private school chains in chile: do better schools scale up?, cato policy analysis no. 682

8/6/2019 Private School Chains in Chile: Do Better Schools Scale Up?, Cato Policy Analysis No. 682

1/16

Executive Summary

There is a persistent debate over the role ofscale of operations in education. Some arguethat school franchises offer educational servic-es more effectively than do small independentschools. Skeptics counter that large, centralizedoperations create hard-to-manage bureaucra-cies and foster diseconomies of scale and thatsmall schools are more effective at promotinghigher-quality education. The answer to this

question has profound implications for U.S.education policy, because reliably scaling up thebest schools has proven to be a particularly dif-ficult problem. If there are policies that wouldmake it easier to replicate the most effectiveschools, systemwide educational quality could

be improved substantially.We can gain insight into this debate by ex-

amining Chiles national voucher program.This paper uses fourth-grade data to compareachievement in private franchise, private inde-pendent, and public schools in Chile. Our find-ings suggest that franchises have a large advan-tage over independent schools once student andpeer attributes and selectivity are controlled for.

We also find that further disaggregating schoolfranchise widens the larger franchise advantage.We conclude that policies oriented toward cre-ating incentives for private school owners tojoin or start up a franchise may have the poten-tial for improving educational outcomes.

Private School Chains in ChileDo Better Schools Scale Up?

by Gregory Elacqua, Dante Contreras, Felipe Salazar,and Humberto Santos

No. 682 August 16, 2011

Gregory Elacqua is the director of the Public Policy Institute, School of Business and Economics, UniversidadDiego Portales; Dante Contreras is professor of economics at the School of Business and Economics, Universidadde Chile; Felipe Salazar is a researcher at the Center for Comparative Education Policy; and Humberto Santosis a researcher at the Public Policy Institute, School of Business and Economics, Universidad Diego Portales.


2/16

Introduction

The optimal scale of operations of schoolsis a hotly debated issue in current education-

al policy reform discussions. One view is thatlarger schooling operations offer educationalservices more efficiently than small indepen-dent schools. Proponents argue that increas-ing the size of schooling operations wouldlower per pupil costs and free up resourcesfor use at the school and classroom levels.1Researchers also claim that large privateschool franchises promote the creation ofsound institutional environments in mem-ber schools.2

Advocates of large school chains also

maintain that such chains have an easiertime gaining credibility and legitimacy inthe community,3 and that larger schoolingoperations have more opportunities to ac-cess private investments to expand than dosmaller ones.4

These assertions have sparked two differ-ent trends in school management: consoli-dating public school districts and increas-ing public funding for private and charterschool franchises and Educational Manage-ment Organizations. Despite a growing pop-

ulation, more than 100,000 school districtshave been eliminated since 1938, a declineof nearly 90 percent.5 There is also a grow-ing number of private and charter schoolpartnerships and Educational ManagementOrganizations in the United States.6 For in-stance, Edison Schools, the United Stateslargest for-profit manager of schools, hasbecome one of the nations largest charterschool management organizations and hasincreased from slightly more than 200 char-ter schools in 1995 to more than 3,600 char-

ter schools in 2006.Critics fear that these reforms could have

potential negative unintended consequenc-es. They argue that large, centralized opera-tions will create hard-to-manage bureaucra-cies and foster diseconomies of scale due toassociated problems of managing complexorganizations and maintaining a sense of

order.7 Opponents are also concerned thatlarge schooling operations will make it dif-ficult to create a sense of community amongstudents, parents, teachers, and adminis-trators.8 Others are concerned that larger

schooling operations encourage more stan-dardization and less innovation.9

Some have argued that reducing the sizeof schooling operations is a more effectiveway to improve educational outcomes. Theyclaim that small schools can improve thequality of education by creating intimatelearning communities where students areencouraged by educators who know them.10

Small school advocates also argue thatsmaller schools foster cooperation betweenteachers, school administrators, and parents

and higher trust in the school community.11

Following these insights, many currentproposals for reform in the United Statesshare a vision of small, autonomous schoolsencouraged to strengthen school communi-ties.12 The small schools movement has alsomade significant progress in recent yearsFor example, the Bill and Melinda GatesFoundation invested more than $1 billion todivide large urban high schools in the UnitedStates. These resources partly funded the cre-ation of 197 small high schools in New York

