Customs and INS: Random Inspection Programs Can Be Strengthened  
(03-DEC-01, GAO-02-215R).					 
								 
Prior to the events of September 11, 2001, GAO had initiated a	 
review of the U.S.  random inspection programs--Customs'	 
Compliance Measurement Examination (COMPEX) and Immigration and  
Naturalization Service's (INS) Inspections Traveler Examination  
(INTEX). These programs help Customs and INS assess the nature	 
and extent of enforcement risks at ports of entry. They measure  
these agencies' effectiveness by comparing violations found	 
during targeted inspections with violations found during the	 
inspection of random samples. GAO found that both Customs and INS
inspectors performing COMPEX and INTEX inspections did not always
adhering to guidance on sample selection and did not always	 
conduct inspections with the minimum level of thoroughness	 
required. As a result, statistical data generated by the programs
may not reliably reflect the extent to which travelers who seek  
entry into the U.S. are in violation of customs or immigration	 
laws. GAO also noted that the COMPEX and INTEX programs both draw
from the same population of international travelers, have similar
purposes and goals, and oftentimes incorporate Customs and INS	 
inspectors who work side by side, particularly at land border	 
ports of entry. Because of this, Customs and INS might realize	 
some efficiencies if the two random inspection programs were	 
combined.							 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-02-215R					        
    ACCNO:   A02532						        
  TITLE:     Customs and INS: Random Inspection Programs Can Be       
Strengthened							 
     DATE:   12/03/2001 
  SUBJECT:   Illegal aliens					 
	     Inspection 					 
	     Statistical data					 
	     Immigration information systems			 
	     Customs Compliance Measurement			 
	     Examination Program				 
								 
	     INS Inspections Traveler Examination		 
	     Program						 
								 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Testimony.                                               **
**                                                              **
** 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-02-215R
     
GAO- 02- 215R Random Inspection Programs United States General Accounting
Office

Washington, DC 20548

December 3, 2001 Mr. Robert C. Bonner Commissioner U. S. Customs Service

Mr. James W. Ziglar Commissioner Immigration and Naturalization Service

Subject: Customs and INS: Random Inspection Programs Can Be Strengthened
Prior to the tragic events of September 11, 2001, we had initiated a review
of the U. S. Customs Service?s and the Immigration and Naturalization
Service?s (INS) random inspection programs- Customs? Compliance Measurement
Examination (COMPEX) and INS? Inspections Traveler Examination (INTEX). The
purpose of COMPEX and INTEX is to help Customs and INS assess the nature and
extent of enforcement risks at ports of entry. The programs measure these
agencies? effectiveness by comparing violations found during targeted
inspections with violations found during the inspection of random samples.
Because circumstances and inspectional activities and priorities have
dramatically changed at our nation?s ports of entry since September 11, we
intend to do no further work on the random inspection programs at this time.
However, our observations of each program?s operations raised some concerns
about the design and implementation of COMPEX and INTEX and suggested some
opportunities to combine the two programs. This report presents our concerns
and suggestions for strengthening the programs once normal operations
resume.

Summary

We found that both Customs and INS inspectors that performed COMPEX and
INTEX inspections were not always adhering to guidance on how to properly
select samples and were not always conducting inspections with the minimum
level of thoroughness required. As a result, the statistical data generated
by the programs may not reliably reflect the extent to which travelers who
seek entry into the U. S. are in violation of Customs or immigration laws.
We also noted that the COMPEX and INTEX programs both draw from the same
population of international travelers, have similar purposes and goals, and
oftentimes incorporate Customs and INS inspectors who work side by

Page 2 GAO- 02- 215R Random Inspection Programs

side, particularly at land border ports of entry. Because of this, Customs
and INS might realize some efficiencies if the two random inspection
programs were combined. Consequently, we are recommending that Customs and
INS (1) consider refresher training or other reminders to inspectors on
proper random inspection selection procedures and (2) explore opportunities
for combining the random inspection programs. Customs and INS officials
concurred with these recommendations.

