Law Enforcement Support Center: Name-Based Systems Limit Ability to
Identify Arrested Aliens (Letter Report, 08/21/95, GAO/AIMD-95-147).

Pursuant to a congressional request, GAO reviewed Immigration and
Naturalization Service (INS) initiatives for identifying which arrested
individuals are aliens, focusing on whether: (1) the Law Enforcement
Support Center (LESC) can use existing databases to identify as aliens
individuals arrested for aggravated felonies; (2) other INS initiatives
will allow identification of aliens arrested for aggravated felonies;
and (3) criminal alien information in two INS databases is complete and
accurate.

GAO found that: (1) LESC allows law enforcement agencies continuous
access to INS data, but LESC electronic searches cannot conclusively
identify aliens arrested for aggravated felonies, since name-based data
can be easily falsified; (2) LESC has initiated enforcement actions on
1,935 aliens, but released 920 additional aliens for various reasons;
(3) at least 46 of the 920 aliens released had been arrested for
aggravated felonies; (4) INS is planning to fully implement its INS
Identification System in 1999 to provide a unique and effective
identifier for aliens encountered for enforcement or benefit purposes,
but it will include only known criminal aliens currently in INS
databases; (5) INS has recently implemented an initiative to identify
criminal aliens entering the country illegally, based on fingerprint
data for criminal aliens deported from California; and (6) INS omitted
important criminal alien information from certain databases, causing
them to be incomplete and inaccurate.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  AIMD-95-147
     TITLE:  Law Enforcement Support Center: Name-Based Systems Limit 
             Ability to Identify Arrested Aliens
      DATE:  08/21/95
   SUBJECT:  Data integrity
             Illegal aliens
             Criminals
             Arrests
             Data bases
             Law enforcement
             Intergovernmental relations
             Computerized information systems
             Immigration or emigration
             Crimes or offenses
IDENTIFIER:  INS Deportable Alien Control System
             INS Central Index System
             California
             INS National Automated Immigration Lookout System
             INS Nonimmigrant Information System
             INS Student and Schools System
             Phoenix (AZ)
             Maricopa County (AZ)
             INS Identification System
             National Law Enforcement Telecommunications System
             Interstate Identification Index
             INS Alien Documentation, Identification, and 
             Telecommunications System
             California Criminal History System
             
**************************************************************************
* This file contains an ASCII representation of the text of a GAO        *
* report.  Delineations within the text indicating chapter titles,       *
* headings, and bullets are preserved.  Major divisions and subdivisions *
* of the text, such as Chapters, Sections, and Appendixes, are           *
* identified by double and single lines.  The numbers on the right end   *
* of these lines indicate the position of each of the subsections in the *
* document outline.  These numbers do NOT correspond with the page       *
* numbers of the printed 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.                                          *
*                                                                        *
* A printed copy of this report may be obtained from the GAO Document    *
* Distribution Facility by calling (202) 512-6000, by faxing your        *
* request to (301) 258-4066, or by writing to P.O. Box 6015,             *
* Gaithersburg, MD 20884-6015. We are unable to accept electronic orders *
* for printed documents at this time.                                    *
**************************************************************************


Cover
================================================================ COVER


Report to the Ranking Minority Member, Subcommittee on National
Security, International Affairs and Criminal Justice, House Committee
on Government Reform and Oversight

August 1995

LAW ENFORCEMENT SUPPORT CENTER -
NAME-BASED SYSTEMS LIMIT ABILITY
TO IDENTIFY ARRESTED ALIENS

GAO/AIMD-95-147

Law Enforcement Support Center


Abbreviations
=============================================================== ABBREV

  CAL - California
  CAL/DOJ - California Department of Justice
  CIS - Central Index System
  DACS - Deportable Alien Control System
  DOJ - Department of Justice
  ENFORCE - Enforcement Case Tracking System
  FBI - Federal Bureau of Investigation
  IDENT - INS Identification System
  IIR - Institute for Intergovernmental Research
  IRM - information resource management
  INS - Immigration and Naturalization Service
  LEA - law enforcement agency
  LESC - Law Enforcement Support Center
  NAILS - National Automated Immigration Lookout System
  NCIC - National Crime Information Center
  NIIS - Nonimmigrant Information System
  NLETS - National Law Enforcement Telecommunications System
  STSC - Student and Schools System

Letter
=============================================================== LETTER


B-260745

August 21, 1995

The Honorable Karen L.  Thurman
Ranking Minority Member, Subcommittee on
 National Security, International Affairs
 and Criminal Justice
Committee on Government Reform and Oversight
House of Representatives

Dear Ms.  Thurman: 

This report responds to the former chairman's request that we review
current Immigration and Naturalization Service (INS) initiatives for
identifying which arrested individuals are aliens and evaluate the
reliability of INS' criminal alien\1

information.  The Anti-Drug Abuse Act of 1988 (Public Law 100-690)
required the Attorney General to devise and implement a system to
make available to federal, state, and local authorities, on a 24-hour
basis, the investigative resources of the INS to determine whether
individuals arrested for aggravated felonies\2 are aliens.  To meet
this requirement, INS established the Law Enforcement Support Center
(LESC), whose pilot operations began in July 1994.  When individuals
arrested for aggravated felonies identify themselves as being
foreign-born, the local law enforcement agency (LEA) sends a request
to LESC to determine whether these individuals are aliens\3 and,
thus, possibly subject to deportation.  LESC staff then query five
existing INS databases to determine if the individual arrested is an
alien and send a response to the LEA. 


Based on the former chairman's request and subsequent agreements with
his office, we determined whether

  LESC, using INS' existing name-based\4 databases, can identify
     individuals arrested for aggravated felonies as aliens;

  other INS initiatives will allow identification of aliens arrested
     for aggravated felonies; and

  criminal alien information in two of INS' databases is complete and
     accurate. 


--------------------
\1 For purposes of this report, the term "criminal alien," refers to
any alien who has been convicted of certain serious crimes in the
United States.  Under provisions of the Immigration and Nationality
Act
(8 U.S.C.  S 1251, 1252(a)), aliens who are convicted of certain
serious crimes including aggravated felonies, crimes of moral
turpitude, multiple crimes, and drug and firearms offenses may be
taken into custody and deported, typically following the end of the
alien's incarceration for the underlying sentence. 