City alone.13Much of the existing empirical evidence

on the optimal size of schooling operationsis mixed and often clouded by methodologi-cal limitations.14 The research that examinesthe benefits of private school franchises ver-sus small independent schools also suffersfrom thin data because it derives from theevaluation of small-scale programs. For in-stance, in its evaluation of Edison Schools,which was not a randomized study, research-ers found that the performance of these

schools varies.15 Similarly, the evaluations ofthe small high schools funded by the GatesFoundation also suggest that there is widevariation in the quality of these schools.16

The evidence on private school franchis-es and small private independent schools islimited because most educational systemsonly provide funding to public schools.17

2

The evidence onprivate schoolfranchises and

small private

independentschools is limited

because mosteducational

systems onlyprovide funding

to public schools.


3/16

We can gain insight into the distinctstrands of arguments on the optimal size ofschooling operations by examining Chilesnational voucher program. In 1981 Chilebegan financing public and most private

schools with vouchers. Private voucherschools currently account for over 50 per-cent of total enrollment and about one thirdof these schools belong to private voucherschool franchises. This paper compares theachievement of fourth-graders in privatevoucher franchise, private voucher indepen-dent, and public schools. The results present-ed in this study provide suggestive evidencethat, all else being equal, private voucher fran-chise schools are more effective than privatevoucher independent and public schools.

Background on Chile

During the 1980s, the military governmentin Chile (1973 to 1990) instituted a sweepingeducation reform package. First, the govern-ment decentralized the administration of pub-lic schools, transferring responsibility for pub-lic school management from the Ministry ofEducation to municipal governments, whose

maximum authority is the mayor. Second,the government introduced a flat per-pupil voucher scheme. Municipalities and privateschools that did not charge tuition started toreceive a per-student voucher for every child at-

tending their schools. As a result, enrollmentlosses began to have a direct effect on theireducation budgets. Elite private schools thatcharged tuition continued to operate withoutpublic funding.

Education in Chile has become increas-ingly privatized since the voucher reformswere introduced. In 1979, 12 percent of Chil-ean K12 students attended private schoolsthat received some public subsidy, and an-other 7 percent attended unsubsidized pri-vate schools. By 1990, 32 percent of students

attended private voucher schools. By 2009,enrollment in such schools had reached 48percent. Adding in the 7 percent of studentsin elite private nonvoucher schools leavesa majority of Chilean students in privateschools (see Figure 1).

The essential features of the national voucher system have remained in place forover three decades. The democratic govern-ments in power since 1990 have chosen to fo-cus on improving the quality of poor schools

3

In 1981 Chilebegan financingpublic and mostprivate schools

with vouchers,and privatevoucher schoolscurrentlyaccount for over50 percent oftotal enrollment

Figure 1Enrollment Share in Private and Public Schools (19792009)

Sources: Government of Chile, Ministry of Education, http://ded.mineduc.cl/DedPublico/archivos_de_datos and

http://www.simce.cl/index.php?id=448, and authors calculations.

Percenta

geo

ftotalprimaryan

d

secon

daryenro

llmen

t


4/16

4

through direct resource investments, whilemaintaining the organizational and fundingcomponents introduced in the 1980s.18

The data presented in Figures 2a and 2bsuggest that the private voucher school sec-

tor is essentially a cottage industry. Morethan 70 percent of private voucher schoolsare independent schools that do not belongto a franchise. Private voucher school fran-chises, which are defined in Chile as schools

Figure 2aDistribution of Primary Schools, by Private Voucher School Category (2008)

Figure 2b

Distribution of Primary Students, by Private Voucher School Category (2008)



