Choice, Charters, and Public School Competition

When historians review the changes in schools over the twentieth century, two things are likely to stand out. First, there has been a dramatic consolidation of school districts, leaving fewer but significantly larger districts. Second, schools have become more unionized. Historians may not see the

School choice comes in a variety of forms ranging from deciding where to live to home schooling. This Commentary contrasts two important forms of public-school choice that have developed: open enrollment programs and charter schools.

* The Concept of School Choice

In the twentieth century, the United States led the world's schooling transformation and went from a small, elite schooling system to one that was significantly changed in breadth and depth. Universal schooling with later and later schooling ages became the norm throughout the country.

But with these positive changes came others that made the schools less accessible to the direct control of local parents. First, school districts consolidated. In 1937, there were 119,000 separate public school districts. Today, there are fewer than 15,000.

One important result of this consolidation is less choice in public schooling. With a limited number of larger districts, the decision making and management of school districts effectively moves away from the local population. Larger districts also mean that there are more diverse preferences for what the parents want in their schools, and compromises must be made across the various interests.

Second, as districts consolidated, funding also changed dramatically. In 1930, less than ? percent of revenues for elementary and secondary schools came from the federal government, and less than one-fifth came from states, leaving over 80 percent to be raised locally. By 2000, the local share was down to 43 percent with both federal and state shares rising. As states have become more prominent in the funding of schools, they have also moved toward more centralization of decisions about the operations of the schools-with the result that school decisions have further migrated away from the parents and local voters and toward state bureaucracies.

If the parents directly control the funding of the schools, parents can exert some influence about what the school does, simply because the schools realize that their support is directly related to pleasing the parents. But, with large and diverse districts and with funding decisions made increasingly elsewhere, the incentives for schools to please the parents fall.

Parents can of course move to new school districts, and their ability to do so can exert a degree of competitive pressure on public schools. But this view of shopping across alternative jurisdictions does have important limitations. For a variety of reasons, the public schools might not look too different from each other. State restrictions, the limited viewpoints of school personnel, and other things could lead schools to be quite similar in approach, curriculum, and goals. Job locations can also dramatically limit the number of relevant school districts.

Restoring the ability of parents to enter easily into the schooling process will depend crucially on developing and sustaining new ways for them to exercise choices. The fundamental underlying the idea of expanded choice in public schooling is that freeing up consumer demand can have a variety of beneficial effects. Consumers can select the alternative that best meets their interests and desires. Importantly, since few consumers like overpriced goods, such demand pressure could lead to efficiency and innovation in education. If one school did not provide good value, it would tend to lose students to a competitor that offered more for the level of spending. It is precisely these incentives that are most important in assessing the issue of school choice.

Milton Friedman was one of the first scholars to promote the idea of expanded choice in schools. He argued in Capitalism and Freedom that the government may want to intervene in the area of education for a variety of reasons, but none of the potential reasons, including ensuring a minimal level of education in the population or enabling children of the poor to attend schools, requires that the government actually run the schools (Friedman, 1962). Friedman identified the now obvious alternative to government-run schooling-providing vouchers to parents. These vouchers would transfer funding to the school that a parent chooses.

Yet, with some exceptions in special circumstances such as the Cleveland situation or the early political anomaly of Milwaukee choice, the voucher idea has yet to be met with much policy success. Perhaps the most obvious factor is the rise of teacher unions. When vouchers were first suggested, unions were not pervasive. Their subsequent rise and increase in power has forever changed our ability to introduce any radical policy in schools. Competition in schools in particular implies that the job security of current personnel would be threatened, and any hint of even experimenting with choice has been vigorously attacked by the unions. They have been very effective at resisting any such change, including mounting powerful media campaigns to prevent citizen referenda on vouchers from being adopted.

A particularly effective argument in the public relations war over vouchers has been that giving money to private schools would harm the public schools and that we should instead be working to improve the public schools. A second argument is that private schools are not under the control of the government, so there is no accountability for the government funds that are provided. The sound bite summary has been the mantra of a number of people: "they are for choice but it should be restricted to public school choice." This position has been particularly popular among politicians who want to protect the existing public schools from any competitive pressures while still seeming open to more fundamental reforms of schools.

Yet citizen sentiment for expanded choice has generally increased over time, and this has led to a variety of innovations in school choice that fit the notion of public school choice. Importantly, they are not all the same, and they have very different incentive effects. Two quite different kinds of choice stand out: open enrollment or magnet school plans and charter schools. It is useful to review these in terms of outcomes and incentives.

* Public School Choice

A particularly popular version of public school choice involves an open enrollment plan. For example, any student could apply to go to a different school in their district rather than the one to which originally assigned. Or in a more expansive version, no initial assignment is made, and students apply to an ordered set of district schools. A common version of this has been the use of magnet schools that offer a specialized focus such as college preparatory or the arts.

Forms of open enrollment plans were the response of a number of Southern districts to the desegregation orders flowing from Brown v. Board of Education. In general, simple open enrollment plans were not found to satisfy the court requirements for desegregation of districts, but magnet schools (with racial balance restrictions) became a reasonably common policy approach (Armor, 1995). In 2001-02, three percent of all students attended a magnet school (Hoffman, 2003).

