What are charter schools and do they deliver?

Charter schools are tuition-free, publicly funded schools. Charter school leaders accept greater accountability in exchange for greater autonomy. About 3 million students attend charter schools across 43 states and the District of Columbia. Assessing whether charter schools work is a complicated question. While the last four U.S. presidents—Clinton, Bush, Obama, and Trump—all expressed support for charter schools, support is splintering along party lines.

Traditionally, local school districts operate public schools. They make decisions about how schools run, in some cases through collective bargaining agreements with teachers unions. Most students are assigned to traditional public schools based on the neighborhood in which they live.

Charter schools operate differently. A charter school begins with an application that describes the proposed school’s mission, curriculum, management structure, finances, and other characteristics. That proposal goes to a government-approved authorizing agency such as a school district or a university. Who can serve as an authorizer—like many aspects of charter school law—varies from state to state. If the authorizer approves the proposal, it creates a contract (“charter”) with the school’s governing board that describes the school’s rights, responsibilities, and performance expectations. This is the core of what has been called the “charter school bargain.” School leaders accept greater accountability (e.g., the possibility of closure for poor performance) in exchange for greater autonomy (e.g., the ability to pursue a specialized theme).

Along with accountability and autonomy, the other principle at the foundation of the charter school model is choice. Students are not assigned to charter schools based on where they live. Instead, their families request a seat in the school. Some cities use unified enrollment systems that allow families to request many schools at once, ranked in order of preference, and then assign students to schools using a placement algorithm.

The principles of accountability, autonomy, and choice are interconnected. For example, choice serves as a form of market accountability (since schools must attract families to stay open), while school leaders can use their autonomy to differentiate themselves and create a rich assortment of choices for families.

Since their formative days in the early 1990s, charter schools have drawn support from unusual political coalitions, with different groups seeing different opportunities. Many conservatives like the market competition and limited government control. Many progressives like that all families, not just the wealthy, get to choose their children’s schools. Even some union leaders saw early promise in charter schools as a way to give teachers more voice in how schools run. In fact, the last four U.S. presidents—Clinton, Bush, Obama, and Trump—all expressed support.

However, the politics of charters are changing. Teachers unions have long been charter schools’ most powerful opponents—they are skeptical that charter school teachers actually have more control and frustrated that only 11% of charter schools employ unionized teachers. More generally, support is splintering along party lines as Democrats—especially white progressives—have become increasingly opposed. Some of this polarization predates the Trump administration. For example, many prominent Democrats, including Senators Elizabeth Warren and Bernie Sanders, opposed a 2016 Massachusetts referendum to allow more charter schools (a referendum that was soundly defeated). However, the emergence of President Trump and Secretary of Education Betsy DeVos as the most prominent charter school supporters has accelerated the polarization in public opinion.

The specific critiques of charter schools vary. For example, some dislike that charter schools attract public funding (and students) that otherwise would have gone to traditional public schools. Others dislike that for-profit education management organizations can operate charter schools in some states (about 12% of charter schools had this status in 2016-17, according to the National Alliance for Public Charter Schools). Others worry about a lack of transparency or public control—or that charters haven’t lived up to their promise.

Despite the attention these schools get, only about 6% of U.S. public school students attended a charter school in 2016-17. As illustrated by the light blue line in the figure below, this percentage has increased since the turn of the century, although the rate of growth appears to be tapering off.

Notably, a disproportionate share of charter school students are either Hispanic (33% of all charter students) or Black (26%). More than 10% of Black public school students and 7% of Hispanic public school students attended a charter school in 2016-17, compared to about 4% of white students.

This partly reflects the fact that a much larger share of charter schools (57%) than traditional public schools (25%) are operating in cities. In many cities, including Detroit and Washington, D.C., a majority or near-majority of public school students attend a charter school. At the extreme end of the spectrum, New Orleans has an all-charter public school system.

Assessing whether charter schools work requires clarity about their goals and good measures of how well they achieve those goals. It is not clear that we have either.

The most commonly reported studies compare students’ academic performance in charter versus non-charter schools. These comparisons are methodologically challenging, since charter school students differ from other public school students in ways that could affect their performance. Researchers need to find reasonable comparison groups.

One approach is to examine the random lotteries often used when a school receives more applicants than it can accommodate. Researchers tend to find that lottery winners perform much better than lottery losers, but these comparisons come with major caveats. Only schools with excess demand conduct lotteries, and it’s no surprise that students who get into the most sought-after schools perform well.

The Center for Research on Education Outcomes (CREDO) at Stanford University has taken a different approach. CREDO obtains student-level data from many states and districts and then matches charter school students with traditional public school students who appear similar in many ways (e.g., students of the same race, gender, and poverty status, with similar prior test scores, who previously attended the same schools). Then they compare charter students’ test scores with the scores of their matched non-charter pairs. This method isn’t foolproof—for example, the variables used for matching could miss something important—but it is probably the best approach to date for charter/non-charter comparisons that represent the broad population of charter schools.

The figure below shows some notable results from the CREDO studies. The key takeaway is that charter school students, in general, perform about the same as their matched peers in the traditional public schools, but there is variation across different types of schools and groups of students. For example, students in urban charter schools generally perform better than their matched pairs—likely for an assortment of reasons—while students in online charter schools perform much worse.

However, we should consider whether a charters versus traditional public schools comparison is the right measure in the first place. Part of the rationale for charter schools is that they should generate innovation and competitive pressure that improve charter and non-charter schools alike. If those improvements occur—and there is some evidence for this—these comparisons might understate the benefits of charter schools. If charter schools harm traditional public schools by, for example, reducing funding or creating funding uncertainty—and there is some evidence for this, too—these comparisons might understate the costs of charter schools.

Regardless, test score comparisons paint an incomplete picture of charter school performance. We care about a much broader set of outcomes, including how charter schools affect racial segregation, to what extent they create options for disadvantaged families, and whether they are truly producing innovative school models. The related research is too expansive for this overview but summarized nicely in a 2015 NBER report.

When presidential candidates talk about charter schools, it is more about politics and principle than policy, since education is largely left to the states. However, the federal government plays a role. For example, the U.S. Department of Education administers the Charter Schools Program (CSP), which provides funds to assist with matters such as charter school start-up, facilities acquisition, and replication. Congress allotted $440 million for the CSP in FY 2019—a figure that has increased substantially during the Trump administration. Federal policymakers could increase or decrease overall funding and add or remove stipulations for that funding (e.g., keeping federal funds from reaching for-profit charter schools).

State and local policymakers regularly make consequential decisions about charter schools. These decisions cover a wide range of issues, including funding formulas, caps on the number of schools or seats allowed, determinations of which schools (or types of schools) to open, rules about which students get priority access to high-demand schools, and requirements for open meetings and records.