Private voucher schools

Private voucher students


5/16

5

that belong to a network of schools that areoperated by the same legal private vouch-er school owner (sostenedor), account forabout one-third of private voucher schoolsand enrollments. Although this percentage

has remained relatively stable over time, there

is a slight downward trend in recent years,which is shown in Figures 3a and 3b. Most ofthe franchises are fairly small in scale, and al-most 50 percent of primary private voucherfranchise schools belong to franchises that

have fewer than four schools.Figure 3a

Distribution of Primary Schools, by Private Voucher School Category (200008)

Figure 3b

Distribution of Primary Students, by Private Voucher School Category (200008)

Private voucher schools

Private voucher students



Percen

tageo

fprivate

vouc

hersc

hoo

ls

Percen

tageo

fpriva

tevouc

herstu

den

ts


6/16

6

In this paper,we have only

one publicschool category,

while privatevoucher schools

are classifiedaccording towhether they

are independentor belong to a

franchise.

Private voucher schools are diverse inmembership. For-profit franchises, which areoften controlled by a group of local entrepre-neurs, in many cases with private sharehold-ers,19 represent 42 percent of all franchise

schools in Chile. These for-profit schoolsstand in varying degrees of contrast to in-dependent for-profit voucher schools, mostof which are owned and run by former pub-lic school teachers20 and which account forabout 88 percent of all independent schools.

On the other hand, nonprofit voucherschools, including Catholic,21 Protestant22

and secular organizations,23 represent a sig-nificant percentage of the franchise schools(58 percent) but only 12 percent of indepen-dent schools (see Figures 4a and 4b). Thus,

franchise schools are more likely to belongto religious congregations or nondenomina-tional foundations.

Methodology

Below is a brief summary of the method-

ology we used to compare the performanceof franchised, independent, and publicschools in Chile. A more detailed technicalexposition of our methodology is availablein our forthcoming journal paper on this

subject.

24

Empirical StrategyOur empirical model builds on previous

work by Patrick McEwan.25 We hypothesizethat student achievement, measured as stu-dent performance on standardized tests,can be modeled as a function of studentsocioeconomic characteristics (family back-ground, home resources, and peer groups26)

In this paper, we have only one pub-lic school category, while private voucher

schools are classified according to whetherthey are independent or belong to a fran-chise.27 Additionally, the latter are classifiedby size (number of schools) of the franchiseto which they belong.

We can predict the achievement of a typi-cal student in each school category. Here weuse the mean characteristics of private inde-

Figure 4a

Distribution of Private Voucher Primary Schools, by Ownership Type and Franchise

Status (2008)

Independent schools


7/16

7

A simplecomparisonof studentachievementin privatevoucher andpublic schoolsis unlikely to

give unbiasedestimates of theimpact of privatvoucher schoolson studentachievement.

pendent school students as a definition ofthis typical student. To measure the dif-ference in achievement between two schoolcategories, we subtract one prediction from

another. For example, we may estimate thecorrected difference between private fran-chise voucher schools and private indepen-dent voucher schools. This provides an ap-proximation of the expected increase (ordecrease) in test scores for the average privateindependent school student if he or she wereto attend a private franchise voucher school.

If the control variables perfectly accountfor student and peer demographics, thenthe above strategy yields unbiased results.However, there are variables that cannot

be measured. For instance, private voucherschools may be able to select more quali-fied students, on average, than their publicschool counterparts (school choice bias).Similarly, the average student attending aprivate voucher school may be more likelyto have other attributes (such as having par-ents who place a higher value on education)

than the average student attending publicschool (parental choice bias).

For these reasons, a simple comparison ofstudent achievement in private voucher and

public schools is unlikely to give unbiasedestimates of the impact of private voucherschools on student achievement. In order todiminish parental choice selection bias,28we use the methodology proposed by Lung-Fei Lee.29

DataThe previous models are estimated with

student data from Chiles national stan-dardized test, Sistema de Medicin de la Cali-dad de la Educacin (the System of Measure-

ment of the Quality of Education [SIMCE]),which assesses students in grades 4, 8 and10 in language, mathematics, history, andgeography, and in natural sciences in oddyears. In 2008 SIMCE evaluated 245,607fourth-graders, which represents 95 percentof total enrollment at that level. Student testscores are complemented with parent and

Franchise schools



Figure 4b

Distribution of Private Voucher Primary Schools, by Ownership Type and Franchise

Status (2008)


8/16

8

After controllingfor student andpeer attributes

and selection

bias, we finda positive and

significantprivate voucher

franchise schoolachievement

effect in Spanishand mathematics.

teacher questionnaires, which include so-cioeconomic and background informationon the students, their families, peers, andschools.

