the recognition of state structures and the relations between elites and parties.