Illegal Alien Schoolchildren: Issues in Estimating State-by-State
Costs (23-JUN-04, GAO-04-733).					 
                                                                 
In 1982 the Supreme Court ruled that states and school districts 
cannot deny education to illegal alien children residing here.	 
Issues in estimating the costs of providing education to them are
of interest because (1) policy discussions concerning illegal	 
immigration often focus on cost impacts; (2) potential costs are 
borne mostly at the state and local levels; and (3) the Congress 
could authorize federal reimbursement for benefits provided to	 
illegal aliens, based on estimated state costs or numbers of	 
illegal aliens. The foreign-born population is growing and is	 
concentrated in certain states; the illegal immigrant component  
is thought to be substantial. Concerns about education costs may 
reflect "squeezed" state and local budgets, rising school	 
enrollments, and overcrowded schools. To address the potential	 
for estimating the costs of educating illegal alien		 
schoolchildren, this report (1) identifies major government	 
sources of relevant data, (2) describes a Census Bureau plan for 
developing new information, and (3) outlines costestimation	 
approaches. GAO provided a draft of this report to the National  
Center for Education Statistics, the Department of Homeland	 
Security, and the Census Bureau. The agencies informed GAO they  
had no formal comments. 					 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-04-733 					        
    ACCNO:   A10564						        
  TITLE:     Illegal Alien Schoolchildren: Issues in Estimating       
State-by-State Costs						 
     DATE:   06/23/2004 
  SUBJECT:   Cost analysis					 
	     Data collection					 
	     Data integrity					 
	     Education or training costs			 
	     Elementary school students 			 
	     Illegal aliens					 
	     Intergovernmental relations			 
	     Population statistics				 
	     Projections					 
	     Public schools					 
	     Secondary school students				 
	     State programs					 
	     Cost estimates					 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Product.                                                 **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-04-733

United States General Accounting Office

GAO	Report to the Chairman, Committee on the Judiciary, House of Representatives

June 2004

ILLEGAL ALIEN SCHOOLCHILDREN

                              Issues in Estimating
                              State-by-State Costs

GAO-04-733

Highlights of GAO-04-733, a report to the Chairman of the Committee on the
Judiciary, House of Representatives

In 1982 the Supreme Court ruled that states and school districts cannot
deny education to illegal alien children residing here. Issues in
estimating the costs of providing education to them are of interest
because (1) policy discussions concerning illegal immigration often focus
on cost impacts; (2) potential costs are borne mostly at the state and
local levels; and (3) the Congress could authorize federal reimbursement
for benefits provided to illegal aliens, based on estimated state costs or
numbers of illegal aliens.

The foreign-born population is growing and is concentrated in certain
states; the illegal immigrant component is thought to be substantial.
Concerns about education costs may reflect "squeezed" state and local
budgets, rising school enrollments, and overcrowded schools.

To address the potential for estimating the costs of educating illegal
alien schoolchildren, this report (1) identifies major government sources
of relevant data, (2) describes a Census Bureau plan for developing new
information, and (3) outlines costestimation approaches.

GAO provided a draft of this report to the National Center for Education
Statistics, the Department of Homeland Security, and the Census Bureau.
The agencies informed GAO they had no formal comments.

www.gao.gov/cgi-bin/getrpt?GAO-04-733.

To view the full product, including the scope and methodology, click on
the link above. For more information, contact Nancy Kingsbury at (202)
512-2700 or [email protected].

June 2004

ILLEGAL ALIEN SCHOOLCHILDREN

Issues in Estimating State-by-State Costs

Current government information is not sufficient to directly estimate the
stateby-state costs of educating illegal alien schoolchildren. Although a
variety of data are available, no government source estimates the numbers
of illegal alien schoolchildren for most or all states. Specifically:

o  	States and local areas record data on school enrollment and costs but
not on immigration status. In response to GAO's survey, a few states
estimated costs of educating illegal alien children, based partly on
assumptions.

o  	The National Center for Education Statistics (NCES) maintains
enrollment and cost data-but has no information on immigration status.

o  	The Department of Homeland Security (DHS) developed state-by-state
estimates of the illegal alien population, but the estimates do not break
out age groups and are subject to methodological limitations.

The Census Bureau is developing a plan to estimate the size of the
resident illegal alien population, indirectly by age group and state. This
new information might help in developing state-by-state estimates of the
number of school-age illegal alien children. However, the plan does not
specify the age groups to be estimated, faces technical challenges, and
depends upon future funding. Overall, it is too early to evaluate the
Census Bureau's plan.

The simplest approach to estimating the costs of educating illegal alien
children is to multiply average current per pupil expenditures by the
estimated number of illegal alien schoolchildren separately for each state
(see fig.). At present, government information is insufficient for
developing reliable estimates based on this approach. If the Census
Bureau's plan proves successful, relevant data would be available by
2007-09. Taking account of cost determinants such as variation in local
area expenditures, student needs, and school capacity requires additional
data.

Simplest Approach to State-by-State Cost Estimation

Sources: GAO (analysis), NCES (per pupil expenditure data), Census Bureau
(estimation plans), U.S. Department of Education (photo).

Contents

Letter                                                                   1 
                                  Results in Brief                          2 
                                     Background                             3 
                         Objectives, Scope, and Methodology                10 
                     Currently Available Government Information            12 
           The Census Bureau's Plans for Estimating Illegal Immigrants Are 
                                     Preliminary                           18 
                            Approaches to Cost Estimation                  21 
                                    Observations                           28 
                                   Agency Comments                         28 

  Table

Table 1: Cost Categories Relevant to the School-Capacity Approach

  Figures

Figure 1: Total Foreign-Born Population (Legal and Illegal) in Key States,
1990 and 2000 7 Figure 2: Simplest State-by-State Estimation Approach 22

Abbreviations

ACS American Community Survey
CPS Current Population Survey
DHS Department of Homeland Security
ELL English language learner
INS Immigration and Naturalization Service
NCES National Center for Education Statistics
TPS temporary protected status

This is a work of the U.S. government and is not subject to copyright
protection in the United States. It may be reproduced and distributed in
its entirety without further permission from GAO. However, because this
work may contain copyrighted images or other material, permission from the
copyright holder may be necessary if you wish to reproduce this material
separately.

United States General Accounting Office Washington, DC 20548

June 21, 2004

The Honorable F. James Sensenbrenner, Jr.
Chairman
Committee on the Judiciary
House of Representatives

Dear Mr. Chairman:

Reliable data are needed to support policy decisions in the area of
illegal
immigration, but such data have often been lacking or inadequate.1 In this
report, we respond to your request that we examine the potential for
estimating state-by-state costs of public schooling for illegal alien
children,
based on government information.2

As you know, the U.S. Supreme Court has ruled that states and school
districts cannot deny K-12 education to resident children who are illegal
aliens. Government estimates of the state-by-state costs of educating them
would, however, be policy relevant for a number of reasons, such as the
potential for federal reimbursement to states.

Because government estimates are not available at present, in this report
we (1) identify currently available government information, (2) describe
the Census Bureau's plan for developing new information, and (3) outline
possible approaches to estimating costs and, for each approach, briefly
assess whether needed information is currently available or planned.

GAO collected relevant information through a survey of 20 states, did
outreach through associations of governors and mayors, conducted

1See U.S. Commission on Immigration Reform, Becoming an American:
Immigration and Immigrant Policy (Washington, D.C.: Government Printing
Office, 1997), pp. 194-98. We pointed to this issue in U.S. General
Accounting Office, Immigration Statistics: Information Gaps, Quality
Issues Limit Utility of Federal Data to Policymakers, GAO/GGD-98-164
(Washington, D.C., July 31, 1998), pp. 12-13 and 22-24.

2According to data from the National Center for Education Statistics
(NCES), the proportion of students in private elementary and secondary
schools has changed little over the past 10 years, remaining around 11
percent; we do not include or discuss expenditures by private schools in
this report.