Several independent variables character-

ize student demographics. These include thestudents gender, parents educational at-tainment, self-reported household income,and the number of nonschool-relatedbooks in the students home. We calculatedstudent peer information by averaging in-dividual student information over all of thestudents in a given classroom. We also intro-duce a variable to indicate the relative isola-tion of the school. Although not reportedin the subsequent analysis, we also includedregional dummy variablesrelative to the

Metropolitan Regionin the regressions toaccount for differences across regions.

Empirical Results

A brief summary of the results is providedin Figures 5a and 5b. The zero base line rep-resents the performance of independent (i.e.,nonfranchise) private voucher schools. Theleft panel presents the test score gap betweenpublic schools and private voucher indepen-

dent schools. The right panel displays the gapbetween private voucher franchise schoolsand private voucher independent schools.30The first column in each panel presents theunadjusted difference in test scores. The sub-sequent columns present the differences af-ter accounting for individual and peer attri-butes and selection bias. The test scores werestandardized to a mean of 0 and a standarddeviation of 1, so that the gaps are measuredin terms of standard deviations.

The uncorrected estimates show that the

Spanish and mathematics achievement ofstudents that attend private voucher inde-pendent schools is higher, on average, thanthat of public schools students. However,the first column on the right panel also in-dicates a large unadjusted test score gap be-tween private voucher franchise and privatevoucher independent schools.

After controlling for student and peer attributes and selection bias, we also find a posi-tive and significant private voucher franchiseschool achievement effect in Spanish (0.086standard deviations) and mathematics (0.094

standard deviations). The corrected test scoregap between public and private voucher inde-pendent schools is negative and significantbut smallin the case of Spanish (0.037standard deviations) and negative but not sig-nificant in the case of mathematics.

These results provide some evidence ofthe effectiveness of private school franchisesHowever, a more precise analysis is needed tounderstand the optimal size of a franchiseHere we examine whether larger franchisesare more effective than smaller franchises

Figures 6a and 6b summarize the results forprivate voucher schools, by subject and fran-chise size. The results show that, after con-trolling for student and peer attributes andselection bias, private voucher schools thatbelong to a franchise of four or more schoolshave a more substantial advantage (between0.11 and 0.18 standard deviations) over pri-vate voucher independent schools than private voucher schools that belong to smallerfranchises of two or three schools (0.07 to0.09 standard deviations).31

To support these findings, we comparedtest scores in private voucher franchise andprivate voucher independent schools aftercontrolling for whether or not the privatevoucher school owners were Catholic. It is es-sential to control for the Catholic school ef-fect because previous research in Chile32 andin the United States33 has demonstrated thatCatholic schools, all else equal, outperformpublic schools and other private schools. Bydoing so, we avoid confounding the effectof attending a private franchise school with

the effect of a Catholic school. The results(not reported) do not change the substan-tial findings of our previous analysis, whichsuggests that the positive private voucherfranchise school effect is not related to thereligious affiliation of the schools.34

In order to test for consistency over timewe ran our model with 2002, 2005, and 2006


9/16


10/16

10

Figure 6a

Difference between School Types (by Franchise Size) and Private Voucher Independent Schools for Student with

Average Characteristics of Private Voucher Independent School Student (Fourth Grade 2008)

Standardd

eviations

S

tandardd

eviations

Figure 6b

Difference between School Types (by Franchise Size) and Private Voucher Independent Schools for Student with

Average Characteristics of Private Voucher Independent School Student (Fourth Grade 2008)

Sources: Government of Chile, Ministry of Education, http://ded.mineduc.cl/DedPublico/archivos_de_datos and http://www.simce.cl/index.php?id=448

and authors calculations.

