Objective We compare academic achievement in charter schools versus two types of traditional public school in the State of Michigan over a 10-year period. Charter schools serve as a reform measure for failing public schools, so the natural research question is: Do charter schools generate higher achievement levels than observationally comparable public schools? Methods We have assembled a longitudinal data set spanning academic years 2002/2003-2011/2012 containing proficiency rates on standardized math and reading tests for Grades 4, 7, and 11. Our set of control variables includes demographic measures, free lunch eligibility, school characteristics, funding per student, and locational measures. We model unobserved heterogeneity using random effects estimation. Results We find that Michigan charter schools significantly underperform traditional public schools in both subjects and in all three grade levels early in the study period. These gaps narrow considerably, and in some cases disappear, by the end of the period. Conclusion Michigan is noteworthy among states with charter schools for its deregulated and competitive charter environment. Our results suggest that Michigan charters nevertheless were doing about as well on the state's academic achievement tests as observationally comparable public schools by the end of the period.