This study proposes a political interference hypothesis to explain how political considerations depress the performance of government banks. We define political interference as a situation in which government bank executives are replaced within 12. months after the country's major elections (presidential or parliamentary elections). We classify political and non-political government banks as those that experience or do not experience political interference, respectively. The hypothesis firstly suggests that once government banks undertake political interference, their financial performance deteriorates. That is, political banks display the worst performance, followed by non-political banks and private banks have the best performance. Next, we posit that the impact of political interference is greater in developing countries than in developed countries. Finally, we hypothesize that the underperformance of government banks will be reduced if we remove political interference. By employing bank data from 65 countries from the period of 2003-2007, our hypothesis effectively explains why government banks in developed countries escape relatively unscathed, while those in developing countries suffer significantly.