*** Significant at 1 percent; ** Significant at 5 percent; * Significant at 10 percent.

Spanish

Mathematics


11/16

11

Some of thereasons thatmay explain thepositive private

school franchiseeffect includesubstantialbenefits of scalefor the bulkpurchases ofsupplies andequipment andimplementinginnovations incurriculum.

fourth grade SIMCE test score data. This ex-ercise (see Appendix) shows that our resultsare consistent over time. First, we find thatthe public/private voucher independentschool achievement gap is very narrow, and

in some cases not significant, indicatingthat there is not a significant difference be-tween these types of schools once studentand peer characteristics and selection biasare controlled for. Second, the results indi-cate that the positive effect associated withschool franchises is between 0.086 and 0.108standard deviations. Finally, we find that, allelse equal, schools that belong to a franchiseof four or more schools produce higher stu-dent achievement than schools that belongto smaller franchises (two or three schools).

Conclusions andPolicy Implications

This paper compares the academic a-chievement of fourth graders in privatevoucher franchise, private voucher indepen-dent, and public schools in Chile. Control-ling for individual and peer characteristicsand selection bias, the initial results suggestthat private voucher franchise school stu-

dents consistently outperform comparableprivate voucher independent school stu-dents. Private voucher independent schoolsby far the largest category of private voucherschoolsproduce similar test scores, all elseequal, as public schools.

We also considered the effect of thesize of private voucher school franchises.We find that, after controlling for individual and peer characteristics and selec-tion bias, larger private school franchises(four or more schools) outperform smaller

franchises (two or three schools). Studentachievement is more than 0.10 of a standarddeviation higher on the Spanish and math-ematics tests. On the other hand, schoolsthat belong to smaller franchises outper-form private independent voucher schools,but the differences are smaller. Our resultsare consistent over time and after control-

ling for the effect of the religious affiliationof the school.

Some of the reasons that may explainthe positive private school franchise effectinclude the substantial benefits of scale of

educational professionals and administra-tors,35 the bulk purchases of supplies andequipment, and the costs of implementinginnovations in curriculum.36 Private schoolfranchises may also be more likely to benefitfrom access to credit and private investmentthan smaller private independent schoolsin Chile. In addition, some argue that beingembedded within a larger organization fa-cilitates transactions between parents, teach-ers, administrators, and students37 and in-fluences the development of stronger school

communities.38

Before holding these results up as proofthat private school franchises are more ef-fective than private independent schools, weneed additional information on the factorsthat may influence a school owner to estab-lish a franchise that may determine educa-tional outcomes. For instance, high-achievingschools may be more likely to establish fran-chises (or to join a franchise) than low-qualityschools. In a competitive schooling environ-ment, low-quality schools may be unable to

attract students and additional resourcesneeded to expand operations. Private schoolfranchises may also require superior techni-cal skills to manage than small independentschools.

From a policy perspective, the results ofthis study also suggest that more informa-tion is needed on the factors that influenceschools incentives to establish franchises.For instance, how profitable are privateschool franchises? The data presented in Fig-ure 2 reveal that 70 percent of private vouch-

er schools do not belong to a franchise. Smallprivate independent schools may not haveincentives to establish a franchise if they areable to attract enough students and resourc-es to cover the opportunity costs of operat-ing a school. Survey evidence in Chile sug-gests that many of the independent private voucher school owners are former public


12/16

12

Policies orientedto creatingincentives

for schools

to establishfranchises or to

be managed by anorganization thatruns a network ofschools may have

the potentialfor increasing

educationaloutcomes.

school teachers.39 Data on the characteristicsof school owners would improve our under-standing of the complex decisions involvedin establishing a private school franchise.

The results of this paper offer some in-

sights for the debate in the United States onschool vouchers, the scale of operations ofpublic and private schools, and on the ben-efits of education management organiza-tions. The findings provide some groundsfor optimism about the effects of schoolvouchers and some (but not all) categories of

private schools on student achievement. Forinstance, policies oriented to create incen-tives for schools to establish franchises or tobe managed by an organization that runs anetwork of schools may have the potential

for increasing educational outcomes. How-ever, the results also provide grounds forcaution on the flooding of the educationalmarket with low-quality for-profit indepen-dent schools. We do not find significant dif-ferences in achievement between this groupof schools and public schools.