  Results in Brief

interviews at federal agencies, performed a literature review, and
consulted with experts.

Overall, our review indicates that at present, government information is
not sufficient for directly estimating the state-by-state costs of
providing public schooling to illegal alien children. Specifically:

o  	Government data include a variety of information on school
expenditures and on the foreign-born population. However, estimates of the
numbers of illegal alien schoolchildren, by state, are lacking.

o  	The Census Bureau has outlined a preliminary plan to develop indirect
state-by-state estimates of the resident illegal alien population by age.
Such estimates could help determine the number of schoolchildren in that
population. But even if age-group estimates are successfully developed,
they would not be available before 2007-09.

o  	Approaches to state-by-state cost estimation differ in the extent to
which they attempt to account for various factors that can affect costs,
but all approaches require data on or estimates of the number of illegal
alien schoolchildren in each state.

State governments and school districts routinely record school enrollments
and dollar expenditures. Although a few states responded to our survey
with estimates of the costs of educating illegal alien children, none
actually collect data on children's immigration status.

At the federal level, the National Center for Education Statistics (NCES)
centrally maintains data on school enrollment and expenditures, as well as
certain other information. (For example, NCES has collected data on school
overcrowding for a sample of areas.) But like the states and school
districts, NCES has no data on the immigration status of schoolchildren.

The Census Bureau collects citizenship data on foreign-born U.S.
residents, including children, but does not ask further questions about
immigration status.

The Department of Homeland Security (DHS) has developed indirect
stateby-state estimates of the size of the resident illegal population,
but these

estimates do not break out age groups; the DHS estimates are subject to
methodological limitations as well.3

The Census Bureau has outlined a plan for research to develop age-group
estimates for the resident illegal immigrant population, by geographic
divisions (states and selected counties), with trends across time. It is
too early to judge the quality of this plan, but if it is successfully
implemented, it could help quantify the school-age portion of the resident
illegal alien population. The Census Bureau also suggested that other
government agencies might expand on its planned population estimates-for
example, by developing further estimates on the extent to which illegal
alien children attend public schools or require special programs.

The simplest approach to estimating state-by-state costs multiplies
current average per pupil expenditures for each state by an estimate of
the number of illegal alien children attending school in that state. Other
approaches take account of potentially key cost factors, including cost
variations across local areas within a state and higher costs for students
needing special programs, such as English Language Learner (ELL)
programs.4 Yet another approach assesses the role of enrollment growth,
school capacity, and "incremental costs."5 All these approaches require
estimates of numbers of illegal alien schoolchildren by state.

We provided a draft of this report to the Census Bureau, DHS, and NCES.
The agencies informed us that they had no formal comments, but we received
informal comments from the Census Bureau and NCES on minor technical
points and made changes in the report as appropriate.

Background 	In 1982, the U.S. Supreme Court ruled that it would be
unconstitutional for any state or school district to deny K-12 education
to a child residing in that state or school district on the basis of the
child's being an illegal

3DHS estimates use decennial census data on the foreign-born population
and administrative data on legal immigrants, among other sources. These
estimates are termed "indirect" because information on the illegal
immigrant population is not directly collected; instead, a variety of
other estimates (some marked by uncertainty) are combined to project
state-by-state figures for this population.

4The term English language learner is defined in the Background section of
this report (see

p. 10).

5Theoretically, in schools or classes with excess capacity, enrollment
growth might not cause overcrowding, and incremental per pupil costs might
be lower than average per pupil costs. Otherwise, the opposite would hold
true.

alien.6 Nevertheless, issues in estimating the state-by-state costs of
educating illegal alien children are policy relevant because

1. 	concern about cost impacts is often among the issues raised in debates
and discussions about immigration policy,7

2. 	illegal alien schoolchildren represent a potential cost component that
is particularly relevant to states and local areas because they bear most
of the costs of educating illegal alien schoolchildren,8 and

3. 	the Congress could authorize federal reimbursement of the costs of
providing K-12 education to illegal alien children, based on the
stateby-state costs of educating them or on the estimated numbers residing
in each state.9

The following sections provide background on (1) issues that have
heightened concern about the costs of educating illegal alien
schoolchildren; (2) variation in education expenditures-and immigration
levels-across states and local areas; and (3) other factors that may
contribute to cost variation.

Concern about Costs May Concern about the costs of illegal alien
schoolchildren may be heightened

Be Heightened 	because education costs are high and the illegal immigrant
population is thought to be large.

According to NCES, for the 1999-2000 school year, current expenditures by
primary and secondary public schools-not including capital outlays-

6See Plyler v. Doe, 457 U.S. 202 (1982).

7U.S. Commission on Immigration Reform, U.S. Immigration Policy: Restoring
Credibility. (Washington, D.C.: Government Printing Office, 1994), pp.
109-43.

8U.S. Commission on Immigration Reform, U.S. Immigration Policy: Restoring
Credibility, pp. 143-52.

9Such reimbursement was recently enacted for emergency health benefits.
The Medicare Prescription Drug, Improvement, and Modernization Act of
2003, Public Law 108-173 (Dec. 8, 2003), S:1011, Federal Reimbursement of
Emergency Health Services Furnished to Undocumented Aliens, specifies that
annually, from 2005 through 2008, $167 million will be distributed among
the states, based on the estimated percentage of undocumented aliens
residing in each state. Also see U.S. General Accounting Office,
Undocumented Aliens: Questions Persist about Their Impact on Hospitals'
Uncompensated Care Costs.

GAO-04-472 (Washington, D.C.: May 21, 2004).

totaled about $324 billion. Capital outlays in that school year-for
facilities acquisition and construction-were an additional $35 billion.10
These costs were borne primarily at state and local levels; federal
dollars represented about 7 percent of school revenue.11

The Immigration and Naturalization Service (INS) estimated that the
illegal immigrant population increased to 7 million as of January 2000.12
However, there is no government estimate of the percentage of the illegal
alien population that is of school age or that is now attending public
school in grades K-12.

Some possibly related reasons for concern about costs include the
following:

o  	Many state and local budgets are squeezed, and education costs
represent their largest expenditure item.13

10An additional $22 billion was spent on other categories and programs,
some not part of K12 education. "Current expenditures," as used in the
NCES finance surveys, means current expenditures for public elementary and
secondary education-not including capital outlays. Current expenditures
include instruction expenditures and expenditures for other functions that
support public elementary and secondary education, such as school support
(guidance counselors, nurses), instructional support (libraries, teacher
training), administration, student transportation, and food services.
School districts also spend money on functions that fall outside of public
elementary and secondary education, such as adult education, community
services, and support for private education. NCES collects and tracks
these expenditures separately.

11Local dollars represented 43 percent of total school revenues and state
dollars 50 percent. National Center for Education Statistics, Statistics
in Brief (Washington, D.C.: May 2002),

p. 1. More recently, the No Child Left Behind Act of 2001 (Public Law
107-110, Jan. 8, 2002) may have increased the federal share slightly, at
least in some states. Notably, title III specifies a formula grant
program, based on estimated state-by-state distributions of children who
have limited English proficiency and are recent immigrants.

12U.S. Immigration and Naturalization Service, Office of Policy and
Planning, "Estimates of the Unauthorized Immigrant Population Residing in
the United States: 1990-2000," Washington, D.C., January 2003. (On March
1, 2003, INS was transferred from the Department of Justice to DHS.) We
discuss the 7 million estimate in a later section of this report; for a
fuller discussion, see U.S. General Accounting Office, Overstay Tracking:
A Key Component of Homeland Security and a Layered Defense, GAO-04-82
(Washington, D.C.: May 21, 2004), app. III.

