ÃÍãÏ ÃÈæ ÒäØ
08-03-2012, 01:00 PM
Catholic Schools, Charter Schools, and Urban Neighborhoods
Margaret F. Brinig & Nicole Stelle Garnett
INTRODUCTION
This Article addresses previously unstudied implications of two
dramatic shifts in the American educational landscape. The first shift
is the rapid disappearance of urban Catholic schools. More than
1,600 Catholic elementary and secondary schools, most of them
located in urban neighborhoods, have closed during the last two
decades.1 The Archdiocese of Chicago alone (the subject of our
study) has closed 148 schools since 1984.2 Since the economic and
demographic realities underlying urban Catholic school closures
persist, this trend likely will continue and even accelerate in coming
years. The second shift is the rise of charter schools. In 2010 more
than 1.7 million children were enrolled in 5,400 charter schools in the
United States.3 During the 2009–10 school year, there were
† Fritz Duda Family Chair in Law, University of Notre Dame Law School.
†† Professor of Law, University of Notre Dame Law School.
We are particularly indebted to Sister Mary Paul McCaughey, OP, Superintendent of
Catholic Schools, Archdiocese of Chicago, for her invaluable insights. We received many
helpful comments at the symposium on Understanding Education in the United States: Its
Legal and Social Implications, held at the University of Chicago Law School on June 17 and
18, 2011. Alison Curran, Peter Reed, and Michael Wilde provided excellent research
assistance.
1 See Richard W. Garnett, Treasure A.C.E., Natl Rev Online (Sept 10, 2008), online at
http://www.nationalreview.com/articles/225595/treasure-c-e/richard-w-garnett (visited Oct 21,
2011). See also Peter Meyer, Can Catholic Schools Be Saved?, 7 Educ Next 12, 16 (Spring 2007);
Sol Stern, Save the Catholic Schools!, 17 City J 74, 74–76 (Spring 2007); Mary Ann Zehr, Catholic
Schools’ Mission to Serve Needy Children Jeopardized by Closings, 26 Educ Wk 16–17 (Mar 8,
2005).
2 Paul Simons, Closed School History: 1984–2004 *2 (Office of Catholic Schools
Archdiocese of Chicago 2004), online at http://www.illinoisloop.org/cath_closed_school_84
_04.pdf (visited Oct 21, 2011).
3 All about Charter Schools: Quick Facts (Center for Education Reform 2011), online at
http://www.edreform.com/issues/choice-charter-schools/facts/ (visited Oct 21, 2011).
32 The University of Chicago Law Review [79:31
104 charter schools in the city of Chicago,4 24 of which opened during
the period of our study.5
Although we are intrigued by the questions raised by the
extensive literature on Catholic and charter schools’ strengths as
educational institutions,6 we do not address them here. Instead, we
raise new questions about how Catholic and charter schools function
as community institutions. These questions are important ones.
Catholic schools are vanishing from the urban neighborhoods where
they have operated for decades—in some cases, for over a century—
and are being replaced by educational institutions that did not exist
anywhere in the United States two decades ago.7 Yet virtually
nothing is known about the impact the transition will have on urban
neighborhoods, many of which already struggle with disorder, crime,
and poverty.
This is the third in a series of papers exploring the effects of
Catholic school closures on urban neighborhoods. In previous studies,
we linked Catholic school closures to increased disorder and crime,
4 Illinois Network of Charter Schools, Illinois Public Charter Schools: Profiles 1 (2010),
online at http://incschools.org/docs/INCS_SchoolProfiles2010_Final.pdf (visited Oct 21, 2011).
5 See Office of New Schools, 2009–2010 Charter and Contract Schools Performance
Report (Chicago Public Schools 2010), online at http://www.cps.edu/NewSchools/Documents
/2009-2010_PerformanceReport.pdf (visited Oct 21, 2011).
6 Numerous scholars have demonstrated that Catholic schools tend to outperform their
public counterparts, especially at the challenging task of educating underprivileged minority
students. See, for example, Andrew M. Greeley, Catholic High Schools and Minority
Students 108 (Transaction Books 1982); James S. Coleman, Thomas Hoffer, and Sally Kilgore,
High School Achievement: Public, Catholic, and Private Schools Compared 143–46 (Basic
Books 1982). Scholars are divided about whether charter schools outperform traditional public
schools. See Julian R. Betts and Y. Emily Tang, Value-Added and Experimental Studies of the
Effect of Charter Schools on Student Achievement: A Literature Review 26 (Washington Bothell
2008), online at http://www.crpe.org/cs/crpe/download/csr_files/pub_ncsrp_bettstang_dec08.pdf
(visited Oct 21, 2011); Multiple Choice: Charter School Performance in 16 States 1 (Stanford
June 2009), online at http://credo.stanford.edu/reports/MULTIPLE_CHOICE_CREDO.pdf
(visited Dec 31, 2011). The available evidence suggests charter schools in Chicago do
outperform Chicago public schools. See Achievement and Attainment in Chicago Charter
Schools 21 (RAND 2009), online at http://www.rand.org/content/dam/rand/pubs/technical
_reports/2009/RAND_TR585-1.pdf (visited Oct 21, 2011) (comparing data from Chicago
public schools and Chicago charter schools and concluding “as with [high school] graduation
and college enrollment, results suggest that, for the average charter eighth-grader, attending a
charter [high school] may have positive effects on ACT scores”); Caroline M. Hoxby and
Jonah E. Rockoff, Findings from the City of Big Shoulders: Younger Students Learn More in
Charter Schools, 5 Educ Next 52, 58 (Fall 2005) (comparing Chicago’s public and charter
schools and concluding that “among students who enter in a typical grade, attending a charter
school improves reading and math scores by an amount that is both statistically and
substantively significant”).
7 Chester E. Finn Jr, Bruno V. Manno, and Gregg Vanourek, Charter Schools in Action:
Renewing Public Education 18 (Princeton 2000).
2012] Catholic Schools, Charter Schools, and Urban Neighborhoods 33
and decreased social cohesion, in Chicago neighborhoods.8 This
Article turns to two questions left unanswered in our previous
investigations. First, because we have focused exclusively on school
closures, we remain uncertain whether our results reflect the
beneficial effects of open Catholic schools rather than the negative
effects of school closures. Second, since we have thus far focused only
on Catholic schools, we cannot know whether other kinds of schools
generate similar positive externalities. In this Article, we begin to
answer these questions by comparing the effects of open Catholic and
charter schools on crime rates. Relying on police-beat-level data
provided in Chicago, we find that that police beats with open Catholic
schools have lower rates of serious crime than those without them, and
that open charter schools appear to have no statistically significant
effect on crime. All of these findings hold true even after we control
for numerous demographic variables that would tend to predict
neighborhood decline.
Our findings—that the presence of a Catholic school in a police
beat appears to suppress crime and the presence of a charter school
does not—are important for two related reasons. First, charter
schools are not only growing at an exponential rate but, as the
Catholic school sector contracts, they are coming to replace Catholic
schools as the schools of choice in urban neighborhoods. In many
cases (including fourteen schools in this study), charter schools also
are physically replacing Catholic schools by operating in closed
Catholic school buildings. Second, in education-reform debates,
charter schools frequently are cited as a means of capturing the
benefits of school choice without enlisting private schools through
voucher and tax-credit programs, which arguably threaten both to
drain public school resources and to undermine public values.9 Our
findings, in contrast, suggest that charter schools may be imperfect
substitutes for “complete” school choice. Charter schools may fill the
educational void left by Catholic schools’ disappearance from our
cities—a possibility about which we remain dubious—but, at least
thus far, they do not appear to replicate Catholic schools’ positive
community benefits. A more complete menu of school-choice
options might help preserve these benefits by stemming the tide of
Catholic school closures.
8 See Margaret F. Brinig and Nicole Stelle Garnett, Catholic Schools, Urban
Neighborhoods, and Education Reform, 85 Notre Dame L Rev 887, 953 (2010); Margaret F.
Brinig and Nicole Stelle Garnett, Catholic Schools and Broken Windows, 9 J Empirical Legal
Stud *49 (forthcoming 2012), online at http://ssrn.com/abstract=1564254 (visited Oct 21, 2011).
9 See David Brooks, The Quiet Revolution, NY Times A35 (Oct 23, 2009).
34 The University of Chicago Law Review [79:31
I. PREVIOUS INVESTIGATIONS
In our previous studies, we sought to measure the effects of
Catholic school closures on perceived disorder, perceived social
cohesion, and crime in Chicago neighborhoods. In our initial study,
we relied upon survey data collected for the Project on Human
Development in Chicago Neighborhoods (PHDCN) to measure the
effects of Catholic school closures on perceived disorder and
perceived social cohesion in Chicago neighborhoods.10 In 1994 and
1995, the PHDCN surveyed approximately four thousand Chicago
residents about perceived levels of neighborhood crime, disorder,
and social cohesion.11 After matching each of the 130 Catholic
elementary schools that closed in the city of Chicago between 1984
and 1994 to the PHDCN data, we estimated the effects of a Catholicschool
closure using two-stage least squares regression analysis, a
method that enabled us both to control for numerous demographic
variables and to employ variables predicting school closures
unrelated to demographics.12 Our analysis linked school closures to
neighborhood social cohesion and increased neighborhood
disorder.13
In our second study, we conducted a latent growth analysis of
effects of Catholic-school closures between 1990 and 1996 on the rate
of serious crime in police beats between 1999 and 2005.14 While crime
decreased across the city of Chicago during this period, our analysis
suggested that Catholic-school closures affected the slope of the
decline.15 That is, “crime decreased more slowly between 1999 and
2005 in police beats where Catholic schools closed between 1990 and
1996.”16 As in our initial study, we incorporated a variable—the parish
leadership characteristics that we describe briefly below—to disaggregate
school-closure decisions from neighborhood demographics.17
10 See Brinig and Garnett, 85 Notre Dame L Rev at 902 (cited in note 8).
11 See Robert J. Sampson and Stephen W. Raudenbush, Systematic Social Observation of
Public Spaces: A New Look at Disorder in Urban Neighborhoods, 105 Am J Soc 603, 619–20, 637
(1999).
12 See Brinig and Garnett, 85 Notre Dame L Rev at 923 (cited in note 8).
13 Id at 924–28.
14 Brinig and Garnett, 9 J Empirical Legal Stud at *1–6 (cited in note 8).
15 Id at *3.
16 Id.
17 Id at *13–17.
2012] Catholic Schools, Charter Schools, and Urban Neighborhoods 35
II. CATHOLIC SCHOOLS AND CHARTER SCHOOLS:
A BRIEF OVERVIEW
Before turning to the empirical questions at the heart of this
investigation, we provide a brief overview of the two distinct
institutional forms we study—Catholic schools and charter schools.
A. Urban Catholic Schools
Traditionally, almost all Catholic elementary schools (the
subject of our study) were “parochial.” That is, they were operated
by a Catholic parish led by a Catholic priest, known as the pastor,
who is the chief operating officer for all parish operations, including
the school.18 In the late nineteenth century, Catholic bishops,
responding to widespread nativism and Protestant indoctrination in
the public schools, began to demand that every parish build and
support a school and that all parish members enroll their children in
it.19 As a result, by the middle of the twentieth century, most major
American cities were densely blanketed with Catholic schools. As
political scientist Gerald Gamm has demonstrated, urban Catholics’
attachments to their parishes and schools fostered a strong
geographic “rootedness” that caused them to suburbanize later, and
to resist racial integration more strenuously, than other white urban
residents.20
By the late 1960s, however, shifting urban demographics and
labor-force realities began to threaten the viability of the parochial
school model, at least in urban areas.21 Historically, parochial schools
were entirely funded by the parish and staffed almost entirely by
religious sisters (nuns) who labored for little more than what one
commentator has called a “token wage.”22 In the 1960s, however,
religious vocations plummeted at the same time that Catholics
suburbanized en masse, causing parochial schools to experience
dramatic increases in labor costs just as collection revenues declined
precipitously.23 Gradually, schools built to educate working-class
Catholic children began to assume the role of educating poor, and
18 See Anthony S. Bryk, Valerie E. Lee, and Peter B. Holland, Catholic Schools and the
Common Good 148–65 (Harvard 1993).
19 Id at 23–33.
20 Gerald Gamm, Urban Exodus: Why the Jews Left Boston and the Catholics Stayed 129,
237–47 (Harvard 1999).
21 See John T. McGreevy, Parish Boundaries: The Catholic Encounter with Race in the
Twentieth-Century Urban North 236 (Chicago 1996).