Spanish

Figure A1

Test Score Gap after Controlling for Student and Peer Attributes and Selection Bias

for Different SIMCEa Databases (Fourth Grade)

Appendix

Mathematics

Sources: Government of Chile, Ministry of Education, http://ded.mineduc.cl/DedPublico/archivos_de_datos andhttp://www.simce.cl/index.php?id=448, and authors calculations.aSystem of Measurement of the Quality of Education.*** Significant at 1 percent; ** Significant at 5 percent; * Significant at 10 percent.

Standard

Dev

iations

Above/

Be

low

Indepen

den

tVouc

her

Sc

hoo

lMean

Stan

dard

Dev

iations

Above

/Below

Indepen

dent

Vouc

her

Schoo

lMean


13/16

13

NotesThis policy analysis is based on the results of ourforthcoming paper The Effectiveness of PrivateSchool Franchises in Chiles National VoucherProgram, School Effectiveness and School Improve-ment, pp. 142.

1. John E. Chub, The Private Can Be Public,Education Next, Spring 2001.

2. Robert McMeekin, Networks of Schools,Ed-ucation Policy Analysis Archives 12, no. 16 (May 2003),http://epaa.asu.edu/epaa/v11n16/. McMeekin ar-gues that being part of a franchise provides a shar-ing experience within the network and facilitatesthe flow of information to teachers and adminis-trators on best practices.

3. Priscilla Wohlstetter et al., Improving ServiceDelivery in Education through Collaboration: AnExploratory Study of the Role of Cross-Sectoral

Alliances in the Development and Support ofCharter Schools, Social Science Quarterly 85 (2004):107896. Wohlstetter et al. maintain, on the basisof their research of charter school partnerships inthe United States, that well-established charterschool networks can build credibility for fundrais-ing more easily than small independent charterschools.

4. William C. Symonds, For-Profit Schools:Theyre Spreading Fast. Can Private CompaniesDo a Better Job of Educating Americas Kids?

Business Week, February 7, 2000.

5. National Center for Education Statistics(NCES), Digest of Education Statistics 2002, http://nces.ed.gov/pubs2003/2003060.pdf.

6. Carrie Lips, Edupreneurs: A Survey of For-Profit Education, Cato Institute Policy Analysisno. 386, November 20, 2000, p. 2.

7. Heath Brown et al., Scale of Operationsand Locus of Control in Market-versus Mission-Oriented Charter Schools, Social Science Quarterly85 (2004): 103551; Leanna Steifel et al., HighSchool Size: Effects on Budgets and Performancein New York City,Educational Evaluation and Poli-cy Analysis 22 (2000): 2239.

8. Paul T. Hill, Lawrence C. Pierce, and JamesW. Guthrie,Reinventing Public Education: How Con-tracting Can Transform Americas Schools (Chicago,IL: University of Chicago Press, 1997).

9. Clive Belfield and Henry M. Levin, Vouch-ers and Public Policy: When Ideology Trumps Ev-idence,American Journal of Education 111 (2005):54867.

10. Patricia A. Wasley et al., Small Schools: GreatStrides. A Study of New Small Schools in Chicago,http://www.bnkst.edu/gems/publications/smallschools.pdf; Robert G. Barker and Paul V. Gump,

Big School, Small School: High School Size and StudentBehavior (Stanford, CA: Stanford University Press,1964).

11. Valerie E. Lee and Susanna Loeb, School Sizein Chicago Elementary Schools: Effects on Teach-ers Attitudes and Students Achievement,Ameri-can Educational Research Journal37 (2000): 331.

12. Mary Anne Raywid, Small Schools: A ReformThat Works,Educational Leadership 55 (1998): 3439.

13. Eileen Foley, Approaches of Bill & MelindaGates Foundation-Funded Intermediary Orga-nizations to Structuring and Supporting SmallHigh Schools in New York City, 2010, http://www.eric.ed.gov/PDFS/ED510236.pdf.