\2 Aggravated felonies include murder, narcotics and firearms
trafficking, money laundering, crimes of violence, and other serious
crimes such as alien smuggling and document fraud. 

\3 For purposes of this report, the term alien refers to foreign-born
persons who are not naturalized citizens of the United States.  Legal
aliens can be deported if convicted of serious crimes.  LEA cannot
assume that an individual is foreign-born based on the individual's
name or appearance.  Individuals not stating that they are
foreign-born are considered by LEAs to be citizens. 

\4 INS' case files are established using name and date of birth
provided by the alien. 


   RESULTS IN BRIEF
------------------------------------------------------------ Letter :1

The LESC pilot allows LEAs access to INS data 24 hours a day. 
However, because of limitations inherent in INS' name-based
databases, LESC electronic searches cannot conclusively identify
individuals arrested for aggravated felonies as aliens.  These
searches match the name and date of birth provided by the arrested
individual with those of aliens in the databases.  Unlike
fingerprints, which uniquely identify an individual, these data may
be shared by more than one individual and can be easily falsified. 
Thus, the LESC electronic search can only identify possible matches,
which require an INS investigator to conduct an additional
investigation to conclusively determine whether the arrested
individual is an alien.  Using this process during the pilot, INS
initiated enforcement actions on 1,935 aliens.  However, the LEAs
released 920 additional aliens that INS could have taken enforcement
actions against because INS did not identify them before bond was
posted or the aliens were released on their own recognizance by the
LEAs.  The INS investigator assigned to the pilot estimated that at
least 46 of the 920 were arrested for aggravated felonies--the
population targeted by the mandate. 

Two other initiatives offer possibilities for identifying aliens
arrested for aggravated felonies.  The first initiative is the INS
Identification System (IDENT).  According to INS officials, it is to
be fully implemented in 1999 and is intended to provide a unique
identifier for aliens encountered for enforcement or benefit
purposes, thus allowing a more efficient means of identifying these
aliens.  However, IDENT is to include only (1) known criminal aliens
currently in INS' databases and (2) those aliens for whom INS will
process benefit or enforcement actions after IDENT's implementation. 
IDENT will not include noncriminal aliens currently in INS'
databases, students, tourists, or business people. 

The second initiative is limited to California and to criminal aliens
previously deported from that state.  Under this initiative, INS
provides fingerprint cards of previously deported criminal aliens to
the California Department of Justice (CAL/DOJ), which reads them into
the state automated fingerprint processing system.  When there is a
match, the person's state criminal history file is flagged with a
criminal alien alert.  If these persons are arrested after reentering
the country illegally, the LEAs can immediately identify them as
deportable aliens and contact INS. 

Criminal alien information in INS' Deportable Alien Control System
(DACS) database and the corresponding Central Index System (CIS)
files is incomplete and inaccurate.  According to our statistical
sample of criminal aliens recorded in DACS, important information
contained in INS paper files was missing from, or incorrectly entered
into, the DACS electronic files.  For example, 80 percent of the
records in our sample did not contain all aliases used by the aliens,
and 22 percent contained either misspelled names, incorrect name
order, or incorrect nationality.  Furthermore, DACS did not contain
records of all the paper files it should have included, and some of
the criminal alien files it did contain lacked the information needed
to show that these persons were criminal aliens. 


   BACKGROUND
------------------------------------------------------------ Letter :2

INS, a component of the Department of Justice (DOJ), is responsible
for administering the immigration and naturalization laws relating to
the admission, exclusion, deportation, and naturalization of aliens. 
Under the Immigration and Nationality Act, the Attorney General has
sole authority to determine alien status and to initiate deportation
proceedings.  The act also authorizes the Attorney General to
apprehend and deport aliens who have been convicted of serious
crimes, including aggravated felonies; crimes of moral turpitude;
multiple crimes; and drug and firearm offenses.  These apprehension
and deportation authorities have been delegated to INS. 


      INS DATABASES
---------------------------------------------------------- Letter :2.1

INS has five electronic databases containing the primary information
it relies upon for its day-to-day operations.  The databases are as
follows: 

  The Central Index System (CIS) is the central file for all aliens
     with whom INS has had contact except nonimmigrant\5 aliens.  It
     also contains cross references to other databases in which
     individuals have files. 

  The Deportable Alien Control System (DACS) contains data on all
     aliens who currently are or have been in deportation
     proceedings, including criminal aliens.  DACS also contains
     information on aliens who have been apprehended upon entering
     the country illegally and returned involuntarily to their
     countries.  DACS supports INS' enforcement activities. 

  The National Automated Immigration Lookout System II (NAILS II) is
     a lookout enforcement system which contains data about persons
     of interest to INS, including aliens who are suspected of
     illegal activities and/or have been previously deported.  NAILS
     II is used by inspectors at various ports of entry throughout
     the country. 

  The Nonimmigrant Information System (NIIS) contains arrival,
     departure, and ancillary information pertaining to nonimmigrant
     aliens entering the United States legally on tourist or business
     visas. 

  The Student and Schools System (STSC) is the primary vehicle for
     identifying, locating, and determining the status or benefit
     eligibility of nonimmigrant students and their dependents. 

Appendix I provides a more detailed description of the databases. 
All alien case files in the databases, with the exception of NIIS and
STSC, have corresponding paper files, which contain information on
all service and, if applicable, enforcement actions that INS has
taken or is taking.  The paper file for criminal aliens generally
includes a set of fingerprints, the Federal Bureau of Investigation
(FBI) criminal history, and other identifying information, such as a
photograph.\6


--------------------
\5 Nonimmigrant aliens are those individuals who come to the United
States for a specific period of time, such as tourists, students, and
business people. 

\6 Of 383 criminal alien paper files we reviewed, 290 contained the
FBI criminal history and 277 contained sets of fingerprints. 