13See Elizabeth McNichol and Makeda Harris ("Many States Cut Budgets as
Fiscal Squeeze Continues," Center on Budget and Policy Priorities,
Washington, D.C., 2004), who state that "the amount of state education
spending included in proposed fiscal year 2005 budgets in a number of
states-including California . . . [and] New York . . . falls short of the
amount needed to maintain current-law funding levels or restore cuts made
over the last few years."

o  	School enrollments are rising. From 1990 to 2000, school enrollments
rose about 15 percent-with steeper rises in states like Nevada (with an
increase close to 70 percent), Arizona and Florida (each with an increase
greater than 30 percent), and California, Colorado, and Georgia (each with
an increase of about 25 percent).

o  	Despite a 1990s boom in school construction, one in five schools were
overcrowded in 1999, rising to one in three for schools in western
states.14

    Immigration Levels and Education Costs Vary across Areas

It is likely that the costs of educating illegal alien children vary
across states and local areas because (1) immigration (including
foreign-born persons residing here legally and illegally) is heavily
concentrated in certain areas and (2) per pupil school expenditures vary
by state and school district.

As shown in figure 1, between 1990 and 2000 foreign-born populations
increased in the top 5 traditional destination states-California, Florida,
Illinois, New York, and Texas. As of 2000, the foreign-born represented 12
to 26 percent of these states' populations. (Illegal immigration may
follow similar patterns; for example, INS estimated that as of January
2000, 2 million illegal immigrants resided in California, 1 million in
Texas, and nearly  1/2 million in New York.) Also, as illustrated in
figure 1, certain smaller states (with somewhat lower concentrations of
immigrants-for example, Georgia, Nevada, and North Carolina have 5 to 16
percent foreign born) have recently experienced growth in their
foreign-born populations.

14U.S. General Accounting Office, School Facilities: Construction
Expenditures Have Grown Significantly in Recent Years, GAO/HEHS-00-41
(Washington, D.C.: Mar. 3, 2000),

p. 7, and National Center for Education Statistics, The Condition of
Education 2001: Indicator 45, Overcrowding in Schools, NCES-2001-072
(Washington, D.C.: U.S. Government Printing Office, 2001), p. 72. NCES
defined overcrowding as enrollment more than 5 percent above the
facility's initial capacity. About half of all schools were operating at
less than capacity.

Figure 1: Total Foreign-Born Population (Legal and Illegal) in Key States,
1990 and 2000

Million 9 8 7 6 5 4 3 2 1 0

 Traditional destination states of foreign-born population States with highest
                percentage of growth in foreign-born population

                          1990 foreign-born population

                          2000 foreign-born population

Source: Census Bureau.

Note: Data or estimates for illegal aliens are not available from the
Census Bureau. Other state-bystate estimates of illegal aliens are not
included because of data reliability concerns.

States with at least 10 percent foreign born as of 2000 that are not shown
in figure 1 include Arizona, Hawaii, New Jersey, and Rhode Island. In
contrast, states such as Mississippi, Montana, North Dakota, South Dakota,
and West Virginia had 2 percent or less foreign born.

Per pupil school expenditures also vary by state. To illustrate this for
selected states, in the 1999-2000 school year, current per pupil
expenditures (exclusive of capital outlays) averaged about

o  $10,000 in New York and New Jersey

o  $8,000 in Michigan and Wisconsin

o  $7,000 in Illinois, Indiana, and Virginia

o  $6,000 in California, Florida, Georgia, Nevada, and North Carolina

o  $5,000 in Arizona and Arkansas

o  $4,000 in Utah15

Local areas can also vary in terms of both immigration concentration and
per pupil expenditures, despite equalization rules and programs in some
states.16

Two California areas serve as illustration.17 Current per pupil
expenditures in Pasadena Unified School District (near Los Angeles)
averaged about $7,000 for the 1999-2000 school year; about 24 percent of
the population in Pasadena City is foreign born. In contrast, current per
pupil expenditures in Fairfield-Suisun School District averaged about
$5,000; this school district is the largest in Solano County (located
between San Francisco and Sacramento), in which about 17 percent of
residents are foreign-born. A variety of factors affect local-area costs.
For example, among other determinants, teacher salaries may vary according
to the cost of living in different areas of a state.

Other Factors Are Related Other factors that add to the complexity of
fully accounting for variation

to Costs 	in education costs include variation in student needs and school
capacity. Even within a single school district, costs can vary among
individual students. That is, expenditures are higher for students who
live in poverty

15Table 167, Total and Current Expenditures Per Pupil in Fall Enrollment
in Public and Secondary Public Education, by Function and State:
1999-2000. NCES Digest of Education Statistics, List of Tables and Figures
for 2002, ch. 2, Elementary and Secondary Education, Revenues and
Expenditures. http://www.nces.ed.gov/programs/digest/ (May 16, 2004).
Note: Excludes capital expenditures. We report costs for the 1999-2000
school year for consistency with decennial census data. The most recent
cost data available from NCES indicate that by 2001-02, per pupil
expenditures had increased by an average of $765 for these states as a
group.

16In 1971, the California Supreme Court ruled in Serrano v. Priest, 487
P.2d 1241 (Cal. 1971), that equalization of revenue should be required
across school districts. Various other states also have equalization
programs. For a summary of "persistent challenges" to fiscal equity or
adequacy, and a list of relevant studies, see Margaret Hadderman, "Equity
and Adequacy in Educational Finance." ERIC Digest.
http://www.ericdigests.org/20021/equity.html (May 19, 2004).

17Table 91, Revenues and Expenditures of Public School Districts Enrolling
More Than 15,000 Pupils, by State: 1999-2000. NCES Digest of Education
Statistics, List of Tables and Figures for 2002, ch. 2, Elementary and
Secondary Education, Schools and School Districts.
http://www.nces.ed.gov/programs/digest/ (May 17, 2004).

or who have limited English proficiency, if schools attempt to meet these
needs (for example, with English Language Learner or ELL programs).

According to the Education Alliance at Brown University, ELL indicates

a person who is in the process of acquiring English and has a first
language other than English. Other terms commonly found in the literature
include language minority students, limited English proficiency (LEP),
English as a second language (ESL) and culturally and linguistically
diverse (CLD).18

In 2000-2001, some school districts enrolled high percentages of students
in various ELL programs. More than 40 percent of Los Angeles students and
about 33 percent of Dallas students are enrolled in ELL programs; of
course, not all students enrolled in these programs are foreign-born.19

The costs of meeting the needs of such students have been variously
estimated. Bringing ELL-enrolled children up to the grade level of sameage
non-ELL-enrolled children has been estimated to potentially increase costs
by an additional 10 to 100 percent over usual per pupil costs; for
students living in poverty (independent of ELL programs), the
corresponding range of estimates is 20 to 100 percent.20 Bringing students

18See the Education Alliance at Brown University,
http://www.lab.brown.edu/tdl/ell.shtml (May 20, 2004). NCES describes ELL
students as "born outside of the United States with a native language
other than English; or . . . from environments where the language is
predominantly non-English; or . . . American Indians and Alaska Natives
whose level of English proficiency may have been affected by a non-English
environment resulting in difficulty speaking, reading, writing, or
understanding the English language." NCES notes that ELL students were
"formerly known as Limited-English-Proficient (LEP)." See National Center
for Education Statistics, Instructions for Completing the Nonfiscal
Surveys of the Common Core of Data: School Universe Survey, Agency Use
Survey, State Nonfiscal Survey, 2003-2004 (Washington, D.C.: U.S.
Department of Education, OMB No. 1850-0067, expires Nov. 30, 2004), p. 94.

19See Beth Antunez, English Language Learners in the Great City Schools:
Survey Results on Students, Languages, and Programs (Washington, D.C.:
Council of Great City Schools, March 2003).

20See U.S. General Accounting Office, School Finance: Per-Pupil Spending
Differences between Selected Inner City and Suburban Schools Varied by
Metropolitan Area, GAO-03-234 (Washington, D.C.: Dec. 9, 2002), pp. 5-6;
William Duncombe, Anna Lukemeyer, and John Yinger, "Financing an Adequate
Education: A Case Study of New York," in William J. Fowler, ed.,
Developments in School Finance: 2001-02 (Washington, D.C.: National Center
for Education Statistics, 2003); and U.S. General Accounting Office,

Bureau of Indian Affairs Schools: Expenditures in Selected Schools Are
Comparable to Similar Public Schools, but Data Are Insufficient to Judge
Adequacy of Funding and Formulas, GAO-03-955 (Washington, D.C.: Sept. 4,
2003), p. 39.