Objectives, Scope, and Methodology

The objectives of our work were to determine (1) the purpose of Customs? and
INS? random inspection programs, (2) how the programs have been designed and
implemented, including the sample selection procedures, and (3) the
feasibility of combining the programs. To accomplish these objectives, we
interviewed Customs and INS officials responsible for overseeing COMPEX and
INTEX at both headquarters and at two ports of entry, Los Angeles
International Airport (LAX) and the San Ysidro land border port near San
Diego, California. In addition, we observed the COMPEX and INTEX inspection
processes at the two ports and analyzed COMPEX and INTEX measurement
results. We conducted our review from July to September 2001 in accordance
with generally accepted government auditing standards. This report provides
the results of our limited work.

Background

Customs and INS are authorized to inspect every person coming through a U.
S. port of entry. Customs? inspection is aimed at ensuring that incoming
travelers do not bring any contraband (e. g., drugs or weapons) into the
country, and INS? inspection is aimed at ensuring that the people have
proper documentation (e. g., passport or visa) and are admissible to enter
the country. Most inspections consist of a brief interview and a cursory
document check that takes only a matter of seconds. If intelligence reports,
targeted enforcement operations, or the inspector?s suspicions lead an
inspector to believe that the traveler may be committing a violation, the
inspector will refer the traveler for a more intensive ?secondary
inspection.? 1 During the secondary inspection the traveler is extensively
questioned and his or her belongings may be examined thoroughly. A very
small percentage of travelers are given these intensive inspections.

Both Customs and INS began their random inspection programs (COMPEX and
INTEX) to provide data that validates their inspection efforts and addresses
the Government Performance and Results Act (GPRA) requirement to identify
outcome measures to portray the benefit the agency provides to the American
public. COMPEX and INTEX are designed to randomly select and intensively
inspect slightly more international travelers than would normally undergo
intensive inspection.

1 Targeted enforcement operations make use of intelligence and investigative
capability to identify highrisk travelers to prevent and disrupt the
smuggling of contraband or other illegal activities.

Page 3 GAO- 02- 215R Random Inspection Programs

Combining the results of these random inspections with the results of
targeted inspections is intended to allow Customs and INS to estimate the
proportion of total violations that each agency identifies. 2

In addition, Customs intended that COMPEX data be used to calculate a
statistically valid measure of the total number of violations. Customs then
compares that measure to the number of violations found during targeted
enforcement inspections to determine the effectiveness of targeted
enforcement operations.

COMPEX began in 1995, several years before INTEX. Customs fully implemented
COMPEX in fiscal year 1997 at 20 airports that each process about a million
passengers a year and at about 70 3 land border ports. At each airport and
land border port that process more than a million passengers a year, Customs
samples about 12,000 passengers a year; about 33 a day. INS initiated INTEX
in fiscal year 1998 at 20 ports- 10 airports and 10 land border ports of
various sizes. In fiscal year 2000, INS added an additional airport and land
border port for a total of 22 ports. At the time of our review, INS was
pilot- testing INTEX and, as of the end of October 2001, had suspended the
program indefinitely. Prior to the suspension, INS officials told us that
they were considering several options concerning INTEX?s future: expanding
the program to more ports, ending the program entirely, or possibly merging
INTEX into COMPEX.

Concerns about COMPEX Design and Implementation

Customs established criteria stating that all passengers must have an equal
chance of being selected for COMPEX inspections. We found, however, that not
all passengers necessarily have an equal chance of being selected. For
example, Customs did not design COMPEX to have inspections performed on
pedestrians and buses at all COMPEX land border ports, including San Ysidro.
By excluding pedestrians and others from the sample population, about one-
third of San Ysidro?s international travelers are excluded from COMPEX. In
comparison, INTEX includes pedestrians and buses in its sample population.
Customs? officials stated that they are exploring how to conduct COMPEX
inspections on these populations in the future.

As a second example, not every passenger arriving at LAX had an equal chance
of being selected. Inspectors were given designated times to make COMPEX
selections from passengers who were exiting Customs? inspection area. At one
international terminal, an inspector who was monitoring two lines of
passengers told us that he always makes a COMPEX selection from one line and
not from the other.