22 Id at 10, 236.
23 Id at 234–40.
36 The University of Chicago Law Review [79:31
frequently non-Catholic, children.24 Dioceses were forced to take on
more of the financial burden of operating urban parish schools at the
same time they were obligated to build new schools to serve
suburbanizing Catholics. At a more retail level, some priests began
to view schools as an unnecessary burden, especially as the non-
Catholic student population increased.25 The urban parochial model
began to unravel, and dioceses began to close schools in large
numbers.26
Between 1984 and 2004, the Archdiocese of Chicago closed
130 elementary and 18 secondary schools.27 In some cases, several
schools closed in the same neighborhood—not surprisingly, given the
density of schools that historically served different ethnic
populations.28 Despite the many closures, the Archdiocese still
operates the largest nonpublic school system in the country, with 218
elementary schools and 40 high schools enrolling over 96,000
students.29 Most of the elementary schools continue to be operated by
parishes, although the Archdiocese retains supervisory authority
over them and substantially subsidizes many of them, either directly
or through a private philanthropic organization known as the Big
Shoulders Fund.30
B. Charter Schools
Although charter schools have roots in a number of older
reform ideas, they have existed in their current form for less than
two decades,31 and the first charter schools in Chicago opened in
1997.32 Charter schools are public-private hybrids.33 Charter schools
resemble public schools since they are open to all who wish to
attend, tuition free, and secular. Charter schools also are more
accountable than private schools and, arguably, even more than
24 Id at 241–42. See also Bryk, Lee, and Holland, Catholic Schools at 52 (cited in note 18).
25 See McGreevy, Parish Boundaries at 236–40 (cited in note 21).
26 See Brinig and Garnett, 85 Notre Dame L Rev at 892–903 (cited in note 8).
27 See Simons, Closed School History at *2 (cited in note 2).
28 See id.
29 See Office of Catholic Schools, Facts about the Catholic Schools in the Archdiocese of
Chicago (Archdiocese of Chicago 2009), online at http://schools.archchicago.org/public
/factsheet.shtm (visited Oct 21, 2011).
30 See Facts & Accomplishments (Big Shoulders Fund 2010), online at http://
www.bigshouldersfund.org/content/?s=477&s2=484&p=491&t=Facts-&-Accomplishments
(visited Oct 21, 2011).
31 See Finn, Manno, and Vanourek, Charter Schools in Action at 13–22 (cited in note 7).
32 See Office of New Schools, 2008–09 Charter and Contract Schools Performance
Report 1 (Chicago Public Schools 2009), online at http://www.cps.edu/NewSchools/Documents
/2008-2009_PerformanceReport.pdf (visited Oct 21, 2011).
33 Id at 5.
2012] Catholic Schools, Charter Schools, and Urban Neighborhoods 37
traditional public schools, because underperforming charter schools
are more likely to be closed.34 Charter schools also have attributes of
private schools. They are created by private entrepreneurial action—
the request of a private entity (the charter “operator”) for
permission to open a school from a governmental entity (the charter
“sponsor”). Like private schools, charter schools also enjoy
operational autonomy from local school officials (although the
precise extent of the autonomy depends upon state law).35 And, like
private schools, they are schools of choice—that is, parents select
them for their children much as they would a private school.36
While many charter schools focus on values or character
education, and some are structured around cultural themes with
religious overtones,37 an important legal feature distinguishing
charter and private schools is that charter schools are secular.38
Despite this restriction, a number of Catholic dioceses have, or are
considering, “converting” Catholic elementary schools to secular
charter schools rather than closing them.39 Since many states prohibit
charter schools from being operated by, or being affiliated with,
religious institutions, and a handful expressly prohibit the conversion
of all private schools to charter schools,40 dioceses must create or
contract with secular charter operators to operate the “converted”
schools.41 Although the Archdiocese of Chicago has not intentionally
34 See Andrew J. Rotherham and Richard Whitmire, Close Underperforming Charter
Schools, Reward Those That Work, US News & World Rep (June 17, 2009), online at
http://www.usnews.com/opinion/articles/2009/06/17/close-underperforming-charter-schoolsreward-
those-that-work (visited Oct 21, 2011).
35 See Finn, Manno, and Vanourek, Charter Schools in Action at 127–47 (cited in note 7).
36 Minnesota enacted the first charter school law in 1991. See id at 18–22.
37 For examples of such schools, see generally National Heritage Academies, Our
Approach (2011), online at http://www.nhaschools.com/About-Us/Pages/Our-Approach.aspx
(visited Oct 21, 2011); Great Hearts Academies, Great Hearts Academies, online at
http://greatheartsaz.org (visited Oct 12, 2011).
38 See Benjamin Siracusa Hillman, Note, Is There a Place for Religious Charter Schools?,
118 Yale L J 554, 560 (2008). See also Sarah Lemagie, ACLU Settles with State, School Sponsor:
TiZA, A Charter School Accused of Promoting Religion in Violation of the Constitution, Has
Not Reached an Agreement with the ACLU, Star Trib 1B (Feb 8, 2011); Abby Goodnough,
Hebrew Charter School Spurs Florida Church-State Dispute, NY Times A1 (Aug 24, 2007).
39 See, for example, God and Times Tables, Economist 38 (May 15, 2010); Catholic
Schools Get Final OK to Become Charters, Indianapolis Bus J Online (Apr 8, 2010), online at
http://www.ibj.com/catholic-schools-get-final-ok-to-become-charters/PARAMS/article/19166
(visited Oct 21, 2011); Javier C. Hernandez, City Tries New Tactic to Convert Catholic Schools
to Charter Schools, NY Times A22 (Apr 22, 2009); Bill Turque, 7 Catholic Schools in D.C. Set
to Become Charters: Funding Sources Are Still Unclear, Wash Post B01 (June 17, 2008).
40 See, for example, Hernandez, City Tries New Tactic, NY Times at A22 (cited in note 39).
41 See Andy Smarick, Catholic Schools Become Charter Schools: Lessons from the
Washington Experience 9–14 (Seton Education Partners 2009), online at http://
www.setonpartners.org/Seton_DC_Case_Study_FINAL.pdf (visited Oct 21, 2011); Dana
Brinson, Turning Loss into Renewal: Catholic Schools, Charter Schools, and the Miami
38 The University of Chicago Law Review [79:31
converted any of its schools to charter schools, a number of Chicago
charter schools do operate in buildings that formerly housed
Catholic schools, fourteen of which are included in our study.42
During the 2009–10 school year, there were 104 charter schools
(or, technically, 38 charters school operating on 104 campuses) in
Chicago.43 Twenty-eight of these schools opened during the period of
our study.44 These schools are institutionally diverse. They include
elementary schools, junior high schools, and secondary schools, as
well as nontraditional age groupings (for example, grades 6–12).45 At
least two are single-sex schools,46 and several have themed curricula.47
Charter schools enroll a higher proportion of African American
students (65 percent)48 than does the district as a whole (45 percent),49
and a smaller proportion of Hispanic, white, and Asian students.50
Chicago’s charter schools also enroll a slightly higher proportion of
low-income students (85.6 percent)51 than does the Chicago Public
Schools as a whole (83.3 percent).52 The Chicago Public Schools
reports that 63.4 percent of charter school students are “from the
neighborhood.”53
III. CHARTER SCHOOLS, CATHOLIC SCHOOLS, AND CRIME
In this Part, we turn to two questions raised, but unanswered, by
our previous findings. First, we seek to understand whether—as we
Experience 3 (Seton Education Partners 2010), online at http://publicimpact.com/publications
/Seton_Miami_Case_Study.pdf (visited Oct 21, 2011).
42 See Peter Meyer, Catholic Ethos, Public Education: How the Christian Brothers Came
to Start Two Charter Schools in Chicago, 11 Educ Next 40, 43–45 (Spring 2011).
43 Illinois Network of Charter Schools, Illinois Public Charter Schools: Profiles at 1 (cited
in note 4).
44 See id at 2–17.
45 See id.
46 See id at 15, 17.
47 See, for example, Illinois Network of Charter Schools, Illinois Public Charter Schools:
Profiles at 12, 15 (cited in note 4) (describing Prairie Crossing Charter School as “an
environmentally themed school” and Springfield Ball Charter School as having “a theme of
literacy and numeracy”).
48 See id at 1.
49 See Office of New Schools, 2009–2010 Charter and Contract School Performance
Report at *2 (cited in note 5).
50 Compare Illinois Network of Charter Schools, Illinois Public Charter Schools: Profiles
at 1 (cited in note 4), with Office of New Schools, 2009–2010 Charter and Contract School
Performance Report at *2 (cited in note 5).
51 See Julie Woestehoff and Monty Neill, Chicago School Reform: Lessons for the
Nation *39 (Parents United for Responsible Education 2007), online at http://pureparents.org
/data/files/Final report.pdf (visited Oct 21, 2011).
52 See Office of New Schools, 2009–2010 Charter and Contract School Performance
Report at *2 (cited in note 5).
53 See Office of New Schools, 2008–09 Charter and Contract Schools Performance Report
at 5 (cited in note 32).
2012] Catholic Schools, Charter Schools, and Urban Neighborhoods 39
strongly suspect—open Catholic schools suppress neighborhood
crime or, alternatively, whether the negative effects of Catholic
school closings result from the loss of a community institution. This
distinction is an important one. A finding that an open Catholic
school is associated with lower crime rates in a police beat would
support our suspicion that Catholic schools generate social capital. It
would also provide concrete evidence that Catholic schools behave
differently for neighborhoods than public schools, since other
scholars have demonstrated a link between open public schools and
increased crime.54 Alternatively, if our findings reflect loss effects we
might tend to suspect that that the losses of other kinds of
community institutions might also erode neighborhood social
controls. In order to test the effects of open Catholic schools on
crime rates, we use regression analysis to compare the rates of crime
in police beats with Catholic schools to those without them.
Second, we seek to begin to understand whether we have been
finding “Catholic school effects” rather than simply “school effects.”
Here, we add charter schools to our analysis, for several reasons.
Charter schools are imperfect proxies for public schools, especially in
Chicago, where many charter schools function as neighborhood
schools.55 In contrast to traditional public schools, moreover, charter
schools are not present in many police beats, making a comparison
between beats with and without schools possible. Moreover, charter
schools drive Catholic-school closures both because they compete
with Catholic schools,56 and because, as Archdiocesan officials
emphasized in our discussions, the revenue from leasing Catholic
school buildings to charter operators incentivizes some pastors to
lobby for school closures.57 Charter schools are frequently offered by
some as an alternative to school choice programs that might stem the
tide of Catholic-school closures. Finally, charter schools fill the
educational void left when Catholic schools close—and they also
frequently fill the physical space once occupied by closed Catholic
schools. For example, a spokesperson for the Archdiocese of Detroit
recently estimated to one of us that approximately 90 percent of the
54 See Dennis W. Roncek and Antoinette LoBosco, The Effect of High Schools on Crime
in Their Neighborhoods, 64 Soc Sci Q 598, 609–10 (1983). See also Dennis W. Roncek and
Donald Faggiani, High Schools and Crime: A Replication, 26 Sociological Q 491, 501 (1985).
55 See Finn, Manno, and Vanourek, Charter Schools in Action at 17 (cited in note 7).
56 See Samuel G. Freedman, Lessons from Catholic Schools for Public Educators, NY
Times A17 (May 1, 2010).
57 Interview with Sister Mary Paul McCaughey, Superintendent of Catholic Schools,
Archdiocese of Chicago (Mar 20, 2009) (“McCaughey interview”) (on file with authors).
40 The University of Chicago Law Review [79:31
Archdiocese’s closed Catholic-school buildings are currently
occupied by charter schools.58
Our analysis tends to confirm our suspicion that we are finding a
“Catholic school effect” on neighborhood health. We find that beats
with Catholic schools have consistently lower rates of serious crime
and, in contrast, that charter schools are not correlated in a
statistically significant way with crime rates in either direction.
A. Data
We rely on multiple sources of data. The Archdiocese of Chicago
provided information on closed and open Catholic schools, including
their location, name, and parish affiliation. For detailed information
about parish and school leaders, we relied upon The Official Catholic
Directory.59 Information on clergy abuse came from an official
Archdiocesan report,60 from a “victims’ rights” organization that
collects accusations (including unsubstantiated ones),61 and from
newspaper accounts. Data on charter schools came from the Chicago
Public Schools’ Office of New Schools and from the Illinois Network
of Charter Schools.62 To parallel our information on Catholic schools,
we restricted our analysis to charter elementary schools located in the
city of Chicago proper. We excluded high schools as well as
freestanding middle schools. Demographic information comes from
the 2000 census, and the Chicago Police Department provided data on
the incidence of six major crimes (aggravated assault, aggravated
battery, murder, burglary, robbery, and aggravated sexual assault) at
the police-beat level from 1999–2005.63
58 See also Marisa Schultz, DPS Schools Get New Life as Charters: But Critics Say Their
Lure Costs District Money, Detroit News A6 (Oct 2, 2010); Michelle Martin, Charter Schools
Not “Catholic,” Catholic New World (May 27, 2001), online at http://www.catholic
newworld.com /archive/cnw2001/052701/charter_052701.html (visited Oct 21, 2011).
59 Official Catholic Directory 1999–2005 (P.J. Kenedy & Sons 2005) (providing yearly
updated archdiocesan entries approved by each archdiocese).
60 Archdiocesan Priests with Substantiated Allegations of Sexual Misconduct with Minors
(Archdiocese of Chicago 2011), online at http://www.archchicago.org/c_s_abuse/report_032006
/list.pdf (visited Oct 21, 2011).
61 BishopAccountability.org: Documenting the Abuse Crisis in the Roman Catholic
Church (2011), online at http://www.bishop-accountability.org/Who_We_Are/ (visited Oct 21,
2011).
62 See Office of New Schools (Chicago Public Schools 2011), online at http://www.cps.edu
/newschools/Pages/ONS.aspx (visited Oct 21, 2011); Illinois Network of Charter Schools,
Alphabetical Directory (2011), online at http://incschools.org/charters/find_a_charter_school
/full_list (visited Oct 21, 2011).