      LESC PILOT OPERATIONS
---------------------------------------------------------- Letter :2.2

The pilot phase of LESC started in July 1994 and is scheduled to last
15 months and will conclude September 30, 1995.  Its objectives are
to develop an interim site in Burlington, Vermont, and, using INS
staff temporarily detailed to the site, field test LESC's ability to
respond to inquiries, initially from the Phoenix Arizona Police
Department and, starting in November 1994, from the Maricopa County,
Arizona, Sheriff's Department.  This phase is estimated to cost $1.4
million.  During fiscal years 1996 and 1997, LESC will accept queries
from California, New York, Florida, Texas, and Illinois in addition
to Arizona.  In its fiscal year 1996 budget, INS has requested 39
staff positions to operate LESC and estimates that LESC operating
expenses will total $3.4 million and $3.6 million for fiscal years
1996 and 1997, respectively. 

An interim independent assessment\7 of the LESC pilot operating
results through February 1995 concluded that the LESC concept was
viable and recommended that the pilot continue so that a full
assessment can be made at the conclusion of the pilot test.  In
October 1995, the Justice Management Division of DOJ plans to perform
a full evaluation of LESC's first year of operation.  If this
evaluation indicates that LESC should be expanded nationwide, INS
plans to conduct this expansion over a 3-year period, subject to
Justice and Office of Management and Budget approval.  In addition,
INS officials told us that a site survey will be done based on GSA's
site selection criteria before a permanent location is selected. 

The pilot links LEAs in Phoenix and Maricopa County with LESC in
Burlington, Vermont, so that the LEAs can have access to information
in INS' five databases to determine whether a person who has been
arrested for an aggravated felony is an alien.  The INS Phoenix
District Office provides investigative support for all LEAs in
Arizona.  In Phoenix, when individuals are arrested for aggravated
felonies and identify themselves as foreign-born, the LEA official
initiates a query to LESC.  However, queries from Maricopa County
(the largest LEA in Arizona) are not limited to aggravated felonies;
they are initiated for all foreign-born individuals arrested for any
offense.  These queries are initiated automatically when information
that an individual is foreign-born is entered into the booking
terminal.  INS officials told us that the decision to include all
arrests was made because the LESC staff needed a much greater number
of inquiries than it was receiving from the Phoenix police department
so that they could more effectively test the mechanics of the LESC
electronic computer system. 

The two LEAs initiate queries to LESC through the existing National
Law Enforcement Telecommunications System (NLETS) network, which LEAs
routinely use for interagency exchange of criminal justice and
related information.  Figure 1 shows how queries are routed between
LEAs and LESC. 

   Figure 1:  Flow of LEA Query
   and Response

   (See figure in printed
   edition.)

In Burlington, Vermont, an LESC staff member receives printed
teletyped queries from the LEAs and enters the name, date of birth,
and other information submitted by the LEAs into computer software
that initiates a sequential search of the five INS databases in the
mainframe at the DOJ Data Center in Dallas, Texas.  The search will
either reveal that no information that matches the name and date of
birth of the individual who has been arrested is available or will
list the name or names of any persons in the databases who are
possible matches.  The LESC staff member compiles the information and
teletypes a response to the LEA which states, "This is not an INS
detainer." The response ends with the following message:  "For
further information, contact (agent's name), Phoenix INS [District
Office], (agent's telephone number)." The response from LESC travels
back to the LEA over the same path that the initial query used. 

The LESC staff member concurrently sends a copy of the response to
the INS Phoenix District Office.  There, an investigator assigned to
the pilot is responsible for analyzing the response.  In the early
months of the pilot, before Maricopa County was added, the INS
investigator first obtained the paper files of the individuals
proposed as possible matches.  If the paper files were not in the
Phoenix District Office, the investigator queried CIS to determine
their location and called that location to expedite the file's
transfer.  According to the Phoenix INS investigator, files are
usually transferred within 24 hours for expedited requests. 

After obtaining the paper file, the Phoenix INS investigator reviewed
it for unique identifying information, such as the fingerprint card. 
In some cases, the investigator would compare the arrested person's
fingerprint card, obtained from the LEA, to the fingerprint card in
the INS file.  If identifications could not be made from the file
data, the investigator interviewed the individuals to determine their
alien status.  The Phoenix investigator told us he could usually
determine whether the individuals were aliens from interviews, since
INS investigators are trained to identify aliens through
interrogations. 

If the arrested person is identified as an alien subject to
deportation, INS will issue a detainer, which requires the LEA to
notify INS before releasing the individual from custody.  Therefore,
if the LEA plans to release individuals on bond or their own
recognizance at any point in the criminal proceedings, INS will be
notified in time to take these individuals into custody and begin the
deportation process on the basis of any previous convictions or
deportations.  Conversely, if the LEA prosecutes and convicts aliens
of aggravated felonies, INS can either initiate deportation
proceedings on the basis of these convictions while they are
incarcerated or take them into custody and begin deportation
proceedings after they have served their sentences. 


--------------------
\7 DOJ contracted with the Institute for Intergovernmental Research
(IIR) to conduct this assessment.  IIR is a nonprofit law enforcement
research and management service. 


   SCOPE AND METHODOLOGY
------------------------------------------------------------ Letter :3

In order to determine whether LESC can identify individuals arrested
for aggravated felonies as aliens, we interviewed the INS
investigator in the Phoenix District Office and police officers in
the Phoenix Police Department and the Maricopa County Sheriff's
Office.  In addition, we reviewed over 300 responses sent to LEAs by
LESC during November and December 1994.  We obtained the results of
LESC operations from July 1994 through May 1995. 

To determine whether other current INS initiatives will allow
identification of aliens arrested for aggravated felonies, we
interviewed senior managers and information resource management (IRM)
officials at INS headquarters and reviewed documentation pertaining
to two initiatives--an INS identification system and a project
between INS and the California Department of Justice.  We did not
perform an in-depth review of these initiatives to assess such issues
as feasibility of implementation schedules, cost effectiveness, and
appropriateness in meeting mission needs. 