We also found that inconsistent procedures may affect the quality of COMPEX
inspections and subsequent statistics. Customs? criteria states that a
thorough inspection must be conducted for COMPEX results to be valid. San
Ysidro officials

2 One traveler could be charged with one or more violations. 3 Shortly
before releasing this report, Customs officials told us that they now
conduct COMPEX

inspections at 38 land border ports which consist of about 90 percent of the
land border traffic.

Page 4 GAO- 02- 215R Random Inspection Programs

said, and our observation confirmed, that COMPEX inspectors may not always
conduct inspections with the minimum required level of detail. For example,
we observed inspectors briefly examining a vehicle selected for a COMPEX
inspection. It did not appear that the inspectors interviewed the occupants
or conducted computer checks, both mandatory parts of a COMPEX inspection.

Statistics from fiscal year 2000 COMPEX data show that Customs estimated
that it apprehended 15.3 percent of serious violations at the land border
ports and 20.4 percent at the airports. In addition, for fiscal year 2000,
Customs estimated that its targeted inspections were 15.6 times more
effective than the random COMPEX inspections at the airports and 12.6 times
more effective than the random COMPEX inspections at the land border ports.
However, because of the issues that we identified concerning random sample
selection, these data may not reliably estimate the extent to which
travelers who seek entry into the United States are in violation of Customs
laws.

Concerns about INTEX Design and Implementation

Under the INTEX pilot program design, the data being collected cannot be
used to make estimates about all ports of entry because the ports conducting
INTEX inspections were not randomly selected, according to the INS?
Statistics Branch Director. Furthermore, under INTEX, only five travelers a
day are inspected at each large port, three at each medium port, and one at
each small port. In addition, the ports do not perform the inspections every
day; LAX officials said that they do not conduct INTEX inspections about 1
day each week, and San Ysidro officials said that inspections are not
conducted on Sundays and holidays.

In addition, INTEX statistics show that no port conducted all of its
scheduled inspections. John Fitzgerald Kennedy International Airport
conducted the largest proportion of its scheduled INTEX inspections - 95
percent - in fiscal year 1999, while the San Juan, Puerto Rico, Airport
performed the smallest proportion - 45 percent. INS? INTEX Fiscal Year 2000
Annual Report cited several factors that could account for the
discrepancies. For example, the report noted that port of entry management
support for the program varied. Another factor noted was that port of entry
workloads and staffing varied among the ports and that some ports may not
have enough inspectors to perform the INTEX inspections.

On the basis of our observations at the two ports of entry, we had
additional concerns about INTEX implementation. Since INTEX is used to
measure actual rates of people incorrectly admitted into the United States,
it is important that inspections are done randomly. Each person entering at
the port should have an equal probability of being selected for an INTEX
inspection. 4 At LAX, INTEX inspections are conducted at only one of the
four international passenger arrival terminals, which processes about two-
thirds of the international flights. Consequently, about one- third of the
population of international passengers at LAX are excluded from INTEX. At

4 Using Statistical Sampling, (GAO/ PEMD- 10.1.6, revised May 1992).

Page 5 GAO- 02- 215R Random Inspection Programs

San Ysidro, we also identified another issue that could affect INTEX sample
selections. While a computer program that schedules daily INTEX inspections
specifies a time and passenger lane from which to make a sample selection,
the inspector we observed did not follow the instructions. The inspector
said that he did not receive any formal training on how to conduct INTEX
inspections and had not been provided with the INTEX training manual.

Statistics from fiscal year 2000 INTEX data show that INS estimated that
they prevented 20.3 percent of inadmissible aliens from entering the country
at the 11 land border ports and prevented 42.4 percent from entering the
country at the 11 airports. However, because of the issues that we
identified concerning random sample selection, these data may not reliably
estimate the extent to which travelers who seek entry into the United States
are in violation of immigration laws.

Opportunities to Combine Random Inspection Programs

Customs and INS both conduct intensive random inspections from the same
population of travelers. In addition, at land border ports, Customs and INS
inspectors work side by side when conducting initial inspections of
travelers entering the United States and are cross- trained to be alert for
violators of each other?s laws and regulations. Customs and INS are also
trying to coordinate more of their activities under the Border Coordination
Initiative (BCI). The BCI is a comprehensive, coordinated border management
strategy between Customs, INS, and other agencies to improve interagency
coordination, particularly their joint efforts along the Southwest border.

