350651_Pa616_1stClass.qxp 4/1/2008 3:29 PM Page 1 No. 616 April 16, 2008 ������� Dismal Science The Shortcomings of U.S. School Choice Research and How to Address Them by John Merrifield Executive Summary Pressing questions about the merits of market ments of market systems, including profit, price accountability in K-12 education have spawned a change, market entry, and product differentia- large scholarly literature. Unfortunately, much of tion—factors that are normally central to any dis- that literature is of limited relevance, and some of cussion of market effects. In essence, researchers it is misleading. The studies most widely cited in have drawn conclusions about apples by studying the United States used intense scrutiny of a few lemons. small-scale, restriction-laden U.S. programs—and To address the need for credible evidence on a handful of larger but still restriction-laden for- the effects of genuine education markets, econo- eign school choice expansions—to assert general mists should look to simulation models, indirect conclusions about the effects of “choice,” “com- evidence such as outcomes in similar industries, petition,” and “markets.” The most intensely and school systems abroad that enjoy varying studied programs lack most or all of the key ele- degrees of market accountability. _____________________________________________________________________________________________________ John Merrifield is a professor of economics at the University of Texas–San Antonio, editor of the Journal of School Choice, and author of Parental Choices as an Education Reform Catalyst: Global Lessons (2005) and The School Choice Wars (2001). PA Masthead.indd 1 2/9/06 2:08:34 PM 350651_Pa616_1stClass.qxp 4/1/2008 3:29 PM Page 2 Prices determined The difficulty of charging tuition when by supply and Introduction the government offers “free” schooling is the main barrier to market entry faced by private demand are a key The free market is the primary form of schools in most of the world. That difficulty attribute of organization and accountability for most of not only impairs the effectiveness of private the economy, and increased reliance on mar- schools, it affects the kinds of private schools markets, but they 7 kets was the common denominator of suc- that can exist. The presence of “free” govern- are almost cessful 20th century economic reforms.1 ment schools artificially favors certain kinds unheard of in These facts, coupled with widespread dissatis- of private schools, such as those that faction over the quality of existing education- K-12 education. al services, have raised interest in market- • have large endowments; inspired education reforms that fall under the • have access to church subsidies; catch-all term “school choice.” • can offer inexpensive services that pub- But despite considerable effort to discover lic schools cannot, such as devotional the effects of “school choice,” much that could religious instruction; and qualify as evidence has not been directly stud- • can offer high-value subject or pedagog- ied, and much relevant information is not ic specializations that appeal to only a widely known. The limited educational choic- small share of the population. es available in the United States have been heavily researched, telling us how certain test Indeed, that artificially skewed and nar- scores respond when some families make use rowed menu of school types describes most of current alternatives to their assigned public of the existing U.S. private education sector. school.2 But such constrained choices provide But while U.S. school choice research has little evidence about how true education mar- focused on very limited existing reforms and kets might transform the status quo. Indeed, the small and highly skewed niche of American Harvard economist Caroline Hoxby (whose private schools, scholars have often used their work has focused on U.S. schooling) believes findings to make sweeping claims about com- that “[market-based] reforms would propel petition and market accountability in educa- American schools into wholly unknown terri- tion.8 Much of the alleged direct evidence of, tory.”3 A RAND study determined that “none for example, the effects of “large-scale voucher of the important empirical questions has been programs,” or market accountability through answered definitively. Even the strongest full-fledged competition, is thus wrong, mis- [U.S.] evidence is based on programs that have leading, or irrelevant.9 Misconceptions are been operating for only a short period of time more stubborn foes than ignorance, so undo- with a small number of participants.”4 ing the effects of this imagined evidence is a The novelty and minuteness of existing U.S. considerable challenge, and it is the key aim of school choice programs are not the only fac- this review.10 I also describe the sorts of studies tors that limit their value in assessing the mer- that can be undertaken (and in some cases, its of free-market education. Several key already have been undertaken) to truly deter- aspects of market accountability are virtually mine the merits of market education. absent from those programs: price change, easy market entry, and the profit motive, among others. Prices determined by supply and What Is “School Choice”? demand are a key attribute of markets, but they are almost unheard of in K-12 education—even Before diving into an analysis of the “school under most school choice programs.5 Further- choice” research, the term itself merits clarifica- more, existing private schools’ tuition rates are tion. Except when choice occurs through resi- greatly distorted by the taxpayer-funded com- dential relocation, “school choice” means a pol- petition from “free” public schools.6 icy that improves access to alternatives to the 2 350651_Pa616_1stClass.qxp 4/1/2008 3:29 PM Page 3 assigned public school. Improved access, which choice among “comprehensively uniform” typically results from a drop in the relative schooling options.14 Because school districts “price” of some alternative, may be universal or cannot assign children to specialized schools, limited to target groups, such as low-income choice by residential relocation cannot prompt families or children in public schools formally school systems to focus schools on specific designated as failing.11 Therefore, “school subject themes, or specific types of students, or choice evidence” is mostly what we know about specific teaching styles. the effects of changing the relative prices of Supplementing residence choice are eleven actual and potential schooling options. Note ways (#2–#12) to enhance choice, which means that subsidies like vouchers and tax credits can selectively or universally reducing the relative lower prices from families’ perspective while prices of the schooling options, sometimes by raising them from the school operators’ per- the creation of new schools. These school spective. By reducing disposable income, school choice options are as follows: taxes increase the relative price of private (#2) Chartered, nondistrict public schools, schooling. Interpretations of school choice which are authorized by law in 40 states.15 effects need to take account consistently of how Most charter laws let entities other than exist- much policy reforms change the relative prices ing school districts authorize publicly funded, to families and educators, and what fraction of less-regulated alternatives to assigned public Much of the educators and schoolchildren had access to the schools.16 As unzoned alternatives, charter alleged evidence price changes. If a price decrease significantly owners can pursue topical or pedagogical spe- of the effects increases interest in alternatives to the assigned cialization, but their inability to turn away public school, an entrepreneurial response may mission-incompatible children can under- of “large-scale yield a much-changed menu of schooling mine such efforts. Every charter law creates voucher 17 options. two price controls. Charter schools may not programs” Since the array of policy options for ex- charge tuition, which leaves it to the political panding school choice is poorly understood, process to set the amount paid to charter is wrong, the next section briefly describes the policy operators. A ban on parental copayments pre- misleading, or options that lower the cost, or enhance the vents charters from offering services that cost benefits, of leaving the assigned neighbor- more than the state’s per-child payment (cre- irrelevant. hood public school.12 The third section dis- ating a price “ceiling”), while the guaranteed cusses false evidence and false assumptions state funding gives operators no incentive to used to interpret evidence, while the fourth charge less for services that cost less than the outlines sources of valid evidence. per-child payment (creating a price “floor”). (#3) Choice among district public schools. All but five states (Alabama, Maryland, Nevada, Varieties of School Choice North Carolina, and Virginia) pay at least statu- tory lip service to “public school choice.”18 In Most U.S. children have an assigned public the states that don’t undermine it with restric- school based on their home address.13 Private tions or allow districts to opt out, open enroll- school tuition can severely strain family bud- ment among district public schools creates at gets, so residence choice is the most used least temporarily (until the best schools fill) school choice strategy (#1). The costs of this some alternatives