To determine whether criminal alien information in INS' databases is
complete and accurate, we focused our review on DACS, INS' repository
for information on identified criminal aliens.  We defined our
universe as all files for identified criminal aliens recorded in DACS
for which the corresponding paper files were located in 17 INS
locations.  The 17 INS locations were selected because they
represented areas with the most known criminal aliens and because of
their proximity to our regional offices.  This defined universe
represented 79 percent of the total number of criminal alien
electronic files recorded in DACS.  We selected a statistically valid
random sample of DACS electronic files from our defined universe and
obtained the corresponding CIS electronic files.  We compared these
electronic files to the corresponding paper case files to determine
accuracy and completeness.  In addition, we selected a judgmental
sample of paper files of aliens for whom deportation proceedings had
begun or had been completed to determine whether they had been
entered into DACS.\8

Our review was conducted between December 1993 and June 1995 in
accordance with generally accepted government auditing standards.  We
requested comments on a draft of this report from the Attorney
General or her designee.  On June 15, 1995, the INS Associate
Commissioner for Enforcement, the Assistant Commissioner for
Investigation, and the Assistant Commissioner for Data Systems
Division provided us with oral comments, which are discussed in the
Agency Comments and Our Evaluation section of this report.  INS
officials also provided us with updated information on the status and
results of the LESC, IDENT, and CAL/DOJ initiatives following the
period of our review.  In addition, the FBI provided technical
comments on factual matters in our report.  These INS and FBI
comments have been incorporated where appropriate.  Appendix II
provides a more detailed discussion of our scope and methodology. 


--------------------
\8 A judgmental sample, rather than a statistical sample, was
required because INS did not know the total number of paper files for
aliens for whom deportation proceedings had begun or had been
completed. 


   NAME-BASED DATABASES LIMIT
   LESC'S ABILITY TO IDENTIFY
   ALIENS
------------------------------------------------------------ Letter :4

Although LESC gives LEAs access to INS information, limitations
inherent in INS' name-based databases delay identification of
arrested persons as aliens until an INS investigator can make a
conclusive determination.  LESC searches INS files using the name and
date of birth of arrested individuals to attempt to match them with
aliens in its databases.  These searches are inconclusive because
names and dates of birth in the files are not unique to an individual
and can be easily falsified. 

Unlike LEAs, who establish criminal history files based on
fingerprints, INS establishes paper and electronic files for
aliens--for both enforcement and service purposes--based on the name,
date of birth, and other personal information that an individual
provides.  However, some individuals may have the same or similar
names and dates of birth as other persons.  In addition, aliens who
commit crimes often provide aliases and other false information in
their encounters with INS.\9 As a result, INS may unknowingly create
multiple files for the same individual and, during a search by name,
may not locate all information on the individual.  In our statistical
sample of 383 criminal aliens in DACS, we found that 317, or about 83
percent, had used one or more aliases and 184, or 48 percent, had
supplied more than one birth date to INS.  The following tables
provide more detailed information on the use of multiple aliases and
birth dates by the criminal aliens in our sample. 



                           Table 1
           
            Aliases Used by Statistical Sample of
                     383 Criminal Aliens

                                       Number of
Aliases used\a                            aliens  Percentage
------------------------------------  ----------  ----------
None                                          66          17
One or more                                  317          83
Two or more                                  267          70
Ten or more                                   72          19
Twenty or more                                19           5
Thirty or more                                 6           2
------------------------------------------------------------
\a For those criminal aliens who used aliases, each grouping is
inclusive.  For example, those who used 30 or more aliases are also
included in those who used 20 or more, 10 or more, and 2 or more
aliases. 



                           Table 2
           
            Birth Dates Used by Statistical Sample
                    of 383 Criminal Aliens

                                       Number of
Birth dates used\a                        aliens  Percentage
------------------------------------  ----------  ----------
One                                          199          52
Two or more                                  184          48
Five or more                                  45          12
Ten or more                                    5           1
------------------------------------------------------------
\a For those criminal aliens who used more than one birth date, each
grouping is inclusive.  For example, those who used 5 or more birth
dates or 10 or more birth dates are also included in those who used 2
or more birth dates. 

From the inception of the LESC pilot in July 1994 through May 31,
1995, LESC received 6,738 queries.  For 56 percent of the queries, no
record was found in the electronic files.  INS officials told us that
the two major reasons why no records were found were that (1) the
alien gave a different name or date of birth than that recorded in
INS' databases or (2) the alien entered the country illegally and
thus was not recorded in any of INS' databases.  The automatic search
process provided a possible match for about 10 percent of the
queries.  However, the LESC staff was able to provide possible
matches for an additional 34 percent of the queries by searching each
database separately using variations of the name.  Then the INS
investigator either reviewed the paper file or interviewed the
individual to conclusively determine that the individual was an
alien. 

The time required for LESC's electronic search and the subsequent
investigation limits INS' ability to detain aliens arrested for
aggravated felonies.  According to Phoenix police officers, they
usually need confirmation that arrested individuals are aliens and an
INS detainer within 8 hours of arrest to ensure that these persons
are not released before INS can take action. 

As a result of the combined electronic search and subsequent
investigations from July 1994 through May 1995, INS took enforcement
action on 1,935 aliens.\10 During this period, however, LEAs released
920 aliens that INS would have detained because INS did not identify
them as aliens before bond was posted or they were released on their
own recognizance.  Of the 920 individuals, the Phoenix investigator
estimated that between 5 and 10 percent (46 to 92) of the released
aliens had been arrested for aggravated felonies--the population
targeted by the mandate.  The remaining individuals that had not been
arrested for aggravated felonies had either been previously deported
or had prior convictions and INS could have taken enforcement action
against them if they had not been released. 


--------------------
\9 Aliases include all names used by an individual that differ from
the name used for the first encounter with LEAs or INS. 

\10 Deportation proceedings were initiated against 114 aliens, 844
aliens agreed to return to their country of origin, 13 were presented
for prosecution for illegal reentry, and INS placed detainers on an
additional 964 aliens. 


   OTHER INS INITIATIVES OFFER
   POSSIBILITIES FOR IDENTIFYING
   ALIENS USING FINGERPRINTS
------------------------------------------------------------ Letter :5

Two other INS initiatives use fingerprints as unique identifiers for
aliens.  In the first, INS is developing the INS Identification
System (IDENT), an automated fingerprint database of aliens INS
processes for either enforcement or benefit purposes.  According to
INS, once IDENT is implemented, the fingerprint of an arrested
individual could be matched against INS' automated fingerprint
database to obtain all available information on that individual, thus
allowing INS to determine the correct course of action to take.  In
the second initiative, the California Department of Justice (CAL/DOJ)
is flagging the state criminal histories of previously deported
criminal aliens based on fingerprint cards supplied by INS.  This
allows LEAs to detain these individuals until INS takes them into
custody.  Because both of these initiatives use fingerprints rather
than names and dates of birth, they offer possibilities for more
effectively identifying aliens.  However, INS will continue to rely
on a name search capability for millions of aliens not covered by the
fingerprint initiatives. 