We asked Customs and INS officials, both in headquarters and at the two
ports of entry, whether it might be beneficial to combine the COMPEX and
INTEX programs. Most Customs and INS headquarters officials appeared
receptive to this idea. In particular, an INS official said that because
INTEX was still a pilot program with sample sizes too small to make reliable
estimates about the population, it would make sense to merge the INTEX
program into the larger, more established COMPEX program. Other Customs and
INS officials, however, expressed concerns. Some INS officials believe that
a slightly different philosophy behind the two programs could hinder a
potential merger. COMPEX, for example, focuses on the proportion of total
violations detected through targeting as compared to random (COMPEX)
inspections. INTEX is geared toward estimating how effectively INS
identifies inadmissible aliens attempting to enter the country. However, as
indicated earlier, Customs also uses COMPEX to estimate the total number of
violations being committed, similar to INTEX. Some Customs and INS officials
also expressed a concern involving logistics; for example, the separation of
Customs and INS work areas at some ports of entry and the lack of mutual
access to each other?s computers. Several officials said, however, that it
would be worth the effort to consider combining the random inspection
programs and working toward overcoming the philosophical and operational
concerns of a combined random inspection program.

Page 6 GAO- 02- 215R Random Inspection Programs Conclusion

Although the concerns we identified at the two selected ports may not be
representative of Customs and INS operations nationwide, we believe that our
observations are indicators of conditions that could be addressed to
strengthen the efficiency and effectiveness of the agencies? random
inspection programs and improve the reliability and completeness of the
statistics from these programs as normal operations return in the future.
The two programs have similar purposes and goals, which might be more
closely aligned to satisfy each agency?s need to measure its inspectional
effectiveness. Moreover, the two programs select travelers for inspection
from the same pool of arrivals at our air and land border ports of entry.
Combining the two programs into one might increase effectiveness and reduce
the burden on the traveling public.

Recommendations for Executive Action

To better coordinate random inspection efforts and to increase the
efficiency and quality of the random inspection process, we recommend that
Customs and INS

-- consider refresher training or other reminders to inspectors conducting
random inspections on proper random selection procedures and

-- explore the feasibility of combining their random inspection programs.

Agency Comments

We provided the Customs Service and INS an opportunity to comment on the
draft of this report. Both Customs and INS officials concurred with the
recommendations and provided some technical comments that we incorporated
where appropriate. In addition, Customs officials said that because of other
pressing priorities, they probably would not be able to address exploring
the feasibility of combining random inspection programs until the second
quarter in 2002, when they will discuss the matter with INS.

INS officials said that INTEX employed a methodology that INS considered
using to measure its operational effectiveness. INS decided, after a pilot
test of the program, that INTEX was not the desirable methodology to
accomplish this goal and to meet the requirements of GPRA. INS subsequently
decided to terminate the program. INS recognizes that it needs to develop a
random inspection program to measure the effectiveness of determining
travelers? admissibility to the United States and to meet GPRA requirements.
In developing the program, INS will look to Customs to harmonize processes
with the intention of developing a joint random inspection program that
would be more economical and less disruptive to the traveling public.

------------

Page 7 GAO- 02- 215R Random Inspection Programs

We will send copies of this report to the Chairman and Ranking Minority
Member of the Senate Committee on Finance; the Chairman and Ranking Minority
Member of the Senate Governmental Affairs Committee; the Chairmen and
Ranking Minority Members of the House Ways and Means Committee and its
Subcommittee on Trade; and the Chairman and Ranking Minority Member of the
House Government Reform Committee; and others upon request.

This report will be available on GAO?s homepage at http:// www. gao. gov. If
you have any questions regarding this report, please contact me on (202)
512- 8777. Darryl W. Dutton, Cheryl L. Gordon, and Bradley W. Hunt made key
contributions to this report.

Richard M. Stana Director, Justice Issues

(440062)
*** End of document. ***