This study focuses on one of the important assumption of DEA, that all DMU is homogenous. The DMUs which are different in the pattern of other DMUs is considered to be the outlier. And this outlier affects the efficiency score as well as determinants. So detection or removal is necessary. Here we used purely nonparametric approach to detect outlier which is introduced by Banker and Gifford (1988), this method screen out all outlets and gives us a more reliable efficiency score. This reliable efficiency score gives us other econometric processing with accuracy or validity and minimized measurement error. This study incorporates the comparison of two standard models of DEA named as CCR and BCC how the difference in results takes place and which model is appropriately regarding this study and helpful to investigate their determinants. This study focuses on the ranking of efficient DMUs as well as sources of these inefficiencies, whether it is inside or outside the system. Along with that this study measures the government public sector investment inefficiencies and their determinants along with that explain the strong theoretical background of government, public sector investment lead growth, theoretical use of macroeconomic Barro model that identifies the concept of inefficiencies in public financing. This study discusses the while the inefficiencies in the government, public financing system is due to political reasons, governmental management flaws or due to some economic causes. This study focuses on SAARC countries because WDI reported negative governmental management index for these countries and this area required a meaningful research. Supervisor:- Dr Saud Ahmad Khan