63 These are all serious (Part I) crimes collected yearly by the Department of Justice and
published as Uniform Crime Reports. See National Atlas of the United States, Summary of the
Uniform Crime Reporting Program (2011), online at http://www.nationalatlas.gov
/articles/people /a_crimereport.html (visited Oct 21, 2011). See also Federal Bureau of
2012] Catholic Schools, Charter Schools, and Urban Neighborhoods 41
B. Explaining School Closures and Openings
We recognize the obvious endogeneity problem we face—that
is, the same factors that predict the location of charter and Catholic
schools also might affect crime rates. In order to separate
demographics and school locations, we sought in our previous
studies to identify variables predicting Catholic-school closures that
were unrelated to neighborhood demographics (or other things
closely associated with crime).64 To do so, we began by asking
Archdiocesan officials what factors drove school closures. While
school-closure decisions are complex, the superintendent of Catholic
Schools, Sister Mary Paul McCaughey, emphasized that, for
struggling schools, the most important factor predicting whether a
school closed was the support of the pastor.65 As she explained, while
school-closure decisions are centralized, the Archdiocese tends to
defer to the pastor’s wishes.66 Pastors who wish to “unload” a school
often get their way,67 and pastors who rally to the school’s defense
often are given a second chance to save it. We therefore directed our
attention to the pastors of the parishes with elementary schools and
found, as she predicted, that certain parish leadership characteristics
were strongly connected with school closings—more so than
neighborhood demographics.68
We do not employ these variables here, however, since we
cannot identify similar variables explaining charter schools’
locations. We assume that charter schools open for many reasons
and that some charter-school operators intentionally locate in poor
urban neighborhoods where crime is more prevalent. To avoid
comparing apples to oranges, we chose not to employ the previously
identified school-closure variables (although they remain predictive
of closures here). This choice limits the strength of our findings,
making it impossible to demonstrate causation—although we
Investigation, Uniform Crime Reports (Department of Justice 2011), online at http://
www.fbi.gov/about-us/cjis/ucr/ucr (visited Oct 21, 2011).
64 See Brinig and Garnett, 9 J Empirical Legal Stud at *13–15 (cited in note 8).
65 See id at *12–13.
66 See id.
67 Id at *13.
68 These include irregularity in parish leadership, meaning that the pastor had been
replaced with a temporary administrator or that a priest at the parish had been accused of sex
abuse. The other characteristic predicting closings was the pastor’s age, although this factor
just missed statistical significance for later closures. Since neither parish “irregularity” nor age
would seemingly have anything to do with demographics or neighborhood crime, we were
comfortable concluding that the parish leadership variables are an appropriate way to address
the endogeneity problem. See Brinig and Garnett, 9 J Empirical Legal Stud at *11–16 (cited in
note 8); Brinig and Garnett, 85 Notre Dame L Rev at 912–20 (cited in note 8).
42 The University of Chicago Law Review [79:31
emphasize our regression analysis does control for neighborhood
demographics.
C. Catholic- and Charter-School Effects on Crime
In controlling for demographics, we include the same
characteristics found by a host of other researchers to explain crime
in Chicago.69 We then matched schools, census tracts, and police
beats using ArcGIS—a mapping program.70 As Table 1 indicates, in
2004, there were Catholic schools in eighty-four distinct police beats
and charter elementary schools in twenty-eight distinct police beats.
Fourteen charter schools were located in closed Catholic schools.
69 See, for example, Andrew V. Papachristos, Tracey L. Meares, and Jeffrey Fagan,
Attention Felons: Evaluating Project Safe Neighborhoods in Chicago, 4 J Empirical Legal
Stud 223, 240 table 1 (2007).
70 The number of census tracts in each beat varied from three to twenty-three, with an
average of more than ten per beat. Visual inspection of these tract-beat matches revealed that
it was nearly impossible to choose a majority or typical tract for many beats, so we included
them all to eliminate subjectivity. One beat (3100) had no people living in it, so the data was
simply excluded, leaving us with 2,902 tract/beat observations for which there were both crime
and census information. Beat 1611 had two Catholic schools but was entered only once for
each tract. Beat 922 had two charters located in one closed school, but again it was entered
only once for each tract.
2012] Catholic Schools, Charter Schools, and Urban Neighborhoods 43
TABLE 1. DESCRIPTIVE STATISTICS
N Minimum Maximum Mean
Standard
Deviation
Charter school (28 distinct
police beats/tracts as
of 2004)
2902 0.00 1.00 0.0096 0.09777
Charter school located in
closed Catholic School
(14 distinct police
beats/tracts as of 2004)
2902 0.00 1.00 0.0048 0.06930
Open Catholic school
(as of 2004, 84 distinct
police beats/tracts)
2902 0.00 1.00 0.0283 0.16573
Total population (2000) 2900 0 15359 3597.01 2671.046
Share of population that is
white (2000)
2900 0.0000 1.0000 0.387388 0.3447462
Share of population that is
nonwhite (2000)
2900 0.0000 1.0000 0.612612 0.3447462
Share of population that is
foreign born (2000)
2900 0.0000 0.7388 0.174533 0.1760852
Share living in same
household 5 years (2000)
2900 0.0000 1.0000 0.537996 0.1638469
Percent in labor force (2000) 2900 0.0000 1.0000 0.590173 0.1396101
Percent below poverty line 2900 0.0000 0.9269 0.219373 0.1599469
Percent ages 15–25 2900 0.0000 0.7047 0.150064 0.0668530
Percent living in rental
housing (2000)
2900 0.0000 1.0000 0.578107 0.2260260
Percent female headed
households (2000)
2900 0.0000 1.0000 0.429929 0.2108200
Percent linguistically isolated
(2000)
2900 0.0000 .6667 0.092068 0.1149682
Median income (2000) ($) 2893 0 127221 37142.81 17830.074
Percent households on public
assistance (2000)
2900 0.0000 1.0000 0.094825 0.1099266
Share of population that is
black (2000)
2898 0.0000 1.0000 0.438765 0.4357780
Share of population that is
Hispanic (2000)
2898 0.0000 1.0000 0.221450 0.2837054
Total crime 1999 2815 0.00 599.00 286.5805 106.73678
Total crime 2000 2829 0.00 546.00 273.8996 99.65107
Total crime 2001 2829 1.00 543.00 260.1467 101.41132
Total crime 2002 2829 40.00 555.00 235.3241 94.63097
Total crime 2003 2829 36.00 568.00 221.8271 93.57419
Total crime 2004 2829 24.00 550.00 221.7236 95.05772
44 The University of Chicago Law Review [79:31
N Minimum Maximum Mean
Standard
Deviation
Total crime 2005 2829 33.00 569.00 253.7656 99.82036
Crime rate 1999 2785 0.00 2685.71 26.2389 110.71567
Crime rate 2000 2799 0.00 3100.00 25.3433 112.12574
Crime rate 2001 2799 0.00 1892.86 18.1555 81.02117
Crime rate 2002 2799 0.28 1671.43 17.1429 71.25692
Crime rate 2003 2799 0.36 1514.29 16.1274 68.99457
Crime rate 2004 2799 0.32 1992.86 14.9863 66.69850
Crime rate 2005 2799 0.19 1864.29 14.9499 66.25297
Crime rate (1999–2005) 2750 2.84 15500.00 152.6250 656.99935
Natural logarithm of crime
rate (1999–2005)
2750 1.05 9.65 4.0445 1.10275
Valid N (listwise) 2743
Prior work indicates that the relationship between crime rates71
and other characteristics is best represented by taking the natural
logarithm of the crime rates.72 For some equations, the demographic
variables were reduced to three factors.73 In others, they were each
entered separately. While crime was declining in Chicago between
1999 and 2005, the crime rate, controlling for demographic factors,
was lower in each year in those beats with Catholic schools than in
those that did not include them.
71 That is, crime divided by the census-tract population. To keep the logarithms positive,
this quotient was multiplied by one hundred.
72 See, for example, Brinig and Garnett, 9 J Empirical Legal Stud at *26 (cited in note 8);
Papachristos, Meares, and Fagan, 4 J Empirical Legal Stud at 245–46 (cited in note 69) (using
“[t]he log of the beat-level homicide rate” in order to “improve model fit and account for
nonlinearity” in a study attempting to evaluate the impact of “Project Safe Neighborhood”
initiatives on neighborhood-level crime rates in Chicago); Sampson and Raudenbush, 105 Am
J Soc at 621 (cited in note 11).
73 The technique, called principal component analysis, uses regression results to reduce a
large set of possibly correlated variables into a subset of uncorrelated variables. Because there
are fewer variables, the coefficients for the different types of schools are larger, though
statistical significance and direction do not change. The same technique has been used on
Chicago crime data in earlier work. See Sampson and Raudenbush, 105 Am J Soc at 621–23 &
n 19 (cited in note 11).
2012] Catholic Schools, Charter Schools, and Urban Neighborhoods 45
FIGURE 1. COMPARISON OF POLICE BEATS WITH AND WITHOUT
OPEN CATHOLIC SCHOOLS
To produce Figure 1, we first separated the data into the police
beats where there were and were not open Catholic schools. Then
we used regression analysis to predict the average crime rate for each
of the seven years. In effect, we held constant the presence or the
absence of a charter school (located or not in a closed Catholic
school), plus the computed composite-socioeconomic Table 1 details.
The coefficients are displayed as Model 1 in Table 2; the mean
adjusted predicted values for each year make up the points in
Figure 1. The other two models of Table 2 are very similar, and
graphing them would produce nearly identical results. Model 2
displays regression coefficients, standard errors, and statistical
significance for an equation where instead of grouping the
demographic characteristics, the characteristics are broken out into
original census data for each police beat and tract combination.
Figure 1 used the variables from Model 1 from Table 2 below for
each of the seven years of crime rate data, comparing results for
cases in which there was and was not an open Catholic elementary
school. The difference between Model 2 and Model 3 is that the
former uses individual race characteristics while the latter groups
them together.
TABLE 2. REGRESSION RESULTS: CRIME RATES AND
.0000
.5000
1.0000
1.5000
2.0000
2.5000
3.0000
1999
2000
2001
2002
2003
2004
2005
Adjusted Predicted Mean Value of Logged
Crime Rate
Year
Open Catholic
School
No Open Catholic
School
46 The University of Chicago Law Review [79:31
PRIVATE SCHOOLS
Model 1 Model 2 Model 3
Beta
(Standard error)
Beta
(Standard error)
Beta
(Standard error)
Constant 4.036
(.028)***
5.870
(.164)***
5.750
(.167)***
Open Catholic
school
-0.264
(.109)*
-0.120
(.068)*
-0.160
(.069)*
Charter school 0.154
(.210)
-0.015
(.130)
0.026
(.133)
Charter in
Catholic
0.045
(.282)
0.140
(.175)
0.181
(.179)
PCA1
(deprivation)
0.470
(.018)***
--- ---
PCA2
(immigration)
-0.162
(.018)***
--- ---
PCA3
(stability)
-0.220
(.018)***
--- ---
Total
population
2000
--- 0.000
(.000)***
0.000
(.000)***
Share nonwhite
2000
--- --- 0.926
(.060)***
Share foreign
born 2000
--- -0.397
(.161)*
-0.395
(.147)**
Same
household
2000
--- -0.399
(.108)***
-0.198
(.109)
Percent in labor
force
--- -0.709
(.157)***
-0.624
(.161)***
Percent below
poverty line
--- -1.163
(.144)***
-1.064
(.147)***
Percent renter --- 0.202
(.087)**
0.183
(.088)*
Percent female
head 2000
--- -1.186
(.119)***
-1.143
(.119)***
Percent
linguistically
isolated 2000
--- -0.301
(.218)
0.218
(.214)
Median income
2000
--- -5.233E-6
(.000)***
-5.855E-6
(.000)***
2012] Catholic Schools, Charter Schools, and Urban Neighborhoods 47
Model 1 Model 2 Model 3
Beta
(Standard error)
Beta
(Standard error)
Beta
(Standard error)
Percent
households
with public
assistance
2000
--- 1.875
(.176)***
1.696
(.180)***
Percent Black
2000
--- 0.987
(.064)***
---
Percent
Hispanic 2000
--- 0.978
(.067)***
---
R2 (adjusted) .259 .714 .702
F 240.812 428.332 .432
Note: *** signifies p < .001, ** signifies p < .01, and * signifies p < .05.
In other words, regardless of how we account for demographic
variables that generally predict crime, an open Catholic elementary
school in a beat is associated with a statistically significant decrease
in the rate of crime.74 Although the percentage difference varied by
year, the crime rate in police beats with Catholic schools was, on
average, at least 33 percent lower than police beats without them.
Charter schools appear to have no statistically significant effect on
crime in either direction, although, in a few years, regressions for
individual crimes suggest a statistically significant link between
charter schools and elevated rates of aggravated assault and
aggravated battery. To the extent we can note anything about the
charter schools operating in closed Catholic schools, the direction of
the coefficients is not encouraging (that is, crime seems to increase).
74 In the simplest model of all, looking at the correlation between open Catholic
elementary schools and the logged crime rate, the coefficient is −.086 at p < .001. A model that
simply considers race and income generates an adjusted R2 of .141 (F=119.934), with the
coefficients as follows. The effect of the Catholic school being open, standardized, is almost
exactly the same as the effect of income.