  Objectives, Scope,
  and Methodology

characterized by both poverty and limited English proficiency up to
average levels of achievement could potentially increase average costs by
a larger amount-perhaps 30 to 200 percent over average per pupil costs.

School capacity to absorb new students is an issue because some schools
are overcrowded while others are operating below capacity. This suggests
variation across schools in the added costs associated with enrolling
additional children, such as illegal alien children.

Finally, we note that a broad array of costs additional to dollar costs
may potentially be relevant. For example, overcrowded facilities may be
related to growth in enrollments and, thus, to immigration. If sufficient
temporary space is not added, class-size-and, potentially, the quality of
education-may be affected.21

Our objectives were to address three questions regarding the potential for
estimating the costs of educating illegal alien schoolchildren K-12, by
state, using government information:

1. 	What information is currently available from state and local
governments or federal agencies on either the specific costs of educating
illegal alien schoolchildren or related topics, such as overall education
costs or the illegal alien population?

2. 	What is the Census Bureau's plan for developing new information on the
illegal alien population in the future?

3. 	What are possible approaches to estimating the costs of educating
illegal alien schoolchildren, and are government data available to support
such approaches?

The main methods we used to address these questions included:

o  	A mail survey sent to 22 states, which we selected to include (1) the
major immigration-destination states, identified in terms of numbers of

21See Ricardo R. Fernandez and P. Michael Timpane, "Bursting at the Seams:
Report of the Citizens' Commission on Planning for Enrollment Growth,"
Office of the Chancellor, Board of Education, New York City, 1995, and
Francisco Rivera-Batiz and Lillian Marti, "A School System at Risk: A
Study of the Consequences of Overcrowding in New York City Public
Schools," Institute for Urban and Minority Education, Teachers College,
Columbia University, New York, New York, 1995.

foreign-born residents; (2) the main new-growth states, which we
identified on the basis of large percentage increases in numbers of
foreign-born residents; and (3) some other states with lower percentages
of foreign-born residents (questions 1 and 3).22

o  	Telephone or e-mail inquiries to six local government or school
district officials (question 1).23

o  	Outreach to states and local areas through the National Governors
Association and the U.S. Conference of Mayors (question 1).

o  	Discussions with officials and staff at NCES, DHS (which incorporates
the former INS), and the Census Bureau, plus a collection of relevant
documents from these agencies (questions 1 and 2).24

o  	Analysis, including a review of literature and of the results from our
state survey, local area checks, and discussions with federal agencies- as
well as consultation with experts in economics and school funding
(question 3).25

In discussions with your staff, we agreed to limit the scope of this
report to (1) government estimates and government sources of information,
(2) children of illegal aliens born outside the United States, and (3)
issues concerning the estimation of gross costs, rather than attempting to

22We asked for specific information on enrollments of illegal alien
schoolchildren, the costs of educating these children, and opinions on the
factors that should be considered in estimating these costs. The 20 states
that responded were Arizona, Arkansas, California, Colorado, Illinois,
Indiana, Iowa, Kentucky, Michigan, Nebraska, Nevada, New Jersey, New York,
North Carolina, Pennsylvania, Tennessee, Texas, Utah, Virginia, and
Wisconsin. The 2 that did not respond were Florida and Georgia. We
supplemented our mailing to the 22 states with e-mails, faxes, and
telephone calls.

23We contacted local officials in (1) two traditional destinations for
immigrants (Los Angeles County, California, and New York City), (2) the
fastest growing urban area in the United States (Las Vegas, Nevada), (3)
two suburban counties (Arlington, Virginia, and Fairfax, Virginia), and
(4) one city on the Mexico-United States border (Nogales, Arizona).

24Other than NCES, DHS, and the Census Bureau, we could not identify any
federal agency as a primary source of information on illegal alien
schoolchildren, whether numbers or costs.

25These include F. Howard Nelson of the American Federation of Teachers
Educational Foundation; Frank H. Johnson, an NCES statistician; and GAO's
Chief Economist as well as GAO's Center for Economics staff.

quantify net impacts.26 Additionally, we did not consider preschool,
postsecondary, or after-school programs.

With respect to data reliability, we discussed NCES data quality
procedures with an NCES statistician specializing in NCES surveys of
school expenditures and enrollments, and we reviewed NCES's survey
documentation. We determined that the NCES data were sufficiently reliable
for use in estimating school costs.27 As we discuss in the body of the
report, other federal information concerning estimates of illegal aliens
is not as reliable.

We conducted our work in accordance with generally accepted government
auditing standards from January 2003 to May 2004.

Looking across local, state, and federal levels of government, we found
that a wealth of data are available on school enrollments and dollar
expenditures, but with few exceptions, specific information on illegal
alien schoolchildren has not been developed.

                              Currently Available

Government
Information

    State and Local Data Sources

Of the 22 state governments we surveyed, only 3 provided information on
the costs of schooling illegal alien children. Seventeen states said that
they did not have such information, and 2 states (Florida and Georgia) did
not respond.

Texas, Pennsylvania and North Carolina multiplied their states' average
current per pupil expenditures for the 1999-2000 school year by estimates
of the number of illegal alien schoolchildren. Specifically:

26We excluded from consideration the costs of educating children born in
the United States (and, therefore, U.S. citizens) to parents who were
illegal immigrants. Net costs of providing benefits to a specific
population would consider the full range of benefits received by that
population as well as taxes paid and other potential contributions.
Because we are not examining net costs, we do not address the issue of
life-cycle analysis. Barry Edmonston and Ronald Lee, eds., discuss
life-cycle analysis briefly in Local Fiscal Effects of Illegal
Immigration: Report of a Workshop (Washington, D.C.: National Academy
Press, 1996), p. 25.

27NCES data do not break out certain costs that are needed for some
approaches to cost estimation.

o  	Texas state government staff told us that they first obtained an
estimate of the number of school-aged illegal alien children younger than
18 living in Texas, from a nongovernment organization.28 The

Texas staff then multiplied this figure by the upper and lower bounds

of a range based on alternative assumptions about the percentages of

such children attending public schools-66.8 to 74.8 percent.29

o  	Pennsylvania's state government staff told us told us that they
estimated the number of illegal immigrant schoolchildren in terms of a
range by (1) accepting DHS's estimate for the resident illegal immigrant
population in their state and (2) assuming that schoolchildren represent
10 to 18 percent of this population.30 The staff then multiplied

the upper and lower bounds of the range by average per pupil

expenditures-and also specified an estimate of additional

expenditures for supplemental services such as ELL programs.

o  	North Carolina did not provide documentation of its specific methods
of estimation.

We did not evaluate the estimation procedures these states used. The
annual cost estimates that they provided to us ranged from $50 million to
$87.5 million in Pennsylvania to $932 million to $1.04 billion in Texas.31

Of the 17 states that said they did not have information on the costs of
schooling illegal alien children, 6 indicated that it would be illegal to
ask about children's immigration status. Three of these 6 mentioned the
1982 U.S. Supreme Court decision we cited earlier; these states'
interpretation

28At the request of the Texas state demographer, an immigration expert at
The Urban Institute, Washington, D.C., estimated this figure indirectly,
using census and other data. We did not evaluate that estimate.

29Texas based this range on data, indicating that, overall, 74.8 percent
of Texas children under age 18 attend school and 66.8 percent attend
public school.

30They told us that 18 percent of households in Pennsylvania have
school-age children and that they expected this figure to be lower for the
resident illegal alien population because it is likely that some families
of illegal aliens remain in their native land.

