Charter schools are a relatively new phenomenon in American education. Since the first charter school opened in Minnesota in 1991, they have expanded to 42 states and represent 6.2% of all public schools in the country. This growth has been attributed to a number of factors, chief among them evidence that charter schools can improve performance (Lamdin and Mintrom, 1997). While there is a substantial evidence for relative performance benefits of charter schooling (e.g. CREDO, 2015) far less research been conducted on the efficiency of charter schools relative to traditional public schools. What research there is has produced both positive (e.g. Wolf et. al., 2014) and negative results (e.g. Carpenter and Noller, 2010). What can account for the disparity in these findings? In this paper, I make the case that differences in charter efficiency may be accounted for by differences in their level of autonomy from the school district. I base this argument on economic theories that the devolution of power to the lowest level possible tends to produce gains in efficiency (Johnson, 1991; Duncombe and Yinger, 1997). Those that are "on the ground" are thought to be more effective at monitoring expenditures, and allocations of resources have to pass through less "red tape (Hess 2006)." In addition, more autonomous charter schools better fit the original purpose of charter schools in devolving power from centralized authorities (Budde, 1996). In order to test this theory, I take advantage of a unique situation that exists in Milwaukee, Wisconsin in which three types of charter schools with varying levels of autonomy operate simultaneously. Using school type as a proxy for autonomy, I find that more independent charter schools are more efficient than traditional public schools and charter schools with less autonomy.