Variable Beta Standard Error Beta (standardized)
Constant 3.603*** .100 ---
Open Catholic
school
(as of 2004)
-0.371*** .113 -0.058
Percent black in
census tract
(2000)
0.987*** .074 0.399
Percent Hispanic in
census tract
(2000)
0.591*** .097 0.156
Median income
(2000)
-3.749E-6*** .000 -0.059
Dependent variable: log crime rate; p < .001 for all coefficients.
48 The University of Chicago Law Review [79:31
We cannot at this point say that opening a new Catholic school
(or re-opening a closed one) would decrease crime. We also cannot
say whether individual charter schools—including, perhaps
especially, charter schools that mimic the educational program of
Catholic schools—suppress crime, or whether charter schools will, as
they become established, have the same positive effects as Catholic
schools. We also cannot say whether “converting” Catholic schools
to charter schools will maintain Catholic schools’ positive effects—
although our findings here suggest that simply operating a
nonsectarian charter in a closed Catholic school does not.
IV. EXPLAINING CATHOLIC SCHOOLS’ POSITIVE EXTERNALITIES
Our findings suggesting that charter schools do not suppress
crime are not inconsistent with previous studies linking public
schools with disorder and crime.75 Our research suggests, however,
that urban Catholic elementary schools have the opposite effect—
that is, that they suppress disorder and serious crime. We can, at this
point, only speculate about possible explanations for Catholic
schools’ positive externalities.
A. The “Night Watchman” Explanation
One study linking public schools and crime found that public
elementary schools appeared to generate more crime than public high
schools. The authors speculated that unsupervised playgrounds may
serve as recreational hangouts for teenagers or staging areas for
illicit activities.76 Perhaps, therefore, Catholic school facilities simply
are more secure than charter schools. Or perhaps Catholic schools
are more likely to generate what Jane Jacobs famously termed “eyes
upon the street.”77 In most parishes, for example, the pastor lives onsite
and may serve a “night watchman” function.78 Catholic schools
75 See, for example, Lisa Broidy, Dale Willits, and Kristine Denman, Schools and
Neighborhood Crime *6–12 (Justice Research Statistics Association 2009), online at
http://www.jrsa.org/ibrrc/background-status/New_Mexico/Schools_Crime.pdf (visited Oct 21,
2011); Caterina Gouvis Roman, Schools as Generators of Crime: Routine Activities and the
Sociology of Place *111–18 (National Criminal Justice Reference Service 2002); Roncek and
Faggiani, 26 Sociological Q at 501 (cited in note 54); Roncek and LoBosco, 64 Soc Sci Q at 609–10
(cited in note 54).
76 See Paula M. Kautt and Dennis W. Roncek, Schools as Criminal “Hot Spots”:
Primary, Secondary, and Beyond, 32 Crim Just Rev 339, 349–53 (2007).
77 See Jane Jacobs, The Death and Life of Great American Cities 34–35 (Random
House 1961).
78 In our study, however, the Archdiocese closed only a handful of parishes, so the pastor
remained on-site even after the school was shuttered.
2012] Catholic Schools, Charter Schools, and Urban Neighborhoods 49
also might be more likely to sponsor community activities during
after-school hours that draw adults into the neighborhood.79
B. The Student-Body Explanation
Our results also might reflect the fact that Catholic and charter
schools enroll different types of students. A greater degree of
institutional diversity existed among charter schools than Catholic
schools during the period of our study. Most (but not all) of the
Catholic schools in our study enrolled grades kindergarten (or
prekindergarten) through eight. Most of the charter schools did not;
some extended through fifth or sixth grades; and others were limited
to the middle school years (although we excluded these from our
analysis). The educational-psychology literature on “school
transitions” suggests that students perform better—in terms of
behavior, academic achievement, and self-esteem—in K–8 schools.80 If
older students are more likely to generate disorder—and researchers
have linked the greater incidence of crime near public middle and high
schools with the presence of large numbers of adolescents81—then
Catholic schools’ practice of combining elementary- and middleschool
students may generate positive neighborhood externalities.
Moreover, Catholic schools’ control over student-body
composition is frequently cited as contributing to their relative
educational success, as is the fact that better-educated, highly
motivated parents are more likely to choose Catholic schools for
their children.82 Both factors may help explain Catholic and charter
schools’ divergent neighborhood effects. Charter schools exercise far
less enrollment discretion than Catholic schools. They generally must
conduct a lottery for admissions, although they may give priority to
students residing within their attendance boundary (if one is
designated).83 Charter schools may also find it more difficult to expel
disruptive students who may “act out” both inside and outside the
classroom setting. Moreover, although Catholic school tuition is very
low relative to that of other types of private schools,84 a decision to
79 Charles W. Dahm, Parish Ministry in a Hispanic Community 238–51 (Paulist 2004).
80 See Jonah E. Rockoff and Benjamin B. Lockwood, Stuck in the Middle: How and Why
Middle Schools Harm Student Achievement, 10 Educ Next 68, 69–75 (Fall 2010).
81 See, for example, Roncek and LoBosco, 64 Soc Sci Q at 601 (cited in note 54).
82 See Bryk, Lee, and Holland, Catholic Schools at 16, 46–54 (cited in note 18).
83 See, for example, Office of New Schools, Lottery Guidelines for Charter Schools *1
(Chicago Public Schools 2011), online at http://cps.edu/NewSchools/Documents/Lottery
GuidelinesForCharterSchools.pdf (visited Oct 21, 2011).
84 In 2007–08, the average tuition at a Catholic elementary school was $4,944; the
average tuition at a nonsectarian private elementary school was $15,945. See Council for
50 The University of Chicago Law Review [79:31
send a child to a Catholic school signals a threshold level of parental
motivation—and motivated parents may be better able to control
their children’s behavior before and after school. Catholic schools
also frequently place demands on parents that public schools do not
or even cannot: many either require parents to volunteer in the
school or provide parents with the option of volunteering in order to
reduce tuition burdens.85 These requirements may generate a stable
flow of responsible adults in the neighborhood who help keep
disorder and crime in check.
That said, it is important not to overstate the explanatory value
of these factors. For example, enrollment in a charter school also
signals parental motivation. A parent selecting a charter school must
opt out of the traditional public school system and choose among a
range of alternatives.86 Furthermore, most urban Catholic schools
provide significant financial assistance, which enables them to enroll
students from disadvantaged backgrounds. Indeed, the available
evidence suggests that the educational benefits of Catholic schools
are greatest for the poorest, most disadvantaged students.87
C. The Neighborhood-Network Explanation
A third explanation is suggested by William Fischel’s defense of
local public schools.88 Fischel argues that parent networks at
neighborhood public schools enable “community-specific social
capital.”89 As Fischel observes, “My approach to social capital
formation simply requires that parents get to know other parents. . . .
[A]nd sending your child to a local school does that more effectively
than any other means.”90 As he acknowledges, however, the
neighborhood-network benefits of public schools likely are reduced in
major cities, where intradistrict public school choice is commonplace.
Indeed, given the prevalence of public school choice in Chicago—
where more than one-third of all public school elementary students
American Private Education, Facts and Studies (2011), online at
http://www.capenet.org/facts.html (visited Oct 21, 2011).
85 See Bryk, Lee, and Holland, Catholic Schools at 306–08 (cited in note 18).
86 Jack Buckley and Mark Schneider, School Choices, Parental Information, and Tiebout
Sorting: Evidence from Washington, DC, in William A. Fischel, ed, The Tiebout Model at Fifty
101, 104 (Lincoln Institute of Land Policy 2006).
87 See Greeley, Catholic High Schools at 107–08 (cited in note 6); Coleman, Hoffer, and
Kilgore, High School Achievement at 143–46 (cited in note 6).
88 See William A. Fischel, The Homevoter Hypothesis: How Home Values Influence
Local Government Taxation, School Finance, and Land-Use Policies 154–55 (Harvard 2001).
89 William A. Fischel, Why Voters Veto Vouchers: Public Schools and Community-
Specific Social Capital, 7 Econ Gov 109, 112–17 (2006). See also Fischel, The Homevoter
Hypothesis at 142–43, 154–55 (cited in note 88).
90 Fischel, 7 Econ Gov at 116 (cited in note 89).
2012] Catholic Schools, Charter Schools, and Urban Neighborhoods 51
attend a school outside their geographic attendance boundaries—
students attending a Catholic school may be more likely to live in the
surrounding neighborhood than public school students, and Catholic
schools may be more likely to generate local social capital.91
That said, although we suspect that many Catholic schools are
neighborhood schools, most of the city’s charter schools are as well.
Illinois law authorizes the designation of an attendance boundary for
charter schools and permits charter schools to give priority to students
residing within their boundaries.92 During the 2008–09 school year,
Chicago charter schools, on average, drew approximately 63 percent
of their students from the surrounding neighborhood, although the
percentage of neighborhood students ranged from a low of 6.7 percent
to a high of 100 percent.93 Despite this fact, however, charter schools
do not appear to serve the same social-capital-generation function as
their Catholic school counterparts—or, if they do, the social capital
does not translate into reduced crime rates.
D. The Longevity Explanation
All of the Catholic schools that remained open during the
period of our study had been open since the 1930s (though some of
them had received children originally attending other schools closed
since 1984). In contrast, none of the charter schools opened before
1997.94 Over time, as charter schools become more integrated into
neighborhoods, they also may produce similar effects. It is also
possible that Catholic schools, by virtue of their longevity in a
community, will continue to produce positive effects even if they are
“converted” to charter schools. We simply cannot speculate, based
upon our data, about either possibility.
E. The “Last Vestige of Civilization” Explanation
Our results also might reflect the unfortunate reality that, in
some neighborhoods, a Catholic school was one of the last remaining
functional community institutions. As a Catholic bishop who served
as a priest in the Archdiocese of Chicago (and who attended one of
the closed schools in our study) told one of us privately, in some
neighborhoods, a Catholic school was the last vestige of civil
91 See Julie Berry Cullen and Brian A. Jacob, Is Gaining Access to Selective Elementary
Schools Gaining Ground? Evidence from Randomized Lotteries, in Jonathan Gruber, ed, The
Problems of Disadvantaged Youth: An Economic Perspective 43, 51–52 (Chicago 2009).
92 See 105 ILCS § 5/27A-4(d).
93 See Office of New Schools, 2008–09 Charter and Contract Schools Performance Report
at 5 (cited in note 32).
94 Id at 1.
52 The University of Chicago Law Review [79:31
society.95 It is hardly surprising, he remarked, that when the school
disappeared, the neighborhood rapidly declined.96
F. The “Catholic School Effect” Explanation
This fact leaves us to wonder whether our results are suggestive
of another possibility—namely, that what goes on inside a school
does in fact affect what happens outside it. That is, we ask whether
the human-capital- and social-capital-generating functions of a
school intersect. According to James Coleman’s classic formulation,
social capital “inheres in the structure of relations between actors
and among actors,” and institutions that foster these relationships
are incubators of social capital.97 Coleman used schools to illustrate
this conception of social capital, arguing that successful schools
tended to be distinguished by parents’ connections to their children’s
school and to the parents of their children’s peers.98 These
connections, he reasoned, “closed the loop” between schools,
teachers, and parents, thus guaranteeing the enforcement of
appropriate norms.99 Coleman further argued that these kinds of
connections—and the norm-enforcement authority that they
enabled—explained Catholic high schools’ extremely low drop-out
rates.100 Perhaps they also generate positive externalities beyond the
classroom walls. For example, in their influential book, Catholic
Schools and the Common Good, Anthony Bryk and his colleagues
linked Catholic high schools’ educational successes to the fact that
these schools were intentional communities, with high levels of trust
between students, parents, teachers, and administrators.101 In more
recent work, Bryk has argued that neighborhood factors, including
the level of collective efficacy and social capital in a community, are
critical inputs to urban public schools’ success (or failure).102 Bryk’s
work suggests that there may be significant feedback effects between
what goes on in a school and what occurs in the surrounding
community.103
95 Confidential interview with Catholic bishop who formerly served as a priest from the
Archdiocese of Chicago (“Priest Interview”) (on file with author).
96 Id.
97 James S. Coleman, Social Capital in the Creation of Human Capital, 94 Am J
Sociology S95, S98 (1998).
98 See id at S105–08.
99 See id.
100 See id at S114–15.
101 See Bryk, Lee, and Holland, Catholic Schools at 307–08, 313–14 (cited in note 18).
102 See Anthony S. Bryk, et al, Organizing Schools for Improvement: Lessons from
Chicago 177–78 (Chicago 2010).
103 See Bryk, Lee, and Holland, Catholic Schools at 282–85 (cited in note 18); Bryk, et al,
Organizing Schools at 196 (cited in note 102).
2012] Catholic Schools, Charter Schools, and Urban Neighborhoods 53
That said, many charter schools employ educational strategies
that closely approximate the Catholic school formula, including a
highly structured school day, traditional curriculum, high levels of
parental involvement, and an emphasis on building an educational
community between the various school stakeholders.104 Since our
analysis does not distinguish between different charter schools’
educational strategies, we cannot say whether schools employing this
formula positively impact neighborhoods in the way that our study
suggests Catholic schools do—or whether they might come to do so
over time.