      IDENT IS INTENDED TO
      IDENTIFY ALIENS BY
      FINGERPRINT
---------------------------------------------------------- Letter :5.1

In calendar year 1995, INS began to implement IDENT, a
fingerprint-based identification system that uses images of the right
index finger to classify and identify individuals.  According to the
November 14, 1994, IDENT draft project plan, "The solution [to INS'
current identification problems] is to move away from names and use
individually-unique biometrics.  The most reliable and commonly used
biometrics are the fingerprints." IDENT will use the single
fingerprint as an identifier to retrieve the proper case information
related to an individual. 

According to INS officials, IDENT workstations equipped with
computerized devices to capture fingerprint images will be installed
throughout INS by 1999.  As these workstations are deployed, INS
staff will begin to develop the IDENT database by scanning and
storing fingerprint images when individuals are first processed for
either enforcement or benefit purposes.  In all subsequent
encounters, INS plans to check these individuals' fingerprints
against its fingerprint database to verify identity.  INS officials
told us IDENT is designed to support (1) criminal alien lookout
checks for all enforcement arrests, (2) verification and
authentication of benefit applicants, (3) prevention of recidivism
(repeated illegal entry into the country), (4) trend analysis of
border apprehensions, and (5) identification and verification of
holders of INS-issued identification cards.  INS estimates that, over
the next few years of operation, IDENT's automated fingerprint
database will contain up to 1.5 million recidivists, 450,000 criminal
aliens, and 25 million benefit applicants. 

The preliminary IDENT 3-year implementation schedule calls for IDENT
to be installed in INS' southwest border sites by the end of calendar
year 1995, in INS' district offices and the four service centers by
the end of 1996, and in the major ports of entry by the end of 1997. 
Therefore, IDENT should be deployed to most existing INS locations
nationwide by 1997.  In fiscal years 1998 and 1999, IDENT will be
deployed to certain remaining INS locations yet to be determined. 

INS began a field test in San Diego, California, in October 1994 to
evaluate the speed and accuracy of fingerprint technology in
identifying aliens who repeatedly enter the country illegally. 
According to an INS official, as of June 13, 1995, prints associated
with 310,261 apprehensions at the California border had been
registered in a fingerprint database.  An analysis of the
fingerprints showed that about one-third were repeat offenders. 
According to an INS official, the fingerprint technology tested
required an average of less than 2 minutes to determine whether the
fingerprints of an individual apprehended matched any of those
registered in the database. 

According to an INS official, fingerprint information from paper
files of 1,870 identified criminal aliens in INS' databases had been
added to the fingerprint database as of June 1995.  Using this system
in the San Diego area, 41 previously deported criminal aliens had
been identified and detained.  Fingerprint information for an
additional 20,254 criminal aliens was to be added to the criminal
alien data set by August 1, 1995. 

INS plans to spend $28 million to develop and begin deploying this
system servicewide in calendar year 1995.  INS estimates continued
development and deployment costs between 1995 and 1999 at about $50
million and $8 million annually for subsequent centralized
maintenance of the automated fingerprint database. 

If IDENT is implemented as planned, it will provide a unique
identifier for all aliens added to INS' CIS, DACS, and NAILS
databases after fiscal year 1997.  If this information can be made
available to LEAs, INS can issue detainers for those arrested
individuals who have a positive fingerprint match without having to
conduct personal interviews.  At the time of our review, however, no
strategy had been adopted to facilitate this exchange of information. 

IDENT will not include the 20 million noncriminal aliens in INS' CIS
database before the start of the project, except for those who
subsequently come into contact with INS for enforcement or benefit
processing purposes.  These individuals include naturalized citizens
and legal permanent residents.\11 Nor will IDENT include fingerprints
of aliens in the Nonimmigrant Information System (NIIS) or the
Student and Schools System (STSC), which together included about 22
million aliens in 1993. 


--------------------
\11 Legal permanent residents are persons to whom INS has granted the
benefit to remain in the United States legally for an unspecified
period of time. 


      STATEWIDE PROJECT ADDS INS
      CRIMINAL ALIEN INFORMATION
      TO LEA DATABASES
---------------------------------------------------------- Letter :5.2

In September 1994, the State of California and INS initiated a
program, funded by a $250,000 federal grant, to enhance California's
ability to identify criminal aliens.  Under this program INS gives
the California Department of Justice (CAL/DOJ) fingerprint cards of
criminal aliens deported from California.  CAL/DOJ reads INS'
fingerprint cards into California's automated fingerprint processing
system to compare these fingerprints to those in California's
automated fingerprint database.\12 For those for which there is a
match, CAL/DOJ places an alert flag in the individual's state
criminal history file.\13 The criminal alien flag states that the
person has been previously deported and directs the LEA to contact
INS at a central telephone number, which has operators on duty 24
hours a day, 7 days a week. 

In California, many individuals deported as criminal aliens reenter
the country illegally after deportation and become repeat offenders. 
When they are rearrested in California, the LEA, using the
established law enforcement procedures, takes fingerprints, accesses
the person's California criminal history, sees the flag, and contacts
INS.  INS has agreed to send a local INS agent to the LEA within 48
hours of being contacted to take custody of the individual.  If the
underlying offense does not dictate that the subject be prosecuted in
California, INS takes the alien into custody for federal prosecution
or deportation proceedings.  This saves California the costs of
custody and prosecution.  Based on INS data, CAL/DOJ plans to place
an alert flag on the records of approximately 7,000 previously
deported criminal aliens during fiscal year 1995.  As of June 20,
1995, 5,113 alien alert flags had been entered into California's
criminal history system.  This program has enabled the LEAs to
identify 553 individuals as criminal aliens and report them to INS,
who has either taken them into custody or placed a detainer on them. 