31Pennsylvania state government staff estimated a range of $38.1 million
to $66.3 million for current average per pupil expenditures ($7,772) and
possible additions to these costs of $12.2 million to $21.2 million for
ELL and other special programs. These two ranges total $50 million to
$87.5 million.

of this U.S. Supreme Court decision may be overly cautious.32 (Of course,
regardless of legal issues, a state or school district might determine
that asking about immigration status would be inadvisable or
inappropriate.)

The six local governments or school districts that we contacted did not
provide estimates of either the costs of educating or the numbers of
illegal alien schoolchildren, suggesting a lack of widespread local
government ability or effort to make such estimates. However, one school
district, on the southwest border, identified the problem of
schoolchildren who resided in Mexico and crossed the border daily or
weekly to attend U.S. schools in border areas.33

Outreach through the U.S. Conference of Mayors yielded one local area
estimate. San Jose, California, estimated the presence of approximately
20,000 illegal alien school children-noting, however, that no study had
been done. In other responding areas, various city officials said that,
given the 1982 U.S. Supreme Court decision that we mentioned above, they
did not ask about children's immigration status.34 Of course, only a
complete census of school districts could assess the number of local areas
that actually have or have not developed data or specific methodological
procedures for making estimates.

    Federal Data and Estimates

Federal data sources are also limited. NCES and the Census Bureau
routinely collect a variety of education data, including school
enrollments and dollar expenditures, with annual figures for

32The Supreme Court's decision in Plyler v. Doe, 457 U.S. 202 (1982), said
only that children may not be denied education on the basis of their
immigration status. Subsequently, a 1997 district court case (League of
United Latin American Citizens v. Wilson, 997 F.Supp. 1244 (C.D. Cal.
1997)) declared unconstitutional a provision in California's Proposition
187 that required schools to verify the status of schoolchildren because
the Court found that the intent of the requirement was to deny the
students access to a public education. Presumably, a state or local
government could inquire about legal status for another, constitutional
reason (such as to seek federal reimbursement).

33The Superintendent of Nogales Unified School District #1 informally
estimated, from anecdotal information, including observations of children
in Nogales school uniforms crossing the border, that in his school
district such children might number 400 to 600. He said that based on
Arizona's average annual per pupil expenditures of $5,000, he estimates
the annual cost of such children at about $2 million to $4 million.

34These cities include Anchorage, Alaska; Boston, Massachusetts; Dayton,
Ohio; Lincoln, Nebraska; Oklahoma City, Oklahoma; St. Petersburg, Florida;
San Leandro, California; and West Palm Beach, Florida.

o 
numbers of pupils, pupil-teacher ratios, ELL program use, and a measure of
overcrowding and

o 
current per pupil expenditures and capital expenditures for items such as
facilities acquisition, construction, and computer purchases.

NCES maintains these data and makes them available by state and by school
district or local education agency. NCES told us that staff check the
validity of these data by (1) comparing data from the school
district-level survey to data from the state-level surveys and (2)
examining trends for possible anomalies.

NCES tallies education expenditures by set categories. Thus, expenditures
cannot be broken down for specific types of programs-notably, the costs of
ELL programs are not reported separately. Also, NCES collects no data on
the number of foreign-born children enrolled in school or their
immigration status.

The Census Bureau also does not collect data on the legal status of
foreign-born persons residing in the United States. As we stated in a
previous report,

Neither [the decennial] census nor CPS [the Current Population Survey
conducted by the Census Bureau] asks about the legal status of
noncitizens-or whether they are, in fact, here illegally. There are good
reasons for this: such questions fall under the heading of `threatening'
survey questions . . . many respondents might not answer these questions
truthfully; and others might avoid participating altogether if they hear
that such questions

35

will be asked. In addition, . . . Census is concerned about privacy
invasion issues.

DHS has provided indirect, national and state-by-state estimates of the
size of the resident illegal alien population.36 DHS estimated this
population nationally at 7 million as of January 2000 and has published
additional

35GAO/GGD-98-164, p. 39.

36Immigration and Naturalization Service, "Estimates of the Unauthorized
Immigrant Population Residing in the United States: 1990-2000."
(Washington, D.C.: Jan. 2003).
http://uscis.gov/graphics/shared/aboutus/statistics/index.htm (May 21,
2004).

estimates for 42 states and the District of Columbia.37 As defined by DHS,
the illegal population excludes approximately 577,000 aliens, who
constitute several groups, including unauthorized immigrants with pending
applications for legal permanent resident status.38

As we have explained elsewhere,

o 
DHS based its 7 million estimate on two component estimates. The first
component is an estimate of 5.5 million illegal residents who entered the
United States between 1990 and 2000. The estimate was developed by using
the generally accepted residual method.39 The second component is an
estimate of 1.5 million illegal residents who entered before 1990 and were
still here, illegally, in 2000. With respect to this estimate of 1.5
million, DHS has not published an explanation of the base figure or
starting point of its calculations-3.5 million illegal immigrants here as
of 1990. DHS did specify the subsequent steps that it used to determine
how many of the 3.5 million were still here and still illegal as of 2000.
(We note that the 3.5 million base figure is consistent with other
available estimates for 1990.)40

37In addition to estimates for 42 states published in Immigration and
Naturalization Service, "Estimates of the Unauthorized Immigrant
Population Residing in the United States: 19902000," DHS provided us with
unpublished estimates for 4 more states (New Hampshire, South Dakota, West
Virginia, and Wyoming), rounded to the nearest thousand. DHS told us
estimates for the 4 remaining states (Maine, Montana, North Dakota, and
Vermont) rounded to zero.

38There are about (1) 200,000 unauthorized immigrants who have submitted
applications for lawful permanent resident alien status that are pending
and likely to be approved and (2) 377,000 aliens in various groups. The
latter include aliens, mostly from Central America, who otherwise would be
unauthorized residents but are allowed to remain and work in the United
States under various legislative provisions or court rulings, such as
people with Temporary Protected Status (TPS), asylees, and parolees. TPS
derives from the U.S. Attorney General's designating foreign nationals
eligible for temporary refuge. "Asylee" refers to an alien granted asylum,
"parolee" to an alien otherwise inadmissible but allowed to enter
temporarily for urgent humanitarian reasons or significant public benefit.

39The residual method consists of (1) deriving estimates of legal
immigrants, based primarily on DHS's own administrative data, and (2)
subtracting these estimates of legal immigrants from the census figure for
the total number of aliens residing here. DHS's calculations were
completed separately for different years of arrival in the United States;
that is, census data include a report of the year in which each
foreign-born individual came to the United States to live, and DHS
administrative data indicate the year in which each legal immigrant
entered.

40See U.S. General Accounting Office, Overstay Tracking: A Key Component
of Homeland Security and a Layered Defense. GAO-04-82. (Washington, D.C.,
May 21, 2004), app. III.

o 
By definition, DHS's estimates of the resident illegal immigrant
population do not include short-term illegal aliens-for example, persons
here for weeks or months and not likely to be included in the decennial
census (or corrections for undercounts).41

o 
DHS has stated that its national estimates are marked by uncertainty-
noting, for example, "actual trends might be somewhat higher or lower than
those shown"-but has not developed ranges to characterize this
uncertainty.42

DHS's published state-by-state estimates of illegal immigrant residents
cover 42 states and the District of Columbia. These estimates are based,
in part, on comparing legal immigrants' statements about their intended
destinations, as reflected in INS administrative records when they were
admitted to the United States, to decennial census geographic
distributions for the total population of foreign-born residents.43
However, these state-level estimates are uncertain to the extent that
legal immigrants moved to different states from the states of their
intended destinations, at any time before the 2000 census-and some did
move to a different state.44

Information not provided by DHS is also problematic. First, the DHS paper
presenting these indirect estimates does not include ranges so that other
analysts can gauge the degree of uncertainty-and hence their suitability
for specific purposes. For example, DHS does not indicate whether some
state estimates may be highly uncertain. Second, DHS has not published

41Similarly, DHS's estimates do not, by definition, include children who
reside in Mexico and cross the border daily or weekly to attend U.S.
schools in border areas.