.
Margaret F. Brinig & Nicole Stelle Garnett
INTRODUCTION
This Article addresses previously unstudied implications of two
dramatic shifts in the American educational landscape. The first shift
is the rapid disappearance of urban Catholic schools. More than
1,600 Catholic elementary and secondary schools, most of them
located in urban neighborhoods, have closed during the last two
decades.1 The Archdiocese of Chicago alone (the subject of our
study) has closed 148 schools since 1984.2 Since the economic and
demographic realities underlying urban Catholic school closures
persist, this trend likely will continue and even accelerate in coming
years. The second shift is the rise of charter schools. In 2010 more
than 1.7 million children were enrolled in 5,400 charter schools in the
United States.3 During the 2009–10 school year, there were
† Fritz Duda Family Chair in Law, University of Notre Dame Law School.
†† Professor of Law, University of Notre Dame Law School.
We are particularly indebted to Sister Mary Paul McCaughey, OP, Superintendent of
Catholic Schools, Archdiocese of Chicago, for her invaluable insights. We received many
helpful comments at the symposium on Understanding Education in the United States: Its
Legal and Social Implications, held at the University of Chicago Law School on June 17 and
18, 2011. Alison Curran, Peter Reed, and Michael Wilde provided excellent research
assistance.
1 See Richard W. Garnett, Treasure A.C.E., Natl Rev Online (Sept 10, 2008), online at
http://www.nationalreview.com/articles/225595/treasure-c-e/richard-w-garnett (visited Oct 21,
2011). See also Peter Meyer, Can Catholic Schools Be Saved?, 7 Educ Next 12, 16 (Spring 2007);
Sol Stern, Save the Catholic Schools!, 17 City J 74, 74–76 (Spring 2007); Mary Ann Zehr, Catholic
Schools’ Mission to Serve Needy Children Jeopardized by Closings, 26 Educ Wk 16–17 (Mar 8,
2005).
2 Paul Simons, Closed School History: 1984–2004 *2 (Office of Catholic Schools
Archdiocese of Chicago 2004), online at http://www.illinoisloop.org/cath_closed_school_84
_04.pdf (visited Oct 21, 2011).
3 All about Charter Schools: Quick Facts (Center for Education Reform 2011), online at
http://www.edreform.com/issues/choice-charter-schools/facts/ (visited Oct 21, 2011).
32 The University of Chicago Law Review [79:31
104 charter schools in the city of Chicago,4 24 of which opened during
the period of our study.5
Although we are intrigued by the questions raised by the
extensive literature on Catholic and charter schools’ strengths as
educational institutions,6 we do not address them here. Instead, we
raise new questions about how Catholic and charter schools function
as community institutions. These questions are important ones.
Catholic schools are vanishing from the urban neighborhoods where
they have operated for decades—in some cases, for over a century—
and are being replaced by educational institutions that did not exist
anywhere in the United States two decades ago.7 Yet virtually
nothing is known about the impact the transition will have on urban
neighborhoods, many of which already struggle with disorder, crime,
and poverty.
This is the third in a series of papers exploring the effects of
Catholic school closures on urban neighborhoods. In previous studies,
we linked Catholic school closures to increased disorder and crime,
4 Illinois Network of Charter Schools, Illinois Public Charter Schools: Profiles 1 (2010),
online at http://incschools.org/docs/INCS_SchoolProfiles2010_Final.pdf (visited Oct 21, 2011).
5 See Office of New Schools, 2009–2010 Charter and Contract Schools Performance
Report (Chicago Public Schools 2010), online at http://www.cps.edu/NewSchools/Documents
/2009-2010_PerformanceReport.pdf (visited Oct 21, 2011).
6 Numerous scholars have demonstrated that Catholic schools tend to outperform their
public counterparts, especially at the challenging task of educating underprivileged minority
students. See, for example, Andrew M. Greeley, Catholic High Schools and Minority
Students 108 (Transaction Books 1982); James S. Coleman, Thomas Hoffer, and Sally Kilgore,
High School Achievement: Public, Catholic, and Private Schools Compared 143–46 (Basic
Books 1982). Scholars are divided about whether charter schools outperform traditional public
schools. See Julian R. Betts and Y. Emily Tang, Value-Added and Experimental Studies of the
Effect of Charter Schools on Student Achievement: A Literature Review 26 (Washington Bothell
2008), online at http://www.crpe.org/cs/crpe/download/csr_files/pub_ncsrp_bettstang_dec08.pdf
(visited Oct 21, 2011); Multiple Choice: Charter School Performance in 16 States 1 (Stanford
June 2009), online at http://credo.stanford.edu/reports/MULTIPLE_CHOICE_CREDO.pdf
(visited Dec 31, 2011). The available evidence suggests charter schools in Chicago do
outperform Chicago public schools. See Achievement and Attainment in Chicago Charter
Schools 21 (RAND 2009), online at http://www.rand.org/content/dam/rand/pubs/technical
_reports/2009/RAND_TR585-1.pdf (visited Oct 21, 2011) (comparing data from Chicago
public schools and Chicago charter schools and concluding “as with [high school] graduation
and college enrollment, results suggest that, for the average charter eighth-grader, attending a
charter [high school] may have positive effects on ACT scores”); Caroline M. Hoxby and
Jonah E. Rockoff, Findings from the City of Big Shoulders: Younger Students Learn More in
Charter Schools, 5 Educ Next 52, 58 (Fall 2005) (comparing Chicago’s public and charter
schools and concluding that “among students who enter in a typical grade, attending a charter
school improves reading and math scores by an amount that is both statistically and
substantively significant”).
7 Chester E. Finn Jr, Bruno V. Manno, and Gregg Vanourek, Charter Schools in Action:
Renewing Public Education 18 (Princeton 2000).
2012] Catholic Schools, Charter Schools, and Urban Neighborhoods 33
and decreased social cohesion, in Chicago neighborhoods.8 This
Article turns to two questions left unanswered in our previous
investigations. First, because we have focused exclusively on school
closures, we remain uncertain whether our results reflect the
beneficial effects of open Catholic schools rather than the negative
effects of school closures. Second, since we have thus far focused only
on Catholic schools, we cannot know whether other kinds of schools
generate similar positive externalities. In this Article, we begin to
answer these questions by comparing the effects of open Catholic and
charter schools on crime rates. Relying on police-beat-level data
provided in Chicago, we find that that police beats with open Catholic
schools have lower rates of serious crime than those without them, and
that open charter schools appear to have no statistically significant
effect on crime. All of these findings hold true even after we control
for numerous demographic variables that would tend to predict
neighborhood decline.
Our findings—that the presence of a Catholic school in a police
beat appears to suppress crime and the presence of a charter school
does not—are important for two related reasons. First, charter
schools are not only growing at an exponential rate but, as the
Catholic school sector contracts, they are coming to replace Catholic
schools as the schools of choice in urban neighborhoods. In many
cases (including fourteen schools in this study), charter schools also
are physically replacing Catholic schools by operating in closed
Catholic school buildings. Second, in education-reform debates,
charter schools frequently are cited as a means of capturing the
benefits of school choice without enlisting private schools through
voucher and tax-credit programs, which arguably threaten both to
drain public school resources and to undermine public values.9 Our
findings, in contrast, suggest that charter schools may be imperfect
substitutes for “complete” school choice. Charter schools may fill the
educational void left by Catholic schools’ disappearance from our
cities—a possibility about which we remain dubious—but, at least
thus far, they do not appear to replicate Catholic schools’ positive
community benefits. A more complete menu of school-choice
options might help preserve these benefits by stemming the tide of
Catholic school closures.
8 See Margaret F. Brinig and Nicole Stelle Garnett, Catholic Schools, Urban
Neighborhoods, and Education Reform, 85 Notre Dame L Rev 887, 953 (2010); Margaret F.
Brinig and Nicole Stelle Garnett, Catholic Schools and Broken Windows, 9 J Empirical Legal
Stud *49 (forthcoming 2012), online at http://ssrn.com/abstract=1564254 (visited Oct 21, 2011).
9 See David Brooks, The Quiet Revolution, NY Times A35 (Oct 23, 2009).
34 The University of Chicago Law Review [79:31
I. PREVIOUS INVESTIGATIONS
In our previous studies, we sought to measure the effects of
Catholic school closures on perceived disorder, perceived social
cohesion, and crime in Chicago neighborhoods. In our initial study,
we relied upon survey data collected for the Project on Human
Development in Chicago Neighborhoods (PHDCN) to measure the
effects of Catholic school closures on perceived disorder and
perceived social cohesion in Chicago neighborhoods.10 In 1994 and
1995, the PHDCN surveyed approximately four thousand Chicago
residents about perceived levels of neighborhood crime, disorder,
and social cohesion.11 After matching each of the 130 Catholic
elementary schools that closed in the city of Chicago between 1984
and 1994 to the PHDCN data, we estimated the effects of a Catholicschool
closure using two-stage least squares regression analysis, a
method that enabled us both to control for numerous demographic
variables and to employ variables predicting school closures
unrelated to demographics.12 Our analysis linked school closures to
neighborhood social cohesion and increased neighborhood
disorder.13
In our second study, we conducted a latent growth analysis of
effects of Catholic-school closures between 1990 and 1996 on the rate
of serious crime in police beats between 1999 and 2005.14 While crime
decreased across the city of Chicago during this period, our analysis
suggested that Catholic-school closures affected the slope of the
decline.15 That is, “crime decreased more slowly between 1999 and
2005 in police beats where Catholic schools closed between 1990 and
1996.”16 As in our initial study, we incorporated a variable—the parish
leadership characteristics that we describe briefly below—to disaggregate
school-closure decisions from neighborhood demographics.17
10 See Brinig and Garnett, 85 Notre Dame L Rev at 902 (cited in note 8).
11 See Robert J. Sampson and Stephen W. Raudenbush, Systematic Social Observation of
Public Spaces: A New Look at Disorder in Urban Neighborhoods, 105 Am J Soc 603, 619–20, 637
(1999).
12 See Brinig and Garnett, 85 Notre Dame L Rev at 923 (cited in note 8).
13 Id at 924–28.
14 Brinig and Garnett, 9 J Empirical Legal Stud at *1–6 (cited in note 8).
15 Id at *3.
16 Id.
17 Id at *13–17.
2012] Catholic Schools, Charter Schools, and Urban Neighborhoods 35
II. CATHOLIC SCHOOLS AND CHARTER SCHOOLS:
A BRIEF OVERVIEW
Before turning to the empirical questions at the heart of this
investigation, we provide a brief overview of the two distinct
institutional forms we study—Catholic schools and charter schools.
A. Urban Catholic Schools
Traditionally, almost all Catholic elementary schools (the
subject of our study) were “parochial.” That is, they were operated
by a Catholic parish led by a Catholic priest, known as the pastor,
who is the chief operating officer for all parish operations, including
the school.18 In the late nineteenth century, Catholic bishops,
responding to widespread nativism and Protestant indoctrination in
the public schools, began to demand that every parish build and
support a school and that all parish members enroll their children in
it.19 As a result, by the middle of the twentieth century, most major
American cities were densely blanketed with Catholic schools. As
political scientist Gerald Gamm has demonstrated, urban Catholics’
attachments to their parishes and schools fostered a strong
geographic “rootedness” that caused them to suburbanize later, and
to resist racial integration more strenuously, than other white urban
residents.20
By the late 1960s, however, shifting urban demographics and
labor-force realities began to threaten the viability of the parochial
school model, at least in urban areas.21 Historically, parochial schools
were entirely funded by the parish and staffed almost entirely by
religious sisters (nuns) who labored for little more than what one
commentator has called a “token wage.”22 In the 1960s, however,
religious vocations plummeted at the same time that Catholics
suburbanized en masse, causing parochial schools to experience
dramatic increases in labor costs just as collection revenues declined
precipitously.23 Gradually, schools built to educate working-class
Catholic children began to assume the role of educating poor, and
18 See Anthony S. Bryk, Valerie E. Lee, and Peter B. Holland, Catholic Schools and the
Common Good 148–65 (Harvard 1993).
19 Id at 23–33.
20 Gerald Gamm, Urban Exodus: Why the Jews Left Boston and the Catholics Stayed 129,
237–47 (Harvard 1999).
21 See John T. McGreevy, Parish Boundaries: The Catholic Encounter with Race in the
Twentieth-Century Urban North 236 (Chicago 1996).
