Abstract: Smart grids (SGs) have a central role in the development of the global power sector. Cost-benefit analyses and environmental impact assessments are used to support policy on the deployment of SG systems and technologies. However, the conflicting and widely varying estimates of costs, benefits, greenhouse gas (GHG) emission reduction, and energy savings in literature leave policy makers struggling with how to advise regarding SG deployment. Identifying the causes for the wide variation of individual estimates in the literature is crucial if evaluations are to be used in decision-making. This paper (i) summarizes and compares the methodologies used for economic and environmental evaluation of SGs (ii) identifies the sources of variation in estimates across studies, and (iii) point to gap in research on economic and environmental analyses of SG systems. Seventeen studies (nine articles and eight reports published between 2000 and 2015) addressing the economic costs versus benefits, energy efficiency, and GHG emissions of SGs were systematically searched, located, selected, and reviewed. Their methods and data were subsequently extracted and analysed. The results show that no standardized method currently exists for assessing the economic and environmental impacts of SG systems. The costs varied between 0.03 and 1143 M/yr, while the benefits ranged from 0.04 to 804 M/yr, suggesting that SG systems do not result in cost savings The primary energy savings ranged from 0.03 to 0.95 MJ/kWh, whereas the GHG emission reduction ranged from 10 to 180 gCO2/kWh, depending on the country grid mix and the system boundary of the SG system considered. The findings demonstrate that although SG systems are energy efficient and reduce GHG emissions, investments in SG systems may not yield any benefits. Standardizing some methodologies and assumptions such as discount rates, time horizon and scrutinizing some key input data will result in more consistent estimates of costs and benefits, GHG emission reduction, and energy savings.