The Education Competition Index: Quantifying competitive pressure in America’s 125 largest school districts
By Amber M. Northern and Michael J. Petrilli
We at Fordham view this as a healthy development, both because we believe in the fundamental right of parents to choose schools that work best for their children, and because of the large and ever-growing research literaturedemonstrating that competition improves achievement in traditional public schools. That “competitive effects” are largely positive should be seen as good news for everyone, as all of us should root for every sector of American education to improve. And it means that the whole “school choice versus improving traditional public schools” debate presents a false dichotomy; we can do both at the same time. Indeed, embracing school choice is a valuable strategy for improving traditional public schools.
Yet, despite the amount of attention that school choice receives in the media and among policy wonks, politicians, and adult interest groups, the extent of actual competition in major school districts is not well understood. We were curious: Which education markets in America are the most competitive? And which markets have education reformers and choice-encouragers neglected or failed to penetrate?
Those questions prompted this analysis, conducted by David Griffith and Jeanette Luna, Fordham’s associate director of research and research associate, respectively. The study seeks to quantify the extent to which competition is occurring by estimating the number of students enrolled in charter, private, and homeschools in each of the nation’s 125 largest school districts in spring 2020 and then dividing that sum by an estimate of a given district’s total student population (which includes students in traditional public schools). The resulting quotient—the report’s measure of the competition facing a district—is the combined market share of all non-district alternatives. While this is not a perfect measure (we can’t account for inter-district open enrollment, for instance), it is as good an estimate as current data allow.