22 Id at 10, 236.
23 Id at 234–40.
36 The University of Chicago Law Review [79:31
frequently non-Catholic, children.24 Dioceses were forced to take on
more of the financial burden of operating urban parish schools at the
same time they were obligated to build new schools to serve
suburbanizing Catholics. At a more retail level, some priests began
to view schools as an unnecessary burden, especially as the non-
Catholic student population increased.25 The urban parochial model
began to unravel, and dioceses began to close schools in large
numbers.26
Between 1984 and 2004, the Archdiocese of Chicago closed
130 elementary and 18 secondary schools.27 In some cases, several
schools closed in the same neighborhood—not surprisingly, given the
density of schools that historically served different ethnic
populations.28 Despite the many closures, the Archdiocese still
operates the largest nonpublic school system in the country, with 218
elementary schools and 40 high schools enrolling over 96,000
students.29 Most of the elementary schools continue to be operated by
parishes, although the Archdiocese retains supervisory authority
over them and substantially subsidizes many of them, either directly
or through a private philanthropic organization known as the Big
Shoulders Fund.30
B. Charter Schools
Although charter schools have roots in a number of older
reform ideas, they have existed in their current form for less than
two decades,31 and the first charter schools in Chicago opened in
1997.32 Charter schools are public-private hybrids.33 Charter schools
resemble public schools since they are open to all who wish to
attend, tuition free, and secular. Charter schools also are more
accountable than private schools and, arguably, even more than
24 Id at 241–42. See also Bryk, Lee, and Holland, Catholic Schools at 52 (cited in note 18).
25 See McGreevy, Parish Boundaries at 236–40 (cited in note 21).
26 See Brinig and Garnett, 85 Notre Dame L Rev at 892–903 (cited in note 8).
27 See Simons, Closed School History at *2 (cited in note 2).
28 See id.
29 See Office of Catholic Schools, Facts about the Catholic Schools in the Archdiocese of
Chicago (Archdiocese of Chicago 2009), online at http://schools.archchicago.org/public
/factsheet.shtm (visited Oct 21, 2011).
30 See Facts & Accomplishments (Big Shoulders Fund 2010), online at http://
www.bigshouldersfund.org/content/?s=477&s2=484&p=491&t=Facts-&-Accomplishments
(visited Oct 21, 2011).
31 See Finn, Manno, and Vanourek, Charter Schools in Action at 13–22 (cited in note 7).
32 See Office of New Schools, 2008–09 Charter and Contract Schools Performance
Report 1 (Chicago Public Schools 2009), online at http://www.cps.edu/NewSchools/Documents
/2008-2009_PerformanceReport.pdf (visited Oct 21, 2011).
33 Id at 5.
2012] Catholic Schools, Charter Schools, and Urban Neighborhoods 37
traditional public schools, because underperforming charter schools
are more likely to be closed.34 Charter schools also have attributes of
private schools. They are created by private entrepreneurial action—
the request of a private entity (the charter “operator”) for
permission to open a school from a governmental entity (the charter
“sponsor”). Like private schools, charter schools also enjoy
operational autonomy from local school officials (although the
precise extent of the autonomy depends upon state law).35 And, like
private schools, they are schools of choice—that is, parents select
them for their children much as they would a private school.36
While many charter schools focus on values or character
education, and some are structured around cultural themes with
religious overtones,37 an important legal feature distinguishing
charter and private schools is that charter schools are secular.38
Despite this restriction, a number of Catholic dioceses have, or are
considering, “converting” Catholic elementary schools to secular
charter schools rather than closing them.39 Since many states prohibit
charter schools from being operated by, or being affiliated with,
religious institutions, and a handful expressly prohibit the conversion
of all private schools to charter schools,40 dioceses must create or
contract with secular charter operators to operate the “converted”
schools.41 Although the Archdiocese of Chicago has not intentionally
34 See Andrew J. Rotherham and Richard Whitmire, Close Underperforming Charter
Schools, Reward Those That Work, US News & World Rep (June 17, 2009), online at
http://www.usnews.com/opinion/articles/2009/06/17/close-underperforming-charter-schoolsreward-
those-that-work (visited Oct 21, 2011).
35 See Finn, Manno, and Vanourek, Charter Schools in Action at 127–47 (cited in note 7).
36 Minnesota enacted the first charter school law in 1991. See id at 18–22.
37 For examples of such schools, see generally National Heritage Academies, Our
Approach (2011), online at http://www.nhaschools.com/About-Us/Pages/Our-Approach.aspx
(visited Oct 21, 2011); Great Hearts Academies, Great Hearts Academies, online at
http://greatheartsaz.org (visited Oct 12, 2011).
38 See Benjamin Siracusa Hillman, Note, Is There a Place for Religious Charter Schools?,
118 Yale L J 554, 560 (2008). See also Sarah Lemagie, ACLU Settles with State, School Sponsor:
TiZA, A Charter School Accused of Promoting Religion in Violation of the Constitution, Has
Not Reached an Agreement with the ACLU, Star Trib 1B (Feb 8, 2011); Abby Goodnough,
Hebrew Charter School Spurs Florida Church-State Dispute, NY Times A1 (Aug 24, 2007).
39 See, for example, God and Times Tables, Economist 38 (May 15, 2010); Catholic
Schools Get Final OK to Become Charters, Indianapolis Bus J Online (Apr 8, 2010), online at
http://www.ibj.com/catholic-schools-get-final-ok-to-become-charters/PARAMS/article/19166
(visited Oct 21, 2011); Javier C. Hernandez, City Tries New Tactic to Convert Catholic Schools
to Charter Schools, NY Times A22 (Apr 22, 2009); Bill Turque, 7 Catholic Schools in D.C. Set
to Become Charters: Funding Sources Are Still Unclear, Wash Post B01 (June 17, 2008).
40 See, for example, Hernandez, City Tries New Tactic, NY Times at A22 (cited in note 39).
41 See Andy Smarick, Catholic Schools Become Charter Schools: Lessons from the
Washington Experience 9–14 (Seton Education Partners 2009), online at http://
www.setonpartners.org/Seton_DC_Case_Study_FINAL.pdf (visited Oct 21, 2011); Dana
Brinson, Turning Loss into Renewal: Catholic Schools, Charter Schools, and the Miami
38 The University of Chicago Law Review [79:31
converted any of its schools to charter schools, a number of Chicago
charter schools do operate in buildings that formerly housed
Catholic schools, fourteen of which are included in our study.42
During the 2009–10 school year, there were 104 charter schools
(or, technically, 38 charters school operating on 104 campuses) in
Chicago.43 Twenty-eight of these schools opened during the period of
our study.44 These schools are institutionally diverse. They include
elementary schools, junior high schools, and secondary schools, as
well as nontraditional age groupings (for example, grades 6–12).45 At
least two are single-sex schools,46 and several have themed curricula.47
Charter schools enroll a higher proportion of African American
students (65 percent)48 than does the district as a whole (45 percent),49
and a smaller proportion of Hispanic, white, and Asian students.50
Chicago’s charter schools also enroll a slightly higher proportion of
low-income students (85.6 percent)51 than does the Chicago Public
Schools as a whole (83.3 percent).52 The Chicago Public Schools
reports that 63.4 percent of charter school students are “from the
neighborhood.”53
III. CHARTER SCHOOLS, CATHOLIC SCHOOLS, AND CRIME
In this Part, we turn to two questions raised, but unanswered, by
our previous findings. First, we seek to understand whether—as we
Experience 3 (Seton Education Partners 2010), online at http://publicimpact.com/publications
/Seton_Miami_Case_Study.pdf (visited Oct 21, 2011).
42 See Peter Meyer, Catholic Ethos, Public Education: How the Christian Brothers Came
to Start Two Charter Schools in Chicago, 11 Educ Next 40, 43–45 (Spring 2011).
43 Illinois Network of Charter Schools, Illinois Public Charter Schools: Profiles at 1 (cited
in note 4).
44 See id at 2–17.
45 See id.
46 See id at 15, 17.
47 See, for example, Illinois Network of Charter Schools, Illinois Public Charter Schools:
Profiles at 12, 15 (cited in note 4) (describing Prairie Crossing Charter School as “an
environmentally themed school” and Springfield Ball Charter School as having “a theme of
literacy and numeracy”).
48 See id at 1.
49 See Office of New Schools, 2009–2010 Charter and Contract School Performance
Report at *2 (cited in note 5).
50 Compare Illinois Network of Charter Schools, Illinois Public Charter Schools: Profiles
at 1 (cited in note 4), with Office of New Schools, 2009–2010 Charter and Contract School
Performance Report at *2 (cited in note 5).
51 See Julie Woestehoff and Monty Neill, Chicago School Reform: Lessons for the
Nation *39 (Parents United for Responsible Education 2007), online at http://pureparents.org
/data/files/Final report.pdf (visited Oct 21, 2011).
52 See Office of New Schools, 2009–2010 Charter and Contract School Performance
Report at *2 (cited in note 5).
53 See Office of New Schools, 2008–09 Charter and Contract Schools Performance Report
at 5 (cited in note 32).
2012] Catholic Schools, Charter Schools, and Urban Neighborhoods 39
strongly suspect—open Catholic schools suppress neighborhood
crime or, alternatively, whether the negative effects of Catholic
school closings result from the loss of a community institution. This
distinction is an important one. A finding that an open Catholic
school is associated with lower crime rates in a police beat would
support our suspicion that Catholic schools generate social capital. It
would also provide concrete evidence that Catholic schools behave
differently for neighborhoods than public schools, since other
scholars have demonstrated a link between open public schools and
increased crime.54 Alternatively, if our findings reflect loss effects we
might tend to suspect that that the losses of other kinds of
community institutions might also erode neighborhood social
controls. In order to test the effects of open Catholic schools on
crime rates, we use regression analysis to compare the rates of crime
in police beats with Catholic schools to those without them.
Second, we seek to begin to understand whether we have been
finding “Catholic school effects” rather than simply “school effects.”
Here, we add charter schools to our analysis, for several reasons.
Charter schools are imperfect proxies for public schools, especially in
Chicago, where many charter schools function as neighborhood
schools.55 In contrast to traditional public schools, moreover, charter
schools are not present in many police beats, making a comparison
between beats with and without schools possible. Moreover, charter
schools drive Catholic-school closures both because they compete
with Catholic schools,56 and because, as Archdiocesan officials
emphasized in our discussions, the revenue from leasing Catholic
school buildings to charter operators incentivizes some pastors to
lobby for school closures.57 Charter schools are frequently offered by
some as an alternative to school choice programs that might stem the
tide of Catholic-school closures. Finally, charter schools fill the
educational void left when Catholic schools close—and they also
frequently fill the physical space once occupied by closed Catholic
schools. For example, a spokesperson for the Archdiocese of Detroit
recently estimated to one of us that approximately 90 percent of the
54 See Dennis W. Roncek and Antoinette LoBosco, The Effect of High Schools on Crime
in Their Neighborhoods, 64 Soc Sci Q 598, 609–10 (1983). See also Dennis W. Roncek and
Donald Faggiani, High Schools and Crime: A Replication, 26 Sociological Q 491, 501 (1985).
55 See Finn, Manno, and Vanourek, Charter Schools in Action at 17 (cited in note 7).
56 See Samuel G. Freedman, Lessons from Catholic Schools for Public Educators, NY
Times A17 (May 1, 2010).
57 Interview with Sister Mary Paul McCaughey, Superintendent of Catholic Schools,
Archdiocese of Chicago (Mar 20, 2009) (“McCaughey interview”) (on file with authors).
40 The University of Chicago Law Review [79:31
Archdiocese’s closed Catholic-school buildings are currently
occupied by charter schools.58
Our analysis tends to confirm our suspicion that we are finding a
“Catholic school effect” on neighborhood health. We find that beats
with Catholic schools have consistently lower rates of serious crime
and, in contrast, that charter schools are not correlated in a
statistically significant way with crime rates in either direction.
A. Data
We rely on multiple sources of data. The Archdiocese of Chicago
provided information on closed and open Catholic schools, including
their location, name, and parish affiliation. For detailed information
about parish and school leaders, we relied upon The Official Catholic
Directory.59 Information on clergy abuse came from an official
Archdiocesan report,60 from a “victims’ rights” organization that
collects accusations (including unsubstantiated ones),61 and from
newspaper accounts. Data on charter schools came from the Chicago
Public Schools’ Office of New Schools and from the Illinois Network
of Charter Schools.62 To parallel our information on Catholic schools,
we restricted our analysis to charter elementary schools located in the
city of Chicago proper. We excluded high schools as well as
freestanding middle schools. Demographic information comes from
the 2000 census, and the Chicago Police Department provided data on
the incidence of six major crimes (aggravated assault, aggravated
battery, murder, burglary, robbery, and aggravated sexual assault) at
the police-beat level from 1999–2005.63
58 See also Marisa Schultz, DPS Schools Get New Life as Charters: But Critics Say Their
Lure Costs District Money, Detroit News A6 (Oct 2, 2010); Michelle Martin, Charter Schools
Not “Catholic,” Catholic New World (May 27, 2001), online at http://www.catholic
newworld.com /archive/cnw2001/052701/charter_052701.html (visited Oct 21, 2011).
59 Official Catholic Directory 1999–2005 (P.J. Kenedy & Sons 2005) (providing yearly
updated archdiocesan entries approved by each archdiocese).
60 Archdiocesan Priests with Substantiated Allegations of Sexual Misconduct with Minors
(Archdiocese of Chicago 2011), online at http://www.archchicago.org/c_s_abuse/report_032006
/list.pdf (visited Oct 21, 2011).
61 BishopAccountability.org: Documenting the Abuse Crisis in the Roman Catholic
Church (2011), online at http://www.bishop-accountability.org/Who_We_Are/ (visited Oct 21,
2011).
62 See Office of New Schools (Chicago Public Schools 2011), online at http://www.cps.edu
/newschools/Pages/ONS.aspx (visited Oct 21, 2011); Illinois Network of Charter Schools,
Alphabetical Directory (2011), online at http://incschools.org/charters/find_a_charter_school
/full_list (visited Oct 21, 2011).