Although CAL/DOJ is a California initiative, other states can access
California's criminal alien information through the Interstate
Identification Index.\14 LEAs can query the Interstate Identification
Index through the National Crime Information Center (NCIC)\15 to
determine whether an arrested individual has an available criminal
history record.  Thus, LEAs across the nation can access California's
criminal histories and, when the criminal history contains a criminal
alien flag, identify criminal aliens they have arrested. 


--------------------
\12 California's automated fingerprint processing system uses minutia
(fingerprint ridge characteristics) matching technology and an image
system to provide the user with fingerprints of likely candidates for
fingerprint comparisons.  Fingerprint specialists compare the
fingerprints of the likely candidates with the fingerprint that has
been submitted to conclusively determine whether there is a match. 

\13 A state criminal history file is established for each offender
booked into a California facility.  California's Criminal History
System contains California arrests and convictions information on
offenders. 

\14 The FBI sponsors this system which stores names of offenders for
whom criminal histories reside in various state law enforcement
systems, and the FBI provides the records for nonmember states. 

\15 NCIC is the nation's most extensive criminal justice information
system and is maintained by the FBI.  NCIC--established as a service
to all local, state, and federal criminal justice agencies--stores
vast amounts of information, such as criminal history data and
foreign fugitive data, which can be instantly retrieved and furnished
through an NCIC terminal at any authorized agency. 


   CRIMINAL ALIEN INFORMATION IN
   DACS AND CIS IS INACCURATE AND
   INCOMPLETE
------------------------------------------------------------ Letter :6

We found serious problems with the quality of the criminal alien data
in INS' Deportable Alien Control System (DACS) and the corresponding
Central Index System (CIS) files.  First, the electronic files did
not contain complete information on the aliases used by criminal
aliens.  For over 80 percent of the criminal aliens we identified in
our statistical sample, the electronic files did not contain a
complete listing of the aliases available in the paper files.\16 For
example, one individual in our sample had 24 aliases listed in the
paper file but none was recorded in the corresponding DACS and CIS
electronic files.  Consequently, an electronic search using one of
those aliases would not locate that person's file, resulting in a
response that no record existed for the alien. 

Second, 22 percent of the DACS electronic files in our statistical
sample contained either misspelled names, incorrect order of names,
or incorrect nationality.  Based on these results, such errors are
projected to occur in about 22,000 of the over 101,000 electronic
files for criminal aliens in our DACS test population.\17
Consequently, LESC searches using the correct spelling, order of
names, and nationality may not locate the alien electronic file. 

Third, the FBI number--which is linked to fingerprints and uniquely
identifies persons previously arrested for serious offenses--was
missing from the CIS files for most of the criminal aliens in our
sample.  The FBI number was in the paper files for 290 of the
criminal aliens in our statistical sample but was missing from 216
(74 percent) of the corresponding CIS files. 

Fourth, INS did not have both electronic and paper files for all
individuals.  For the 410 criminal aliens in our statistical sample
of criminal aliens recorded in DACS, 27 of the corresponding paper
files could not be located.  Since the paper files are used to verify
the accuracy and completeness of the data in the electronic files,
the data in these files could not be verified.  Nor could the
investigator consult these paper files to conclusively identify these
aliens.  In addition, electronic files alone will not support
deportation hearings. 

We also found instances where paper files existed, but electronic
files did not.  We selected a judgmental sample of 400 paper case
files of aliens for whom deportation proceedings had begun or had
been completed and attempted to find a corresponding electronic file
in DACS.  There were no corresponding electronic DACS files for 77
(19 percent) of the 400 paper case files we selected.  As a result, a
query of DACS would not identify these 77 individuals as aliens who
had been deported or were under deportation proceedings for a
criminal offense. 

Of the 323 DACS electronic files we did find for our sample of paper
files, our review of the paper files showed that 175 of the
electronic files should have contained alert codes\18

designating individuals who had been convicted of criminal
activity.\19 Because 72, or 41 percent, of the 175 electronic files
did not have the alert code, a query of DACS would not inform INS
staff that these individuals were criminal aliens and thus subject to
deportation.  The missing alert codes also prevent INS from knowing
the total number of criminal aliens in its database. 

In oral comments on this report, INS officials acknowledged that
keying errors do occur when information from manually prepared
enforcement forms is entered into DACS.  They stated that in the
future, rather than preparing a paper form, this information will be
entered into the planned Enforcement Case Tracking System (ENFORCE)
Phase I and downloaded to DACS, thus eliminating the need for
duplicate data entry and any resulting key-entry errors.  The ENFORCE
Phase I is currently being prototyped in the McAllen, Texas, and San
Diego, California, Sectors and the Philadelphia District; the
interface will be tested in New York.  According to the Assistant
Commissioner, Data Systems Division, ENFORCE Phase I (and the DACS
interface feature) will be implemented in New York, Texas,
California, Florida, and Illinois after it is successfully prototyped
and tested. 

INS officials acknowledged that they do not have servicewide
procedures directing staff to update the electronic files with
critical new information received after the file has been
established--such as the FBI number or information on aliases.  In
addition, INS did not have procedures for ensuring that data are
entered correctly and completely. 

INS officials stated that they had begun taking action to improve
data reliability.  In April and June 1995, memoranda were issued to
all Regional Directors and the Director of International Affairs
requiring them to establish a method that will ensure timely input of
data into DACS and specifying the data elements that were to be
tracked.  However, these memoranda did not contain specific
procedures for ensuring data accuracy and completeness, nor did they
provide for an independent verification to ensure that such
procedures are followed. 


--------------------
\16 The range of our confidence interval, at a 95-percent confidence
level, is that the actual percentage of DACS electronic files that
did not include all aliases listed in the paper files was between 77
percent and 85 percent; while the corresponding CIS electronic files
did not include all aliases between 79 percent and 87 percent of the
time. 

\17 The range of our confidence interval, at a 95-percent confidence
level, is that the actual number of DACS electronic files that did
not match the paper files for name and nationality was between 17,889
and 26,493. 

\18 These codes identify the type of alert that applies to the alien,
for example, aggravated felon, criminal, or narcotics. 

\19 INS does not classify illegal aliens who have not been convicted
of crimes committed in the United States as criminal aliens. 