42DHS's paper, "Estimates of the Unauthorized Immigrant Population
Residing in the United States: 1990-2000," discusses limitations of its
data, including the complexity of estimating numbers of persons residing
here with legal temporary visas (such as temporary workers and students),
and so forth. The paper does not, however, attempt to validate assumptions
about emigration (how many legal and illegal immigrants leave the United
States).

43DHS calculates state estimates of illegal immigrants by, essentially,
subtracting state-level estimates of legal immigrants-based on these
immigrants' statements about their intended state of residence, on their
applications-from the total number of foreign-born persons residing in
each state, based on the 2000 census.

44See, for example, Marc J. Perry and Jason P. Schachter, Migration of
Natives and the Foreign Born: 1995 to 2000, CENSR-11 Census 2000 Special
Reports (Washington, D.C.: U.S. Census Bureau, August 2003).

any description of analyses conducted to validate its national or state
estimates. Third, DHS's estimates of illegal immigrants do not break out
age groups. This is important because it is not known whether the age
distribution-and thus percentage of school-age children-for the illegal
population is or is not different from that for other groups.45

  The Census Bureau's Plans for Estimating Illegal Immigrants Are Preliminary

The Census Bureau is developing a research plan aimed at eventually
developing new information on the population of illegal immigrants
residing in the United States.46 Census Bureau staff told us that the
objective of the plan is to produce annual estimates of international
migrants according to four statuses: (1) legal immigrants, (2) temporary
migrants, (3) quasi-legal migrants, and (4) unauthorized migrants, "in
order to better inform . . . population estimates." The plan will be based
on an indirect approach that does not involve asking census or survey
questions about immigration status.

Importantly, the Census Bureau anticipates that the new information will
include indirect estimates of illegal aliens in various age groups-perhaps
coming close to identifying the number of resident illegal alien children
who are of school age.47 The Census Bureau staff said that it is not yet
known how many age groups can be reliably broken out. For example, it
might be possible to use 5-year age groups or perhaps five main age groups
(0-17, 18-29, 30-49, 50-64, and 65 and older), but Census Bureau staff
told us that if this is not possible, perhaps only three age groups would
be estimated (0-17, 18-64, and 65 and older).

45A possible additional issue regarding information not provided concerns
the fact that illegal alien schoolchildren may include some
nonresidents-that is, Mexican children who cross the border each school
day to attend U.S. schools. DHS's estimates of the resident illegal alien
population would, by definition, not include schoolchildren who reside
outside the United States.

46Charles L. Kincannon, Director of the Census Bureau, letter to F. James
Sensenbrenner Jr., Chairman of the Committee on the Judiciary, U.S. House
of Representatives, U.S. Congress, Washington, D.C., Aug. 20, 2003. The
preliminary work necessary to carry out this plan was mentioned in a
presentation to a Census Bureau advisory group (see Kevin Deardorff,
"Immigration Research Update," presentation to Census Information
Conference Steering Committee, Feb. 26, 2004).

47By definition, such an estimate would not include schoolchildren who
reside in Mexico and cross the border daily or weekly to attend U.S.
schools in border areas.

While an exact methodology has not yet been specified, the Census Bureau
previously experimented with models for imputing temporary legal status
(for example, possessing a valid temporary work or student visa) to groups
of foreign-born survey respondents.48 This effort used administrative and
other data to suggest the potential numbers of foreignborn residents with
temporary legal status-by age group, as well as by race and country of
birth. The age groups were defined as 0-17, 18-29, 30- 49, 50-64, and 65
and older.

Census Bureau analysts reason that a similar process could be developed
for group imputation of lawful permanent resident status (that is, green
card status).49 If this approach were expanded to cover virtually every
legal immigration status, then the numbers and characteristics of illegal
alien residents could also be estimated, based on a modified residual or
subtraction approach.

The Census Bureau intends to use its new indirect procedures on data from
the new American Community Survey (ACS), scheduled for full implementation
later this year. A variety of other information sources may be tapped. The
Census Bureau further anticipates that

o 
the new information on age groups nationally will be available in 2007,
with estimates for states-and perhaps key counties--potentially available
by 2009 (pending budget approvals needed for the geographic estimates);

o 
key estimates will be made annually, and other estimates (such as
estimates for smaller groups) will be made with 3-year or 5-year
averages;50

48Rachel Cassidy and Lucinda Pearson, "Evaluating Components of
International Migration: Legal Temporary Migrants," Working Paper 60,
Population Division, Census Bureau, January 2002. (Census Bureau working
papers report the results of research undertaken by Census Bureau staff
and undergo a more limited review than official Census Bureau
publications.)

49Characteristics of persons with green cards might also be based on
survey data on greencard holders developed by using one legal-status card
from the "three-card method." See U.S. General Accounting Office, Survey
Methodology: An Innovative Technique for Estimating Sensitive Survey
Items, GAO/GGD-00-30 (Washington, D.C., Nov. 1999), pp. 1 and 5.

50Thus, the size of the group for which an estimate is derived, as well as
the specific estimation methodology, will determine the time period for
initial estimates and trends. For example, trends based on comparing
3-year averages for two points in time with no overlap would not be
available until 2011.

o 
the new estimates will include ranges to indicate a margin of uncertainty;
and

o 
an assessment of the quality of the new estimates will be conducted,
perhaps using an expert panel.

Census Bureau staff told us that its plan does not include estimating the
number of illegal alien children who attend public school or estimating
the children's English proficiency. However, Census Bureau staff suggested
to us that, potentially, other agencies might develop such estimates by
combining the Census Bureau's state-by-state age group estimates of
illegal aliens with other information, assumptions, or models or a
combination of them.

The ACS asks about English proficiency and school attendance. Although
administrative data on legal immigrants do not include this information,
the New Immigrant Survey (NIS)-sponsored by DHS and the National Institute
of Child Health and Human Development, in partnership with other federal
agencies-asks a sample of new green-card holders about their English
proficiency and their school attendance.51

We believe that Census Bureau analysts face a number of technical
challenges in developing reasonably valid and reliable information.
Challenges include (1) assessing coverage of the illegal alien population
in the ACS by age group, as well as coverage of legal foreign-born
residents; (2) developing adequate sources of information to calibrate and
validate a model for estimating immigration status; and (3) assessing and
reporting the levels of uncertainty associated with the estimates.

Additionally, developing this new information resource will depend on
continued funding approvals. Specifically, the Census Bureau's budget
request for 2005 indicates that increased funding (more than requested for

51The NIS asks for the name of the school currently attended rather than
whether the school is public or private. (The NIS is described briefly at
www.pop.upen.edu/nis (May 22, 2004).) Information on a broader spectrum of
foreign-born children legally residing in the United States (for example,
information on their school attendance or their English language
proficiency) might also be developed by using data gathered in a survey
using one or more legal status cards from the three-card method
(GAO/GGD-00-30).

  Approaches to Cost Estimation

2005) will be needed for work during fiscal years 2006 to 2008 to begin
developing migration estimates at the state and local levels.52

Because the Census Bureau's plan is in an early stage of development
(detailed documents are not yet publicly available), it is not yet
possible to judge its quality. However, when more details become
available, the Census Bureau's plan might be compared to DHS's methodology
to determine whether the Census Bureau's approach will be likely to avoid
the weaknesses associated with DHS's.53

We identified four approaches to estimating the dollar costs of illegal
alien schoolchildren by state. These four approaches to estimating dollar
costs are not necessarily mutually exclusive; that is, some are
refinements of, or may be used together with, one or more of the others.54

The four approaches are as follows:

1. state-by-state multiplication;

2. local-area refinement of state-by-state multiplication;

3. student-needs refinement of state-by-state multiplication; and

4.
capacity-based estimation of incremental costs, also a refinement of
state-by-state multiplication.