63 These are all serious (Part I) crimes collected yearly by the Department of Justice and
published as Uniform Crime Reports. See National Atlas of the United States, Summary of the
Uniform Crime Reporting Program (2011), online at http://www.nationalatlas.gov
/articles/people /a_crimereport.html (visited Oct 21, 2011). See also Federal Bureau of
2012] Catholic Schools, Charter Schools, and Urban Neighborhoods 41
B. Explaining School Closures and Openings
We recognize the obvious endogeneity problem we face—that
is, the same factors that predict the location of charter and Catholic
schools also might affect crime rates. In order to separate
demographics and school locations, we sought in our previous
studies to identify variables predicting Catholic-school closures that
were unrelated to neighborhood demographics (or other things
closely associated with crime).64 To do so, we began by asking
Archdiocesan officials what factors drove school closures. While
school-closure decisions are complex, the superintendent of Catholic
Schools, Sister Mary Paul McCaughey, emphasized that, for
struggling schools, the most important factor predicting whether a
school closed was the support of the pastor.65 As she explained, while
school-closure decisions are centralized, the Archdiocese tends to
defer to the pastor’s wishes.66 Pastors who wish to “unload” a school
often get their way,67 and pastors who rally to the school’s defense
often are given a second chance to save it. We therefore directed our
attention to the pastors of the parishes with elementary schools and
found, as she predicted, that certain parish leadership characteristics
were strongly connected with school closings—more so than
neighborhood demographics.68
We do not employ these variables here, however, since we
cannot identify similar variables explaining charter schools’
locations. We assume that charter schools open for many reasons
and that some charter-school operators intentionally locate in poor
urban neighborhoods where crime is more prevalent. To avoid
comparing apples to oranges, we chose not to employ the previously
identified school-closure variables (although they remain predictive
of closures here). This choice limits the strength of our findings,
making it impossible to demonstrate causation—although we
Investigation, Uniform Crime Reports (Department of Justice 2011), online at http://
www.fbi.gov/about-us/cjis/ucr/ucr (visited Oct 21, 2011).
64 See Brinig and Garnett, 9 J Empirical Legal Stud at *13–15 (cited in note 8).
65 See id at *12–13.
66 See id.
67 Id at *13.
68 These include irregularity in parish leadership, meaning that the pastor had been
replaced with a temporary administrator or that a priest at the parish had been accused of sex
abuse. The other characteristic predicting closings was the pastor’s age, although this factor
just missed statistical significance for later closures. Since neither parish “irregularity” nor age
would seemingly have anything to do with demographics or neighborhood crime, we were
comfortable concluding that the parish leadership variables are an appropriate way to address
the endogeneity problem. See Brinig and Garnett, 9 J Empirical Legal Stud at *11–16 (cited in
note 8); Brinig and Garnett, 85 Notre Dame L Rev at 912–20 (cited in note 8).
42 The University of Chicago Law Review [79:31
emphasize our regression analysis does control for neighborhood
demographics.
C. Catholic- and Charter-School Effects on Crime
In controlling for demographics, we include the same
characteristics found by a host of other researchers to explain crime
in Chicago.69 We then matched schools, census tracts, and police
beats using ArcGIS—a mapping program.70 As Table 1 indicates, in
2004, there were Catholic schools in eighty-four distinct police beats
and charter elementary schools in twenty-eight distinct police beats.
Fourteen charter schools were located in closed Catholic schools.
69 See, for example, Andrew V. Papachristos, Tracey L. Meares, and Jeffrey Fagan,
Attention Felons: Evaluating Project Safe Neighborhoods in Chicago, 4 J Empirical Legal
Stud 223, 240 table 1 (2007).
70 The number of census tracts in each beat varied from three to twenty-three, with an
average of more than ten per beat. Visual inspection of these tract-beat matches revealed that
it was nearly impossible to choose a majority or typical tract for many beats, so we included
them all to eliminate subjectivity. One beat (3100) had no people living in it, so the data was
simply excluded, leaving us with 2,902 tract/beat observations for which there were both crime
and census information. Beat 1611 had two Catholic schools but was entered only once for
each tract. Beat 922 had two charters located in one closed school, but again it was entered
only once for each tract.
2012] Catholic Schools, Charter Schools, and Urban Neighborhoods 43
TABLE 1. DESCRIPTIVE STATISTICS
N Minimum Maximum Mean
Standard
Deviation
Charter school (28 distinct
police beats/tracts as
of 2004)
2902 0.00 1.00 0.0096 0.09777
Charter school located in
closed Catholic School
(14 distinct police
beats/tracts as of 2004)
2902 0.00 1.00 0.0048 0.06930
Open Catholic school
(as of 2004, 84 distinct
police beats/tracts)
2902 0.00 1.00 0.0283 0.16573
Total population (2000) 2900 0 15359 3597.01 2671.046
Share of population that is
white (2000)
2900 0.0000 1.0000 0.387388 0.3447462
Share of population that is
nonwhite (2000)
2900 0.0000 1.0000 0.612612 0.3447462
Share of population that is
foreign born (2000)
2900 0.0000 0.7388 0.174533 0.1760852
Share living in same
household 5 years (2000)
2900 0.0000 1.0000 0.537996 0.1638469
Percent in labor force (2000) 2900 0.0000 1.0000 0.590173 0.1396101
Percent below poverty line 2900 0.0000 0.9269 0.219373 0.1599469
Percent ages 15–25 2900 0.0000 0.7047 0.150064 0.0668530
Percent living in rental
housing (2000)
2900 0.0000 1.0000 0.578107 0.2260260
Percent female headed
households (2000)
2900 0.0000 1.0000 0.429929 0.2108200
Percent linguistically isolated
(2000)
2900 0.0000 .6667 0.092068 0.1149682
Median income (2000) ($) 2893 0 127221 37142.81 17830.074
Percent households on public
assistance (2000)
2900 0.0000 1.0000 0.094825 0.1099266
Share of population that is
black (2000)
2898 0.0000 1.0000 0.438765 0.4357780
Share of population that is
Hispanic (2000)
2898 0.0000 1.0000 0.221450 0.2837054
Total crime 1999 2815 0.00 599.00 286.5805 106.73678
Total crime 2000 2829 0.00 546.00 273.8996 99.65107
Total crime 2001 2829 1.00 543.00 260.1467 101.41132
Total crime 2002 2829 40.00 555.00 235.3241 94.63097
Total crime 2003 2829 36.00 568.00 221.8271 93.57419
Total crime 2004 2829 24.00 550.00 221.7236 95.05772
44 The University of Chicago Law Review [79:31
N Minimum Maximum Mean
Standard
Deviation
Total crime 2005 2829 33.00 569.00 253.7656 99.82036
Crime rate 1999 2785 0.00 2685.71 26.2389 110.71567
Crime rate 2000 2799 0.00 3100.00 25.3433 112.12574
Crime rate 2001 2799 0.00 1892.86 18.1555 81.02117
Crime rate 2002 2799 0.28 1671.43 17.1429 71.25692
Crime rate 2003 2799 0.36 1514.29 16.1274 68.99457
Crime rate 2004 2799 0.32 1992.86 14.9863 66.69850
Crime rate 2005 2799 0.19 1864.29 14.9499 66.25297
Crime rate (1999–2005) 2750 2.84 15500.00 152.6250 656.99935
Natural logarithm of crime
rate (1999–2005)
2750 1.05 9.65 4.0445 1.10275
Valid N (listwise) 2743
Prior work indicates that the relationship between crime rates71
and other characteristics is best represented by taking the natural
logarithm of the crime rates.72 For some equations, the demographic
variables were reduced to three factors.73 In others, they were each
entered separately. While crime was declining in Chicago between
1999 and 2005, the crime rate, controlling for demographic factors,
was lower in each year in those beats with Catholic schools than in
those that did not include them.
71 That is, crime divided by the census-tract population. To keep the logarithms positive,
this quotient was multiplied by one hundred.
72 See, for example, Brinig and Garnett, 9 J Empirical Legal Stud at *26 (cited in note 8);
Papachristos, Meares, and Fagan, 4 J Empirical Legal Stud at 245–46 (cited in note 69) (using
“[t]he log of the beat-level homicide rate” in order to “improve model fit and account for
nonlinearity” in a study attempting to evaluate the impact of “Project Safe Neighborhood”
initiatives on neighborhood-level crime rates in Chicago); Sampson and Raudenbush, 105 Am
J Soc at 621 (cited in note 11).
73 The technique, called principal component analysis, uses regression results to reduce a
large set of possibly correlated variables into a subset of uncorrelated variables. Because there
are fewer variables, the coefficients for the different types of schools are larger, though
statistical significance and direction do not change. The same technique has been used on
Chicago crime data in earlier work. See Sampson and Raudenbush, 105 Am J Soc at 621–23 &
n 19 (cited in note 11).
2012] Catholic Schools, Charter Schools, and Urban Neighborhoods 45
FIGURE 1. COMPARISON OF POLICE BEATS WITH AND WITHOUT
OPEN CATHOLIC SCHOOLS
To produce Figure 1, we first separated the data into the police
beats where there were and were not open Catholic schools. Then
we used regression analysis to predict the average crime rate for each
of the seven years. In effect, we held constant the presence or the
absence of a charter school (located or not in a closed Catholic
school), plus the computed composite-socioeconomic Table 1 details.
The coefficients are displayed as Model 1 in Table 2; the mean
adjusted predicted values for each year make up the points in
Figure 1. The other two models of Table 2 are very similar, and
graphing them would produce nearly identical results. Model 2
displays regression coefficients, standard errors, and statistical
significance for an equation where instead of grouping the
demographic characteristics, the characteristics are broken out into
original census data for each police beat and tract combination.
Figure 1 used the variables from Model 1 from Table 2 below for
each of the seven years of crime rate data, comparing results for
cases in which there was and was not an open Catholic elementary
school. The difference between Model 2 and Model 3 is that the
former uses individual race characteristics while the latter groups
them together.
TABLE 2. REGRESSION RESULTS: CRIME RATES AND
.0000
.5000
1.0000
1.5000
2.0000
2.5000
3.0000
1999
2000
2001
2002
2003
2004
2005
Adjusted Predicted Mean Value of Logged
Crime Rate
Year
Open Catholic
School
No Open Catholic
School
46 The University of Chicago Law Review [79:31
PRIVATE SCHOOLS
Model 1 Model 2 Model 3
Beta
(Standard error)
Beta
(Standard error)
Beta
(Standard error)
Constant 4.036
(.028)***
5.870
(.164)***
5.750
(.167)***
Open Catholic
school
-0.264
(.109)*
-0.120
(.068)*
-0.160
(.069)*
Charter school 0.154
(.210)
-0.015
(.130)
0.026
(.133)
Charter in
Catholic
0.045
(.282)
0.140
(.175)
0.181
(.179)
PCA1
(deprivation)
0.470
(.018)***
--- ---
PCA2
(immigration)
-0.162
(.018)***
--- ---
PCA3
(stability)
-0.220
(.018)***
--- ---
Total
population
2000
--- 0.000
(.000)***
0.000
(.000)***
Share nonwhite
2000
--- --- 0.926
(.060)***
Share foreign
born 2000
--- -0.397
(.161)*
-0.395
(.147)**
Same
household
2000
--- -0.399
(.108)***
-0.198
(.109)
Percent in labor
force
--- -0.709
(.157)***
-0.624
(.161)***
Percent below
poverty line
--- -1.163
(.144)***
-1.064
(.147)***
Percent renter --- 0.202
(.087)**
0.183
(.088)*
Percent female
head 2000
--- -1.186
(.119)***
-1.143
(.119)***
Percent
linguistically
isolated 2000
--- -0.301
(.218)
0.218
(.214)
Median income
2000
--- -5.233E-6
(.000)***
-5.855E-6
(.000)***
2012] Catholic Schools, Charter Schools, and Urban Neighborhoods 47
Model 1 Model 2 Model 3
Beta
(Standard error)
Beta
(Standard error)
Beta
(Standard error)
Percent
households
with public
assistance
2000
--- 1.875
(.176)***
1.696
(.180)***
Percent Black
2000
--- 0.987
(.064)***
---
Percent
Hispanic 2000
--- 0.978
(.067)***
---
R2 (adjusted) .259 .714 .702
F 240.812 428.332 .432
Note: *** signifies p < .001, ** signifies p < .01, and * signifies p < .05.
In other words, regardless of how we account for demographic
variables that generally predict crime, an open Catholic elementary
school in a beat is associated with a statistically significant decrease
in the rate of crime.74 Although the percentage difference varied by
year, the crime rate in police beats with Catholic schools was, on
average, at least 33 percent lower than police beats without them.
Charter schools appear to have no statistically significant effect on
crime in either direction, although, in a few years, regressions for
individual crimes suggest a statistically significant link between
charter schools and elevated rates of aggravated assault and
aggravated battery. To the extent we can note anything about the
charter schools operating in closed Catholic schools, the direction of
the coefficients is not encouraging (that is, crime seems to increase).
74 In the simplest model of all, looking at the correlation between open Catholic
elementary schools and the logged crime rate, the coefficient is −.086 at p < .001. A model that
simply considers race and income generates an adjusted R2 of .141 (F=119.934), with the
coefficients as follows. The effect of the Catholic school being open, standardized, is almost
exactly the same as the effect of income.