   CONCLUSIONS
------------------------------------------------------------ Letter :7

Identifying individuals arrested for aggravated felonies as aliens is
critical to joint INS and LEA efforts to prevent the release of these
individuals before INS can take action.  LESC is an attempt to
provide this identification capability; however, this approach is
inherently limited by the existing name-based systems that it depends
upon.  Until INS successfully implements a system that identifies
individuals based on biometric information, INS' ability to make
timely identification of arrested individuals as aliens will continue
to be limited.  INS' planned move to the IDENT automated fingerprint
database is intended to address the need for an improved
identification method for individuals who will be processed for
either enforcement or benefit purposes. 

Further, accurate and complete criminal alien data in INS' DACS and
CIS databases are essential.  Unless INS' data reliability problems
are resolved, INS risks making decisions based on inaccurate and
incomplete information. 


   RECOMMENDATIONS
------------------------------------------------------------ Letter :8

In view of the limitations inherent in name-based databases, we
recommend that the Attorney General direct the Commissioner of INS to
take the following actions before deciding whether to expand LESC
nationwide. 

  Assess whether the information generated by LESC's electronic
     searches justifies the expense and level of resources required
     to expand and maintain a nationwide facility. 

  Determine whether any other alternative would be more effective and
     efficient than LESC in helping identify which arrested
     individuals are aliens. 

To improve the reliability of the criminal alien data in DACS, and
the corresponding electronic files in CIS, we recommend that the
Attorney General direct the INS Commissioner to develop procedures
that will ensure data reliability for both DACS and CIS.  At a
minimum, these procedures should ensure that

  electronic files are created for all known criminal aliens;

  all criminal alien information--including name, date of birth,
     nationality, and aliases used--is entered into the electronic
     files accurately and completely;

  alert codes are included in all criminal alien electronic files;
     and

  the Regional Directors and the Director of International Affairs
     are directed to take appropriate actions to ensure that all
     paper files supporting the criminal alien electronic files are
     located or, if necessary, reconstructed. 

Finally, after these procedures are implemented, we recommend that
the Attorney General direct the Commissioner to develop a strategy to
independently verify that the procedures are followed and that data
reliability is improved. 


   AGENCY COMMENTS
------------------------------------------------------------ Letter :9

We discussed a draft of this report with INS' Associate Commissioner
for Enforcement, the Assistant Commissioner for Investigations, and
the Assistant Commissioner for Data Systems Division.  While these
officials agreed with most of our findings and recommendations, they
took issue with several points.  First, the Associate Commissioner
for Enforcement stated that we had misinterpreted the legislative
mandate for INS contained in the Anti-Drug Abuse Act of 1988 by
stating that INS was required to immediately identify aliens arrested
for aggravated felonies.  He stated that under the congressional
mandate, INS is only required to provide--on a 24-hour
basis--federal, state, and local criminal justice entities with all
available information it has on suspected aggravated felons at the
time of their reported arrest with follow-up action to confirm
identity and alien status in the course of the criminal justice
process. 

We believe INS has misstated the content of our draft.  We agree that
the congressional mandate does not require INS to immediately
determine whether arrested persons are aliens.  Our draft stated only
that INS is limited in its ability to take appropriate enforcement
action against aliens who are arrested for aggravated felonies
because of the lengthy time required by LESC to conduct the
electronic search and by an INS investigator to conclusively identify
the arrested individual as an alien.  To more precisely reflect the
language in the law, we have clarified the section of the report
concerning INS' legislative mandate. 

Second, the Associate Commissioner for Enforcement expressed concern
that the report repeatedly suggested that the LESC approach was
flawed because of its use of name-based methodology and that the
report failed to note that almost all law enforcement systems,
including the FBI's NCIC, presently operate in this manner.  Our
report does point out that INS' existing name-based databases limit
LESC's ability to make timely determinations of whether arrested
persons are aliens.  INS itself has acknowledged these same
limitations and is, for this reason, developing IDENT, a
fingerprint-based identification system to provide quick, accurate
identification of individuals who will be processed for either
enforcement or benefit purposes.  While NCIC does provide for name
searches, it differs from INS databases by relying on
fingerprint-based information to establish case files.  This reduces
the risk--that INS still faces--of creating multiple files on an
individual. 

INS officials disagreed with our recommendation regarding the
evaluation of the LESC pilot.  They stated that we did not need to
recommend that INS include an evaluation of alternatives and the
advisability of expanding LESC nationwide in its final evaluation
because these factors are already part of their planned evaluation
document.  The officials, however, declined to provide a copy of this
document until it is approved by the Department of Justice. 

Finally, in reference to our recommendation on data quality, INS
officials stated that steps have been taken to improve the accuracy
and completeness of the information in DACS and that these steps were
not reflected in the report.  We have modified the report to reflect
that INS (1) plans to deploy an electronic interface to reduce keying
errors and (2) has instructed its field officials to establish a
method of timely and consistent input of data.  However, since INS
has not established specific procedures for ensuring that data in
DACS and CIS are accurate and complete, we have revised the
recommendation to state that INS should develop a strategy to
independently verify that procedures are followed and data
reliability is improved. 


---------------------------------------------------------- Letter :9.1

We are providing copies of this report to the Attorney General;
Commissioner of INS; the Director, Office of Management and Budget;
and other interested parties.  Copies will also be made available to
others upon request. 

Major contributors to this report are listed in appendix III.  Please
contact me on (202) 512-7487 if you need any additional information
or have any further questions concerning this report. 

Sincerely yours,

Linda D.  Koontz
Associate Director, Information Resource
Management/General Government Issues


INS DATABASES QUERIED BY LESC
=========================================================== Appendix I

CENTRAL INDEX SYSTEM (CIS)

CIS a centralized, computer-based information system that serves as
the heart of INS mission support, in both service benefits and law
enforcement.  The Central Index contains data on lawful permanent
residents, naturalized citizens, violators of immigration laws,
aliens with Employment Authorization Document information, and others
for whom the Service has opened alien files or in whom it has a
special interest.  Each physical file with a corresponding electronic
file in DACS should have an electronic file in CIS. 

The major search keys for CIS are A-number and Name.  Variations of
the Name Search are provided by allowing a direct search using Exact
Name or a Sounds-like (Soundex) search using a similar-sounding name
or alias name.  Additionally, the Name Searches allow other
identifying information as secondary search criteria, such as Date of
Birth, Country of Birth, and Files Control Office; Date of Birth is
the most often used secondary search criteria. 