    State-by-State Multiplication Approach

The first approach consists of multiplying current average per pupil
expenditures in each state by an indirect estimate of the number of
illegal alien schoolchildren in that state, as illustrated in fig. 2. This
approach has

52Census Bureau staff told us these plans are documented in Exhibit 13 of
the President's Fiscal Year 2005 Budget. (Exhibit 13 is about the
"Measuring Migration Across U.S. Borders" program of the Intercensal
Demographic Estimates Sub-activity.)

53We note, however, that the Census Bureau's plan to estimate the resident
illegal alien population would (like DHS's estimates) exclude short-term
illegal aliens who are not likely to be included in the decennial census,
or corrections for undercounts. Also, the Census Bureau's estimates would
not include illegal alien schoolchildren residing in Mexico who cross the
border to attend school.

54Estimation of a broad array of potential costs, including overcrowding
and possible reductions in education quality, represents an alternative
direction in assessing cost impacts.

been used in the past to estimate the costs in some states of educating
illegal alien schoolchildren.55

Figure 2: Simplest State-by-State Estimation Approach

Sources: GAO (analysis), NCES (per pupil expenditure data), Census Bureau
(estimation plans), U.S. Department of Education (photo).

A variant of this approach might take account of capital expenditures (for
example, add per pupil capital outlays to per pupil current
expenditures).56

o 
The main advantage of state-by-state multiplication is its simplicity: If
input data are available, cost figures are easily calculated.
Additionally, this approach seems logical and is easily understood.

55For detailed examples of using current average per pupil expenditures
(and the effects of various assumptions about how to define such
expenditures) to estimate states' costs of educating illegal aliens, see
Rebecca L. Clark, and others, "Costs of Providing Public Primary and
Secondary Education to Undocumented Aliens in Seven States," in Fiscal
Impacts of Undocumented Aliens: Selected Estimates for Seven States
(Washington, D.C.: The Urban Institute, September 1994), pp. 61-89, and
descriptions of estimates reviewed in U.S. General Accounting Office,
Illegal Aliens: Assessing Estimates of Financial Burden on California,
GAO/HEHS-95-22 (Washington, D.C.: Nov. 28, 1994), and Illegal Aliens:
National Net Cost Estimates Vary Widely, GAO/HEHS-95-133 (Washington,
D.C.: July 25, 1995).

56A National Academy of Sciences' (NAS) workshop discussed in Edmonston
and Lee, Local Fiscal Effects of Immigration, raised the issue of
including capital costs. Although NCES cautions that annual figures may
vary considerably from year to year, one solution may be to use a summary
figure, such as a 5-year average. The NAS workshop, however, noted "the
applicability of average cost or marginal cost may be time sensitive,"
depending on whether a school system has "significant excess student
capacity" or becomes overcrowded (p. 16).

o 
Its limitations are that it does not account for the many complexities of
costs in this area-which, as we outlined in the background section,
include local-area cost differentials, variation in individual student
needs, differences in school capacity, and potentially other factors.57

Considering its advantages and disadvantages, we believe that
state-bystate multiplication is a logical approach that could yield a
rough approximation of state-by-state costs. As we outlined above,
government data or estimates needed for using this approach are not
currently available. Although NCES provides data on per pupil expenditures
by state, DHS's state-by-state estimates of the illegal immigrant
population do not break out age groups. Moreover, DHS's state-by-state
estimates have methodological limitations that make them uncertain, yet
DHS has not estimated ranges or otherwise characterized the degree of
uncertainty associated with specific estimates.

In the future, the Census Bureau may provide age-group estimates of
illegal immigrants by state.

Although there are a number of challenges, if state-level estimates of
illegal alien children are successfully developed and validated then
reasonably reliable (if somewhat rough) approximations of state-by-state
costs could be calculated, using the state-by-state multiplication
approach. This would be especially the case if others are able to refine
this information by estimating illegal alien children's school attendance.

    Local-Area Refinement of the State-by-State Multiplication Approach

The local-area refinement approach is a logical extension of
state-by-state multiplication and attempts to generate greater precision
by taking account of local variations in school costs.58 Multiplications
would be conducted separately for different areas-for example, cities,
counties, or groups of counties-and then summed across all areas of the
state. Thus,

57For example, the diversity (versus homogeneity) of languages and
backgrounds of illegal immigrant children enrolled in a particular school
district may affect costs.

58In responding to our survey, New York (State Department of the Budget,
Education Unit) pointed to geographic cost differences within states,
noting that "If the unauthorized foreign-born students are clustered in
the regions of the state where education is more expensive, as is most
likely the case in large states like New York and California, using the
average cost in the state as a whole will underestimate the true cost of
educating these students." NAS noted cost variations across geographic
areas within states (Edmonston and Lee, Local Fiscal Effects of Illegal
Immigration, p. 14).

if large numbers of illegal alien schoolchildren attend schools with
higher or lower costs than the state average, this would be reflected in
the cost estimates. This approach is appropriate for states with
significant immigrant populations that exhibit local-area variation in
both school costs and numbers of illegal alien schoolchildren.

o 
The local-area refinement approach has the same advantages as
stateby-state multiplication, and it uses more specific local-area data.

o 
Its limitations are suggested by a review of other methods (discussed
below) that account for differences in student needs or school capacity,
each of which can affect costs.

NCES collects and maintains current per pupil expenditure and capital
outlay data, which are available at the school district level, but
government information on illegal alien schoolchildren is not available.
In the future, the Census Bureau's plan may, if successfully implemented,
help quantify the population of school-age illegal alien children in key
counties.

    Student-Needs Refinement of the State-by-State Multiplication Approach

The student-needs refinement of the state-by-state multiplication approach
can build on either of the two approaches we described above, by
accounting for specific individual student needs that may affect education
costs. As we indicated in the background section, costs may be higher for
children in poverty or with limited English proficiency than for other
children-and both poverty and limited English proficiency may characterize
many illegal alien children. Moreover, of the 20 states responding to our
survey, 5 indicated that because of special needs (such as for ELL
programs), efforts to estimate the costs of educating illegal alien
children should consider these additional cost factors.59 Using a
costfunction approach to statistical analysis allows estimation of cost
factors associated with meeting the needs of students with specific
characteristics.60

59The NAS workshop also pointed to program-specific variations in costs,
for example, ELL (Edmonston and Lee, Local Fiscal Effects of Illegal
Immigration, p. 15).

60See, for example, William Duncombe, "Estimating the Cost of an Adequate
Education in New York," CPR Working Paper 44, Center for Policy Research,
Maxwell School, Syracuse University, Syracuse, New York, February 2002.
http://wwwcpr.maxwell.syr.edu/faculty/duncombe/special%20report/specialreport.htm
(May 22, 2004).

o 
The main advantage of this approach is that by focusing on individual
students, it takes account of more determinants of cost than the other
approaches outlined above.

o 
The main disadvantage is the complexity involved in estimating cost
factors, as well as requirements for data that may not be available.
Additionally, this approach does not account for varying school capacity
(see below).

There are two key issues in implementing the student-needs-refinement
approach:

1.
It requires data based on separate cost accounting for programs designed
to meet special student needs. NCES state-by-state and school-district
data do not include separate information on the cost of either ELL
programs or programs designed to compensate for poverty (and associated
learning disadvantages). However, information on revenues from federal and
state sources for related programs (for example, bilingual education) are
collected in the Annual Survey of Local Government Finances-School
Systems, conducted by the Census Bureau.