14. Matthew Andrews, William Duncombe, and John Yinger, Revisiting Economies of Size inAmerican Education: Are We Any Closer to a Con-sensus? Economics of Education Review 21 (2002):24562. In their extensive review of the literature,

Andrews, Duncombe, and Yinger conclude: boththe claims of supporters of consolidation and de-tractors that claim small is beautiful have not ad-equately been tested using good evaluation meth-ods.

15. Brian Gill et al., State Takeover, School Restruc-turing, Private Management, and Student Achievement

in Philadelphia (Washington: Rand Corporation,2007).

16. American Institutes for Research (AIR) andSRI International, Early College High SchoolInitiative 20032005 Evaluation Report, 2006,http://www.earlycolleges.org/Downloads/ECHS_Eva_2003-2005.pdf; Howard S. Bloom, Saskia LevyThompson, and Rebecca Unterman, Transform-ing the High School Experience: How New YorkCitys New Small Schools Are Boosting Student

Achievement and Graduation Rates, 2010, http://www.mdrc.org/publications/560/full.pdf.

17. Organization for Economic Cooperation and

Development, Education at a Glance (Paris: OECD,2003).

18. See Sergio Martinic and Gregory Elacqua, Fin del Ciclo. Cambios en la Gobernanza del SistemaEducativo (Santiago, Chile: UNESCO, 2010). Seealso Organization for Economic Cooperationand Development, Revisin de Polticas Naciona-les de Educacin Chile (Paris: OECD, 2004). Onlytwo significant modifications have been intro-


14/16

14

duced since 1980. The first was in 1994, when theministry instituted a shared financing schemethat allowed all private voucher schoolsbothelementary and secondaryand public second-ary schools to charge limited tuition. See PedroMontt et al.,Hacia un Sistema Escolar Descentraliza-do, Slido y Fuerte: El Diseo y las Capacidades Hacen

la Diferencia (Santiago, Chile: Ministerio de Edu-cacin de Chile, 2006). The second was in 2008,when the Chilean legislature enacted the Ad-

justed Voucher Law (Ley de Subvencin EscolarPreferencial or SEP). The SEP law recognizes thatit is more costly to educate disadvantaged stu-dents by introducing an extra per-pupil subsidy(50 percent over the base voucher) for students oflow socioeconomic status. See Gregory Elacqua,rsula Mosqueira, and Humberto Santos, LaToma de Decisiones de un Sostenedor: Anlisisa Partir de la Ley SEP, 2009, http://www.cpce.cl/publicaciones/serie-en-foco-educacion, for de-tails on the Ley SEP.

19. Gregory Elacqua, The Politics of EducationReform in Chile: When Ideology Trumps Evi-dence (paper presented at the Latin AmericanStudies Association Annual Conference, Mon-treal, Canada, August, 2007).

20. Javier Corvaln, Gregory Elacqua and Fe-lipe Salazar, El Sector Particular Subvencionadode la Educacin Chilena: descripcin, tipologi-zacin, anlisis de su dinmica y percepcionesfrente a las nuevas regulaciones, 2008, http://ded.mineduc.cl/mineduc/ded/documentos/FONIDE_Informe_CIDE_UDP.pdf.

21. Catholic voucher schools are operated byreligious orders, parishes, archdioceses, and reli-gious foundations.

22. Protestant church schools include Method-ist, Baptist, Seventh Day Adventist, Anglican, Lu-theran, and Presbyterian churches. There are fourprivate voucher schools of other religious orien-tations.

23. Most of the secular nonprofit schools arebranches of foundations that were created forother specific tasks, such as the Rural Social De-

velopment Corporation.

24. Gregory Elacqua, Dante Contreras, FelipeSalazar, and Humberto Santos, The Effective-ness of Private School Franchises in Chiles Na-tional Voucher Program, School Effectiveness andSchool Improvement, forthcoming, pp.142.

25. Patrick McEwan, The Effectiveness of Pub-lic, Catholic, and Non-religious Private Schoolsin Chiles Voucher System,Education Economics9(2001): 10328.