DEPORTABLE ALIEN CONTROL SYSTEM
(DACS)

DACS supports INS' enforcement activities.  It provides information
on the status and disposition of deportation cases and on the
statistics and summary data representing cases by status type and
other activities.  DACS captures deportation data; tracks aliens who
are arrested, detained, or formally removed from the country;
produces deportation reports; and makes the information accessible
on-line to deportation officers and other INS users.  DACS maintains
information on aliens detained by INS and reports on detention
activity. 

NATIONAL AUTOMATED IMMIGRATION
LOOKOUT SYSTEM II (NAILS II)

NAILS II is a lookout enforcement system that contains information
about persons of interest to INS for law enforcement purposes.  It
expedites the determination of traveller admissibility into the
United States at the various ports of entry and identifies
individuals who are suspected of illegal activities.  NAILS II is
used by inspectors at various ports of entry throughout the country. 

NONIMMIGRANT INFORMATION SYSTEM
(NIIS)

NIIS contains arrival, departure, and ancillary information
pertaining to nonimmigrant aliens entering the United States.  It
contains data on the individual's status, identifies individuals who
may have overstayed, and provides statistical information to INS
managers.  It provides for queries based on biographical,
classification, and citizenship data.  There are no physical files to
complement the NIIS electronic file. 

STUDENT AND SCHOOLS SYSTEM (STSC)

STSC is the primary vehicle for identifying, locating, and
determining the status or benefits eligibility of nonimmigrant
students and their dependents.  The data in STSC includes requests
for extensions, change of status, transfers, and employment
authorization.  It also maintains records on approved schools, school
officials, and current or past violations. 


SCOPE AND METHODOLOGY
========================================================== Appendix II

To determine the type of information provided to requesters in
responses from LESC, we performed test queries at both LESC in
Burlington, Vermont, and the Phoenix Police Department in Phoenix,
Arizona.  In addition, we reviewed over 300 LESC responses sent to
LEAs for November and December 1994. 

We interviewed senior managers and IRM officials at INS headquarters
to discuss criminal alien information and initiatives underway to
address existing problems with identifying aliens.  We also reviewed
documentation pertaining to two initiatives--an INS identification
system and a project between INS and the California Department of
Justice.  We did not perform an in-depth review of these initiatives
to assess such issues as feasibility of implementation schedules,
cost-effectiveness, and appropriateness in meeting mission needs.  We
interviewed the Director of LESC, special agents assigned to LESC,
and police officers in the Phoenix Police Department to discuss the
operations at LESC.  In addition, we interviewed FBI officials to
discuss their Integrated Automated Fingerprint Identification System
and the FBI fingerprinting process. 

We tested the accuracy and completeness of criminal alien information
in INS' Deportable Alien Control System (DACS)--the repository for
information on identified criminal aliens--by comparing source
documents in the paper case file to the information in DACS'
electronic file.  Our statistical sample of case files for 410
individuals was selected from DACS, which had electronic files for
959,349 individuals who were or had been in deportation proceedings
as of May 20, 1994.  Of those, we included for testing only those
that had a criminal record in the electronic file, 136,744
individuals.  We then scoped this universe to 17 INS locations
representing 79 percent of the population of individuals with
criminal records recorded in DACS, or 108,502 individuals.  We also
obtained the corresponding CIS electronic files for the 410
individuals in our sample and compared them to the paper case files
and to the DACS electronic files.  In addition, at 16 of the
locations where we performed paper case file reviews, we judgmentally
selected an additional 400 physical case files--25 at each
location--and determined if there was a corresponding electronic file
in DACS. 

For our statistical sample, the sampling method used allowed us to
estimate, at a 95-percent confidence level, the (1) instances of
inaccurate recording of aliases, (2) the number of files without the
FBI number, (3) the number of errors in names or nationality, and (4)
the number of paper case files that could not be located. 

Our projections are expressed as point estimates that fall within
confidence intervals.  This means that if you were to determine an
estimate for 100 different random samples of the same size from this
population, the estimate would fall within the confidence interval 95
out of 100 times.  In other words, the true value is between the
lower and upper limits of the confidence interval 95 percent of the
time. 

Our case file reviews were performed at the following INS district
offices:  Chicago, Illinois; Los Angeles, San Francisco, and San
Diego, California; Phoenix, Arizona; San Antonio, El Paso, and
Houston, Texas; Miami, Florida; New York City, New York; Denver,
Colorado; New Orleans, Louisiana; Newark, New Jersey; and Arlington,
Virginia.  Additional locations included the Varick, New York; El
Centro, California; Florence, Arizona; San Pedro, California; and El
Paso, Texas, Service Processing Centers.  We also performed reviews
at the federal prison in Oakdale, Louisiana. 

We reviewed previous GAO reports pertinent to the Criminal Alien
Program, as well as reports of the Justice Office of the Inspector
General pertinent to fingerprint requirements and the National
Automated Immigration Lookout System II. 


MAJOR CONTRIBUTORS TO THIS REPORT
========================================================= Appendix III

ACCOUNTING AND INFORMATION
MANAGEMENT DIVISION WASHINGTON, D. 
C. 

Antionette Cattledge, Assistant Director
Barbara S.  Oliver, Project Manager
Brian Spencer, Technical Advisor
Tamara J.  Lilly, Computer Specialist
Linda J.  Sellevaag, Communications Analyst

ATLANTA REGIONAL OFFICE

John W.  Randall, Jr., Evaluator

CHICAGO REGIONAL OFFICE

Lenny R.  Moore, Evaluator

DALLAS REGIONAL OFFICE

George Jones, Senior Evaluator
Elaine M.  Coleman, Evaluator
James B.  Smoak, Senior Evaluator

DENVER REGIONAL OFFICE

John A.  Spence, Senior Evaluator

LOS ANGELES REGIONAL OFFICE

Michael P.  Dino, Senior Evaluator
James R.  Russell, Evaluator

NEW YORK REGIONAL OFFICE

George P.  Cullen, Senior Evaluator
Lucine Moore Willis, Evaluator

SAN FRANCISCO REGIONAL OFFICE

Delores J.  Lee, Evaluator
Yola Lewis, Evaluator

