Association for Public Policy Analysis & Management

Charter schools, which are publicly funded but independently operated, are a popular policy effort to improve academic achievement in the United States, especially for disadvantaged children. Part of the rationale for charter schools is that they are allowed greater freedom to innovate in exchange for increased accountability for performance. Prior research on charter schools indicates that they vary widely in terms of quality, and that lower-quality charter schools are more likely to close (Baude et al. 2014, Chingos and West 2015).

Even if the charter sector evolves over time to produce schools with high average quality, the opening and closing of schools may serve to exacerbate educational inequality. This would be the case, for example, if families differ in their ability to discern high-quality charters from low-quality ones. This means that the opening and closing of charter schools, coupled with uncertainty about their quality, may impose school-induced negative impacts on achievement that are disproportionately borne by certain types of families. For example, more advantaged parents may disproportionately flock to new schools judged to be high quality or more quickly leave schools revealed to be low quality.

This paper develops a framework linking household uncertainty and learning about charter school quality with charter school growth patterns to analyze key policy outcomes. In many states with statewide caps on the number of charter schools, it took years for these caps to start binding and restrict new charter school entry. Uncertainty about charter school quality may play a large role in explaining charter school opening patterns and individual charter school growth patterns. To understand these interrelated components, we develop a model of household learning about charter school quality and then nest this model of household learning into an equilibrium framework, where charter school entrepreneurs decide whether or not to enter a particular market to compete with a traditional public school. If better-informed households are more likely to take advantage of better charter schools, then more thorough and targeted information campaigns may be necessary in order to reduce disparities in academic achievement. The degree to which uncertainty reflects demand of charter schools can also help understand how to more effectively increase demand of charter schools if that is a desired outcome by policymakers.

We then use detailed micro-data on households and schools from North Carolina to test our model’s theoretical implications. Preliminary results indicate that charter schools with higher shares of educated parents grow over time, schools that close have lower average test score growth, and charters are more likely to open in larger districts with higher shares of educated parents. This is consistent with the existence of substantial uncertainty about charter school quality, households initially being overly pessimistic about the distribution of charter school quality, and there being variation in the speed at which different types of households learn about charter school quality.