Variable Beta Standard Error Beta (standardized)
Constant 3.603*** .100 ---
Open Catholic
school
(as of 2004)
-0.371*** .113 -0.058
Percent black in
census tract
(2000)
0.987*** .074 0.399
Percent Hispanic in
census tract
(2000)
0.591*** .097 0.156
Median income
(2000)
-3.749E-6*** .000 -0.059
Dependent variable: log crime rate; p < .001 for all coefficients.
48 The University of Chicago Law Review [79:31
We cannot at this point say that opening a new Catholic school
(or re-opening a closed one) would decrease crime. We also cannot
say whether individual charter schools—including, perhaps
especially, charter schools that mimic the educational program of
Catholic schools—suppress crime, or whether charter schools will, as
they become established, have the same positive effects as Catholic
schools. We also cannot say whether “converting” Catholic schools
to charter schools will maintain Catholic schools’ positive effects—
although our findings here suggest that simply operating a
nonsectarian charter in a closed Catholic school does not.
IV. EXPLAINING CATHOLIC SCHOOLS’ POSITIVE EXTERNALITIES
Our findings suggesting that charter schools do not suppress
crime are not inconsistent with previous studies linking public
schools with disorder and crime.75 Our research suggests, however,
that urban Catholic elementary schools have the opposite effect—
that is, that they suppress disorder and serious crime. We can, at this
point, only speculate about possible explanations for Catholic
schools’ positive externalities.
A. The “Night Watchman” Explanation
One study linking public schools and crime found that public
elementary schools appeared to generate more crime than public high
schools. The authors speculated that unsupervised playgrounds may
serve as recreational hangouts for teenagers or staging areas for
illicit activities.76 Perhaps, therefore, Catholic school facilities simply
are more secure than charter schools. Or perhaps Catholic schools
are more likely to generate what Jane Jacobs famously termed “eyes
upon the street.”77 In most parishes, for example, the pastor lives onsite
and may serve a “night watchman” function.78 Catholic schools
75 See, for example, Lisa Broidy, Dale Willits, and Kristine Denman, Schools and
Neighborhood Crime *6–12 (Justice Research Statistics Association 2009), online at
http://www.jrsa.org/ibrrc/background-status/New_Mexico/Schools_Crime.pdf (visited Oct 21,
2011); Caterina Gouvis Roman, Schools as Generators of Crime: Routine Activities and the
Sociology of Place *111–18 (National Criminal Justice Reference Service 2002); Roncek and
Faggiani, 26 Sociological Q at 501 (cited in note 54); Roncek and LoBosco, 64 Soc Sci Q at 609–10
(cited in note 54).
76 See Paula M. Kautt and Dennis W. Roncek, Schools as Criminal “Hot Spots”:
Primary, Secondary, and Beyond, 32 Crim Just Rev 339, 349–53 (2007).
77 See Jane Jacobs, The Death and Life of Great American Cities 34–35 (Random
House 1961).
78 In our study, however, the Archdiocese closed only a handful of parishes, so the pastor
remained on-site even after the school was shuttered.
2012] Catholic Schools, Charter Schools, and Urban Neighborhoods 49
also might be more likely to sponsor community activities during
after-school hours that draw adults into the neighborhood.79
B. The Student-Body Explanation
Our results also might reflect the fact that Catholic and charter
schools enroll different types of students. A greater degree of
institutional diversity existed among charter schools than Catholic
schools during the period of our study. Most (but not all) of the
Catholic schools in our study enrolled grades kindergarten (or
prekindergarten) through eight. Most of the charter schools did not;
some extended through fifth or sixth grades; and others were limited
to the middle school years (although we excluded these from our
analysis). The educational-psychology literature on “school
transitions” suggests that students perform better—in terms of
behavior, academic achievement, and self-esteem—in K–8 schools.80 If
older students are more likely to generate disorder—and researchers
have linked the greater incidence of crime near public middle and high
schools with the presence of large numbers of adolescents81—then
Catholic schools’ practice of combining elementary- and middleschool
students may generate positive neighborhood externalities.
Moreover, Catholic schools’ control over student-body
composition is frequently cited as contributing to their relative
educational success, as is the fact that better-educated, highly
motivated parents are more likely to choose Catholic schools for
their children.82 Both factors may help explain Catholic and charter
schools’ divergent neighborhood effects. Charter schools exercise far
less enrollment discretion than Catholic schools. They generally must
conduct a lottery for admissions, although they may give priority to
students residing within their attendance boundary (if one is
designated).83 Charter schools may also find it more difficult to expel
disruptive students who may “act out” both inside and outside the
classroom setting. Moreover, although Catholic school tuition is very
low relative to that of other types of private schools,84 a decision to
79 Charles W. Dahm, Parish Ministry in a Hispanic Community 238–51 (Paulist 2004).
80 See Jonah E. Rockoff and Benjamin B. Lockwood, Stuck in the Middle: How and Why
Middle Schools Harm Student Achievement, 10 Educ Next 68, 69–75 (Fall 2010).
81 See, for example, Roncek and LoBosco, 64 Soc Sci Q at 601 (cited in note 54).
82 See Bryk, Lee, and Holland, Catholic Schools at 16, 46–54 (cited in note 18).
83 See, for example, Office of New Schools, Lottery Guidelines for Charter Schools *1
(Chicago Public Schools 2011), online at http://cps.edu/NewSchools/Documents/Lottery
GuidelinesForCharterSchools.pdf (visited Oct 21, 2011).
84 In 2007–08, the average tuition at a Catholic elementary school was $4,944; the
average tuition at a nonsectarian private elementary school was $15,945. See Council for
50 The University of Chicago Law Review [79:31
send a child to a Catholic school signals a threshold level of parental
motivation—and motivated parents may be better able to control
their children’s behavior before and after school. Catholic schools
also frequently place demands on parents that public schools do not
or even cannot: many either require parents to volunteer in the
school or provide parents with the option of volunteering in order to
reduce tuition burdens.85 These requirements may generate a stable
flow of responsible adults in the neighborhood who help keep
disorder and crime in check.
That said, it is important not to overstate the explanatory value
of these factors. For example, enrollment in a charter school also
signals parental motivation. A parent selecting a charter school must
opt out of the traditional public school system and choose among a
range of alternatives.86 Furthermore, most urban Catholic schools
provide significant financial assistance, which enables them to enroll
students from disadvantaged backgrounds. Indeed, the available
evidence suggests that the educational benefits of Catholic schools
are greatest for the poorest, most disadvantaged students.87
C. The Neighborhood-Network Explanation
A third explanation is suggested by William Fischel’s defense of
local public schools.88 Fischel argues that parent networks at
neighborhood public schools enable “community-specific social
capital.”89 As Fischel observes, “My approach to social capital
formation simply requires that parents get to know other parents. . . .
[A]nd sending your child to a local school does that more effectively
than any other means.”90 As he acknowledges, however, the
neighborhood-network benefits of public schools likely are reduced in
major cities, where intradistrict public school choice is commonplace.
Indeed, given the prevalence of public school choice in Chicago—
where more than one-third of all public school elementary students
American Private Education, Facts and Studies (2011), online at
http://www.capenet.org/facts.html (visited Oct 21, 2011).
85 See Bryk, Lee, and Holland, Catholic Schools at 306–08 (cited in note 18).
86 Jack Buckley and Mark Schneider, School Choices, Parental Information, and Tiebout
Sorting: Evidence from Washington, DC, in William A. Fischel, ed, The Tiebout Model at Fifty
101, 104 (Lincoln Institute of Land Policy 2006).
87 See Greeley, Catholic High Schools at 107–08 (cited in note 6); Coleman, Hoffer, and
Kilgore, High School Achievement at 143–46 (cited in note 6).
88 See William A. Fischel, The Homevoter Hypothesis: How Home Values Influence
Local Government Taxation, School Finance, and Land-Use Policies 154–55 (Harvard 2001).
89 William A. Fischel, Why Voters Veto Vouchers: Public Schools and Community-
Specific Social Capital, 7 Econ Gov 109, 112–17 (2006). See also Fischel, The Homevoter
Hypothesis at 142–43, 154–55 (cited in note 88).
90 Fischel, 7 Econ Gov at 116 (cited in note 89).
2012] Catholic Schools, Charter Schools, and Urban Neighborhoods 51
attend a school outside their geographic attendance boundaries—
students attending a Catholic school may be more likely to live in the
surrounding neighborhood than public school students, and Catholic
schools may be more likely to generate local social capital.91
That said, although we suspect that many Catholic schools are
neighborhood schools, most of the city’s charter schools are as well.
Illinois law authorizes the designation of an attendance boundary for
charter schools and permits charter schools to give priority to students
residing within their boundaries.92 During the 2008–09 school year,
Chicago charter schools, on average, drew approximately 63 percent
of their students from the surrounding neighborhood, although the
percentage of neighborhood students ranged from a low of 6.7 percent
to a high of 100 percent.93 Despite this fact, however, charter schools
do not appear to serve the same social-capital-generation function as
their Catholic school counterparts—or, if they do, the social capital
does not translate into reduced crime rates.
D. The Longevity Explanation
All of the Catholic schools that remained open during the
period of our study had been open since the 1930s (though some of
them had received children originally attending other schools closed
since 1984). In contrast, none of the charter schools opened before
1997.94 Over time, as charter schools become more integrated into
neighborhoods, they also may produce similar effects. It is also
possible that Catholic schools, by virtue of their longevity in a
community, will continue to produce positive effects even if they are
“converted” to charter schools. We simply cannot speculate, based
upon our data, about either possibility.
E. The “Last Vestige of Civilization” Explanation
Our results also might reflect the unfortunate reality that, in
some neighborhoods, a Catholic school was one of the last remaining
functional community institutions. As a Catholic bishop who served
as a priest in the Archdiocese of Chicago (and who attended one of
the closed schools in our study) told one of us privately, in some
neighborhoods, a Catholic school was the last vestige of civil
91 See Julie Berry Cullen and Brian A. Jacob, Is Gaining Access to Selective Elementary
Schools Gaining Ground? Evidence from Randomized Lotteries, in Jonathan Gruber, ed, The
Problems of Disadvantaged Youth: An Economic Perspective 43, 51–52 (Chicago 2009).
92 See 105 ILCS § 5/27A-4(d).
93 See Office of New Schools, 2008–09 Charter and Contract Schools Performance Report
at 5 (cited in note 32).
94 Id at 1.
52 The University of Chicago Law Review [79:31
society.95 It is hardly surprising, he remarked, that when the school
disappeared, the neighborhood rapidly declined.96
F. The “Catholic School Effect” Explanation
This fact leaves us to wonder whether our results are suggestive
of another possibility—namely, that what goes on inside a school
does in fact affect what happens outside it. That is, we ask whether
the human-capital- and social-capital-generating functions of a
school intersect. According to James Coleman’s classic formulation,
social capital “inheres in the structure of relations between actors
and among actors,” and institutions that foster these relationships
are incubators of social capital.97 Coleman used schools to illustrate
this conception of social capital, arguing that successful schools
tended to be distinguished by parents’ connections to their children’s
school and to the parents of their children’s peers.98 These
connections, he reasoned, “closed the loop” between schools,
teachers, and parents, thus guaranteeing the enforcement of
appropriate norms.99 Coleman further argued that these kinds of
connections—and the norm-enforcement authority that they
enabled—explained Catholic high schools’ extremely low drop-out
rates.100 Perhaps they also generate positive externalities beyond the
classroom walls. For example, in their influential book, Catholic
Schools and the Common Good, Anthony Bryk and his colleagues
linked Catholic high schools’ educational successes to the fact that
these schools were intentional communities, with high levels of trust
between students, parents, teachers, and administrators.101 In more
recent work, Bryk has argued that neighborhood factors, including
the level of collective efficacy and social capital in a community, are
critical inputs to urban public schools’ success (or failure).102 Bryk’s
work suggests that there may be significant feedback effects between
what goes on in a school and what occurs in the surrounding
community.103
95 Confidential interview with Catholic bishop who formerly served as a priest from the
Archdiocese of Chicago (“Priest Interview”) (on file with author).
96 Id.
97 James S. Coleman, Social Capital in the Creation of Human Capital, 94 Am J
Sociology S95, S98 (1998).
98 See id at S105–08.
99 See id.
100 See id at S114–15.
101 See Bryk, Lee, and Holland, Catholic Schools at 307–08, 313–14 (cited in note 18).
102 See Anthony S. Bryk, et al, Organizing Schools for Improvement: Lessons from
Chicago 177–78 (Chicago 2010).
103 See Bryk, Lee, and Holland, Catholic Schools at 282–85 (cited in note 18); Bryk, et al,
Organizing Schools at 196 (cited in note 102).
2012] Catholic Schools, Charter Schools, and Urban Neighborhoods 53
That said, many charter schools employ educational strategies
that closely approximate the Catholic school formula, including a
highly structured school day, traditional curriculum, high levels of
parental involvement, and an emphasis on building an educational
community between the various school stakeholders.104 Since our
analysis does not distinguish between different charter schools’
educational strategies, we cannot say whether schools employing this
formula positively impact neighborhoods in the way that our study
suggests Catholic schools do—or whether they might come to do so
over time.
.