2.
It also requires estimates of the English proficiency and the poverty
status of illegal alien school children. The Census Bureau's plan does not
envision such estimates, but Census Bureau staff told us that other
government agencies or researchers could expand on the information the
Census Bureau provides by, for example, using additional assumptions and
models to achieve indirect estimates of illegal alien children with
limited English proficiency.61

School-Capacity Approach
The incremental cost of adding a student to a classroom, school, or school
district may be much higher or lower than the average per pupil
expenditure, depending on the capacity of the classroom, school, or
district in question. To illustrate, adding one student to a classroom and
school that has the capacity to easily accommodate the student would not

61Refer back to the earlier section entitled The Census Bureau's Plans for
Estimating Illegal Immigrants Are Preliminary.

require additional capital outlays, such as building a new facility.62 In
contrast, if that school were at or above capacity, the reverse might be
true.63

Various experts have suggested that-at least in some circumstances-a
capacity-based approach to cost estimation may produce more meaningful
results than the other approaches. For example, an NAS workshop on the
fiscal effects of immigration indicated that "Over time, [education] costs
are a function of the capacity utilization level, not simply the number of
additional children."64

In general, three cost categories can be distinguished, as shown in table
1. Category 1 costs do not vary with enrollment increases or decreases.
Category 2 costs tend to change in a predominantly linear fashion when
students are added. Category 3 costs, which may be related to growth in
enrollment, tend to increase or decrease in step jumps rather than in a
linear fashion; these may be more difficult to predict.

62Consistent with this is the notion that when enrollments increase,
average per pupil expenditures may initially decline. Of course, if more
students are added, a point may come at which-even for a school with
initial excess capacity-one or more teachers must be added, overcrowding
will occur, and additional classroom space must be acquired or new school
must be built.

63An example is given in Brad R. Humphreys, "A Report on Incremental Costs
and Benefits Associated with Increasing Enrollment at UMBC," Department of
Economics, University of Maryland Baltimore County, Baltimore, Maryland,
May 17, 2003. In preparing this report, we conducted a preliminary
analysis of education costs in new-growth immigration states (including
the 5 states on the right in fig. 1, and 5 other states). We found that
per pupil expenditures in these 10 new-growth states increased by a larger
amount than in other states, taken as a group. However, we could not
conclude that there was a connection between immigration and the change in
per pupil expenditures, because there was considerable overall growth in
these states' total populations. And even after high percentage growth in
the foreign-born populations in these states, foreign-born persons
remained a relatively small percentage of these states' total populations.

64Edmonston and Lee, Local Fiscal Effects of Immigration, p. 16.

Table 1: Cost Categories Relevant to the School-Capacity Approach

Cost category Cost example

1. Fixed costs  o  maintaining a school board

o  salary of the superintendent of schools

2. Variable costs that correspond directly to the number of students  o 
new textbooks  o  other materials purchased for each student

3. Costs incurred when capacity is expanded  o  adding teachers

o  adding on to an existing school or building a new school

Source: GAO analysis.

For a classroom with excess capacity, the only costs associated with
adding a single student would be category 2 costs.65 Where capacity is
limited or already stretched, both category 2 and category 3 are relevant.
Calculating category 3 costs is difficult. However, studies to explore the
incremental costs of illegal alien children might be approached by
developing matched pairs of school districts-that is, school districts
that appear to have many relevant characteristics in common but that
differ with respect to enrollment of illegal alien schoolchildren over
time.

o  	The advantage of the school-capacity approach is that it reflects
important classroom and school factors for which the other approaches do
not account.

o  	The main disadvantages are its complexity and requirements for
extensive, specific data that may not be available.

Separate data for each cost category in table 1 are not necessarily easily
accessed. Although some expenditures within a category (such as paper
supplies, books, and periodicals in category 2) can be broken out in NCES
data, the costs of adding teachers because of expanding enrollments could
probably be estimated only by undertaking a special study, such as that
suggested above. In the future, the Census Bureau may be able to supply
the data needed to (1) help identify candidate counties or districts for
such studies and (2) track trends in the estimated numbers of illegal
alien schoolchildren in these areas over time. NCES school-expenditure
data could then be compared over time-ideally, before and after a period
of

65In fact, the slightly larger class size would, all other things being
equal, tend to result in a decline in average per pupil expenditures.

  Observations

Agency Comments

substantial growth in the illegal alien enrollments for one of the two
school districts.

Considering our findings, we believe that the government information that
is available is not sufficient to reliably quantify the costs of educating
illegal alien schoolchildren. All approaches to estimating these costs
require data or estimates of the number of illegal alien schoolchildren.
Neither state nor local governments collect this information, and federal
agencies do not provide estimates.

Although DHS estimates the resident illegal immigrant population, its
estimates are subject to unspecified levels of uncertainty; further, DHS
estimates do not break out age groups. The Census Bureau plans to develop
age-group estimates, but it is too early to evaluate its plans. If
successfully implemented, the plan calls for national estimates to be made
available by 2007 with state and perhaps local estimates by 2009.

If more data on the numbers and characteristics of illegal alien
schoolchildren were to become available, then it would be appropriate to
conduct an in-depth analysis of the methodological alternatives that could
be supported by those data. If feasible, analyses aimed at producing
comparative estimates, based on more than one approach, could shed
considerable light on the issue of education costs. They could do so by
indicating not only the expenditure levels based on the numbers of illegal
alien pupils and the average per pupil expenditures by state or local
area. They could also do so by providing information on whether estimated
expenditures are appreciably different when adjusted to reflect these
students' needs or the capacity of the schools they attend or both.

We provided a draft of this report to the Census Bureau, DHS, and NCES.
The agencies informed us that they had no formal comments, but we received
informal comments from the Census Bureau and NCES on minor technical
points and made changes in the report as appropriate.

As agreed with your office, we will be sending copies of this report to
the Secretary of the Department of Homeland Security, the Director of the
National Center for Education Statistics, the Director of the Census
Bureau, and various congressional committees. We will also make copies
available to others on request. In addition, the report will be available
at no charge on the GAO Web site at http://www.gao.gov.

If you or your staff would like to discuss any of the issues we present
here,
please call me at (202) 512-2700 or Judith A. Droitcour, who served as
project director on this study, at (202) 512-9145. Other major
contributors
to this report include Eric M. Larson, Mona H. Sehgal, Seyda Wentworth,
and Timothy Carr.

Sincerely yours,

Nancy R. Kingsbury,
Managing Director
Applied Research and Methods

  GAO's Mission

Obtaining Copies of GAO Reports and Testimony

The General Accounting Office, the audit, evaluation and investigative arm
of Congress, exists to support Congress in meeting its constitutional
responsibilities and to help improve the performance and accountability of
the federal government for the American people. GAO examines the use of
public funds; evaluates federal programs and policies; and provides
analyses, recommendations, and other assistance to help Congress make
informed oversight, policy, and funding decisions. GAO's commitment to
good government is reflected in its core values of accountability,
integrity, and reliability.

The fastest and easiest way to obtain copies of GAO documents at no cost
is through the Internet. GAO's Web site (www.gao.gov) contains abstracts
and fulltext files of current reports and testimony and an expanding
archive of older products. The Web site features a search engine to help
you locate documents using key words and phrases. You can print these
documents in their entirety, including charts and other graphics.

Each day, GAO issues a list of newly released reports, testimony, and
correspondence. GAO posts this list, known as "Today's Reports," on its
Web site daily. The list contains links to the full-text document files.
To have GAO e-mail this list to you every afternoon, go to www.gao.gov and
select "Subscribe to e-mail alerts" under the "Order GAO Products"
heading.

Order by Mail or Phone 	The first copy of each printed report is free.
Additional copies are $2 each. A check or money order should be made out
to the Superintendent of Documents. GAO also accepts VISA and Mastercard.
Orders for 100 or more copies mailed to a single address are discounted 25
percent. Orders should be sent to:

U.S. General Accounting Office 441 G Street NW, Room LM Washington, D.C.
20548

To order by Phone: 	Voice: (202) 512-6000 TDD: (202) 512-2537 Fax: (202)
512-6061

To Report Fraud,	Contact: Web site: www.gao.gov/fraudnet/fraudnet.htm

  Waste, and Abuse in E-mail: [email protected]

Federal Programs Automated answering system: (800) 424-5454 or (202)
512-7470

Jeff Nelligan, Managing Director, [email protected] (202) 512-4800

Public Affairs 	U.S. General Accounting Office, 441 G Street NW, Room 7149
Washington, D.C. 20548
*** End of document. ***