26. We include peer group controls because abody of literature has documented the positivespillover effects of having high-ability peers andthe negative effects of being surrounded by disad-

vantaged students. See, for instance, Ron W. Zim-mer and Eugenia F. Toma, Peer Effects in Privateand Public Schools across Countries, Journal of

Policy Analysis and Management19 (2000): 7592.27. We do not include the private nonvoucherschools in this analysis. This set of schools chargehigh tuition, do not receive per-pupil subsidiesand are mainly focused on high-income studentsIn a previous version of this paper, we includedprivate nonvoucher schools in our analysis. Theresults (available upon request) do not changethe substantive conclusions reported here.

28. In order to control for school choice bias,information on school selection practices wouldbe required.

29. Lung-Fei Lee, Generalized Econometric Mod-els with Selectivity,Econometrica51 (1983): 50712.

30. We use private independent schools as theomitted reference category because we are inter-ested in comparing private franchise and inde-pendent school outcomes.

31. In a separate analysis not reported here, weexcluded the largest private voucher school fran-chise in Chile that has 147 schools to make sureit was not confounding our findings. The results(available upon request) do not change the sub-stantive conclusions reported here.

32. McEwan.

33. Anthony S. Bryk, Valerie E. Lee, and PeterB. Holland, Catholic Schools and the Common Good(Cambridge, MA: Harvard University Press, 1993)

34. In addition, our findings are substantivelysimilar when we only consider urban schools inthe Metropolitan Region of Santiago. These re-sults are available upon request.

35. Chubb.

36. William Duncombe and John Yinger, Does

School District Consolidation Cut Costs?Educa-tion Finance and Policy 2 (2007): 34176.

37. McMeekin.

38. Andrew K. Smith and Priscilla WohlstetterReform Through School Networks: A New Kindof Authority and Accountability, Educational

Policy 15 (2001): 499519; Katrina Bulkley andJennifer Hicks, Educational Management Orga-


15/16

15

nizations and the Development of ProfessionalCommunity in Charter Schools, Occasional Pa-per no. 69, 2003, National Center for the Studyof Privatization in Education (NCSPE), Teachers

College, Columbia University, http://www.ncspe.org/publications_files/133_OP69.pdf.

39. Corvaln, Elacqua, and Salazar.


16/16

RECENT STUDIES IN THECATO INSTITUTE POLICY ANALYSIS SERIES

681. Capital Inadequacies: The Dismal Failure of the Basel Regime of Bank

Capital Regulation by Kevin Dowd, Martin Hutchinson, Simon Ashby, and JimiM. Hinchliffe (July 29, 2011)

680. Intercity Buses: The Forgotten Mode by Randal OToole (June 29, 2011)

679. The Subprime Lending Debacle: Competitive Private Markets Are theSolution, Not the Problem by Patric H. Hendershott and Kevin Villani(June 20, 2011)

678. Federal Higher Education Policy and the Profitable Nonprofits by Vance H.Fried (June 15, 2011)

677. The Other Lottery: Are Philanthropists Backing the Best Charter Schools?by Andrew J. Coulson (June 6, 2011)

676. Crony Capitalism and Social Engineering: The Case against Tax-IncrementFinancing by Randal OToole (May 18, 2011)

675. Leashing the Surveillance State: How to Reform Patriot Act SurveillanceAuthorities by Julian Sanchez (May 16, 2011)

674. Fannie Mae, Freddie Mac, and the Future of Federal Housing Finance

Policy: A Study of Regulatory Privilege by David Reiss (April 18, 2011)

673. Bankrupt: Entitlements and the Federal Budget by Michael D. Tanner (March28, 2011)

672. The Case for Gridlock by Marcus E. Ethridge (January 27, 2011)

671. Marriage against the State: Toward a New View of Civil Marriage by JasonKuznicki (January 12, 2011)

670. Fixing Transit: The Case for Privatization by Randal OToole (November 10,

2010)

669. Congress Should Account for the Excess Burden of Taxation by ChristopherJ. Conover (October 13, 2010)

private school chains in chile: do better schools scale up?, cato policy analysis no. 682

Documents