In this study i have put an effort to find the direct and indirect impact of corruption on GDP growth through government size transmission channel. The sample of this study consists of three south Asian economies i.e. those of India, Pakistan and Sri Lanka from 1990 to 2013. I have used cross sectional as well as panel data techniques. My results show that an increase in corruption leads to an increase in government size which, in turn, decreases GDP growth. This study suggests that, in the presence of high corruption level, a larger government size is not favourable for India, Pakistan, or Sri Lanka because it hinders GDP growth. Another aim of this study is to find the long run association (Causality) between corruption and economic growth. To attempt this objective, i have same sample size from 1984 to 2013. This study has conformed that there is short term two way causality running from corruption to GDP growth to corruption. The results of study suggests two policies to the governments of India, Pakistan and Sri Lanka; they can either concentrate more on economic growth which ultimately will reduce corruption or combat corruption rather than focusing on growth