As a general rule, magnet schools produce few of the incentives that lie behind voucher plans. The flows of students are heavily controlled by the common restrictions that space must be available and that other purposes such as racial balance are met. Most important, however, these plans seldom have much effect on incentives in the schools. Under open enrollment, personnel in undersubscribed schools generally still have employment rights and would simply move to another school with more students. Opening open enrollment across districts conceptually provides stronger incentives but unattractive funding consequences, and the "if there is space at the school" clause generally stops all but some token movement.

A different development, charter schools, appears to offer stronger choice incentives. Charter schools differ dramatically by state, but the essential feature is that they are public schools that are allowed to operate to varying degrees outside of the normal public schools. They are schools of choice, surviving through their ability to attract sufficient numbers of students.

These schools can offer true competition to the regular public schools, because they can draw students away from poorly performing regular publics. Employment rights typically do not transfer between charters and regular publics, so there is potentially pressure on school personnel in charters to attract students. The pressure on regular public schools comes from the potential loss of students, which in turn lessens the demand for public schools and their teachers.

Since the nation's first charter school legislation was enacted into law in Minnesota in 1991, some 41 states and the District of Columbia have enacted legislation that provides for charter schools, although some had yet to open any schools by 2004. For the nation as a whole, charter schools increased from a handful in 1991 to close to 3,200 schools serving almost 800,000 students or over 1.5 percent of the public school population in 2004. In some places, charters have become quite significant. For example, in the 2003-04 school year, 17 percent of students in the District of Columbia, 8 percent of students in Arizona, and 4 percent of students in Michigan attended charter schools.

To date, studies of the outcomes of charters have been limited by some serious analytical difficulties. Because students voluntarily choose these schools, inferring the impact of the school, as distinct from the characteristics of the students that are attracted to it, is always difficult (this is referred to as a selection problem). Additionally, because charter schools are largely new, most are still going through a start-up phase, and it takes large inferences to know what they will look like in the steady state.

My own work provides some preliminary estimates of the performance of charters in Texas (Hanushek, Kain, Rivkin, and Branch, 2005). The simplest design that deals with the selection problems is to compare the average learning growth for individual students when in the regular public schools with their own performance in the charters. In this way, charter students become their own control group.

Three things come out of this in terms of quality indicators. First, on average charter schools perform very similarly to the regular public schools. But, second, start-up problems are real, and new charters do not perform as well as more established charters, while more established charters (those over two years in age) on average outperform the regular public schools of Texas. Third, there is a significant distribution of performance across both regular publics and privates. The good are good, and the bad are truly bad.

These findings are consistent with much of the other recent work, although there are some remaining uncertainties. The average North Carolina charter appears less effective than the average traditional public school (Bifulco and Ladd, 2004), while the average Florida charter is on par with the regular public schools after a start-up phase (Sass, 2005). On the other hand, relying upon comparisons between charter applicants in Chicago that were randomly accepted or randomly denied admission, Hoxby and Rockoff (2004) conclude that the three charter schools they observe significantly outperformed their regular school counterparts.

These results await, nonetheless, both the general maturation of more charter schools and the investigation of their performance in different settings.

Another important aspect of competitive markets is enforcing a discipline on the other participants-in this case the regular public schools. Is there any evidence that the regular public schools respond to the pressures of competition? Again, it is very early in the development of charters, but Hoxby (2003) introduces preliminary evidence that there are competitive improvements.

Our Texas study also provides informaftion on competition. If we look at the behavior of parents, we find that they are significantly more likely to withdraw their children from a poorly performing charter as compared to a well performing charter (Hanushek, Kain, Rivkin, and Branch, 2005). This finding is particularly important, because parents are not given information on the value-added of their charter school. The behavior of parents shows, however, that they are good consumers and that they can use the performance data that are available to infer the quality of the school. An early and continual criticism of the voucher idea is that parents are not good consumers, an assertion belied by the data.

* Conclusions

The idea of school choice is a natural extension of arguments about the benefits of competition to education. The clearest form, advocated originally by Milton Friedman and picked up by a wide variety of other people, is to give parents vouchers that allow them to shop for schools. While special circumstances have led to voucher programs in Milwaukee, Cleveland, and the District of Columbia, their growth has been slow and erratic.

On the other hand, alternative forms of choice-flying under the banner of public school choice-have become more pervasive. Some, but not all of these alternatives offer benefits that are similar to vouchers. Most notably, charter schools offer the possibility of options to students and parents that have the ancillary advantages of introducing competitive incentives for schools.

Charter schools are difficult to evaluate. Nonetheless, the best evidence currently available indicates that, after a start-up period, charters have as high if not higher value-added than regular public schools. As with regular public schools, however, there is a wide range of quality in charter schools. But, importantly, parents appear able to recognize the quality of charter schools and to act upon that information by exiting low-quality charters at significantly higher rates than better-quality charters.

Current personnel in the regular public schools resist expansion of charters because they provide what, in their mind, is undesirable competition. This resistance takes a variety of forms. In some states there are strong pressures to limit the number of charter schools. In others, arguments that all schools should have a "level playing field" are used to justify increasingly stringent restrictions on the operations of charter schools. If we are to obtain the benefits from choice and competition, these pressures should be resisted.