Since the September 11, 2001, terrorist attacks, the Department of Justice's (Department) highest priority has been to deter, prevent, and detect future terrorist acts. In part, the Department measures the success of its counterterrorism efforts by reporting hundreds of terrorism-related statistics in its performance plans and statistical reports, such as the:

• number of individuals charged as a result of terrorism investigations,

• number of terrorism convictions,

• number of terrorism-related threats to transportation and facilities, and

• number of terrorism-related threats to people and cities.

Congress and the Department management also use terrorism-related statistics to make operational and funding decisions for Department counterterrorism activities, and to support the Department’s annual budget requests.¹ For these and other reasons, it is essential that the Department report accurate terrorism-related statistics.

The Department’s Office of the Inspector General (OIG) initiated this audit to determine if Department components and the Department as a whole gather and report accurate terrorism-related statistics.

Terrorism-Related Statistics Reported by the Department

In transforming its mission after the terrorist attacks of September 11, 2001, the Department established preventing terrorism and promoting America’s security as its primary strategic goal. The Department relies on its components, such as the Federal Bureau of Investigation (FBI), the Criminal Division, the Executive Office for United States Attorneys (EOUSA), and the United States Attorney’s Offices (USAO), to implement its counterterrorism strategies. The Department and its components collect a variety of terrorism-related statistics measuring these counterterrorism efforts.

We identified the terrorism-related statistics reported by the Department and its components by:

• interviewing officials from the Department, the FBI, the Criminal Division, and EOUSA;

• analyzing budget submissions, congressional testimony, performance plans, and other documents maintained by the Department and its components; and

• viewing the website maintained by the Department at www.lifeandliberty.gov to keep the public informed of the Department’s counterterrorism efforts.

In total, we identified 209 unique terrorism-related statistics that were reported by the Department and its components 602 times from October 1, 2000, through September 30, 2005.² The following chart shows which components reported these 209 statistics.

Source: Interviews with Department officials and Department documents

In testing the accuracy of these terrorism statistics, we excluded from the 209 statistics 17 that were generated by agencies outside the Department. For our initial testing of the remaining 192 statistics, we interviewed Department officials from the FBI, the Criminal Division, EOUSA, and other Department agencies to determine whether internal controls were in place to ensure the 192 statistics were accurately gathered, classified, and reported.

Through these interviews, we learned that the collection and reporting of terrorism-related statistics within the Department is decentralized and haphazard. Often, the official who reported the statistic gathered it through telephone requests or e-mail to other Department staff. Also, for many of the statistics reported, Department officials either had not established internal controls to ensure the statistics were accurately gathered, classified, and reported, or did not document the internal controls used.

After our initial review of the 192 statistics, we excluded 133 statistics from our detailed review because: (1) the sources of the statistics could not be determined;³ (2) the statistics or supporting systems were previously reviewed by the OIG and recommendations were made to correct deficiencies identified; and (3) the statistics were used for informational as opposed to operational purposes.⁴

Of the remaining 59 terrorism‑related statistics, we selected 26 to test whether the statistics were accurate. We selected these 26 statistics based on our assessment of the significance of the statistic to the Department’s counterterrorism efforts and based on the risk associated with reporting the statistic inaccurately.⁵ The statistics selected included 10 from the FBI (reported 13 times), 11 from the USAOs (reported 20 times), and 5 from the Criminal Division (reported 9 times).

To test the accuracy of these 26 statistics, we analyzed documentation and conducted interviews with Department officials to determine if the information reported for each statistic was accurate. In some cases we reviewed documentation for each item counted in the statistic reported. In other cases we reviewed documentation for a sample of the items counted.⁶

As summarized in the following table, we determined that the FBI, EOUSA, and the Criminal Division did not accurately report 24 of the 26 statistics we reviewed.

Summary Results: Review of Selected
Terrorism-Related Statistics Reported by
the Department and its Components

Some of these statistics were significantly overstated or understated, while others were overstated or understated by minor amounts. The statistics were inaccurately reported for various reasons, including that the components: (1) could not provide support for the numbers reported for the statistics; (2) could not provide support of the terrorism link used to classify statistics as terrorism-related; and (3) could not provide documentation to show that some items counted in the statistic reported occurred in the period reported or the evidence provided showed that some items counted in the statistic reported did not occur in the period reported. We summarize our findings by component in the following sections.

FBI Terrorism Statistics

As shown in the following chart, the FBI did not accurately report 8 of the 10 FBI statistics we reviewed.

OIG Analysis of Terrorism-Related Statistics Reported by the FBI [Image Not Available Electronically]

Source: FBI budget requests and performance plans, OIG interviews with FBI staff, and analysis of documents provided by FBI staff

As shown in the previous charts, we found that the FBI significantly overstated the number of:⁷

• number of terrorism-related convictions during FY 2004,

• Counterterrorism Threat Assessments issued during FY 2004, and

• Presidential Terrorist Threat Reports issued during FY 2004.

By contrast, the FBI significantly understated the number of:

• Intelligence Assessments issued during FY 2004,

• Intelligence Bulletins issued during FY 2004,

• Intelligence products produced/disseminated during FY 2004, and

• terrorist threats tracked during FY 2003 and FY 2004.

We determined that the FBI statistics were inaccurate for the following reasons.

• The number of terrorism-related convictions was overstated because the FBI initially coded the investigative cases as terrorism-related when the cases were opened, but did not recode cases when no link to terrorism was established.

• The differences between the numbers of Intelligence Information Reports reported in the budget and performance plan and the FBI’s database likely resulted from the wording of how the database was queried by the FBI. An FBI official also told us that the number of Intelligence Information Reports reissued, corrected, and recalled, or additional information added to the Intelligence Information Reports, could also have affected the accuracy of the number of Intelligence Information Reports tracked in the database.

• The inaccurate reporting of the number of Intelligence Assessments issued likely occurred because: (1) FBI staff did not enter Intelligence Assessments in the database in a timely manner, and (2) the database was queried by calendar year or other time periods instead of by fiscal year.

• The inaccuracies in the number of Intelligence Bulletins issued were attributable to a lack of formalized procedures on how to collect, verify, and report the Intelligence Bulletins.

• The inaccurate reporting of the number of intelligence products issued occurred because no internal controls are in place for gathering and verifying the number of Intelligence Bulletins, Intelligence Assessments, and Intelligence Information Reports that have been posted to websites by the four FBI divisions. In addition, the FBI did not validate the accuracy of the information posted by the divisions to ensure that all the intelligence information products are posted.

• The number of terrorism threats tracked in FY 2003 and 2004 was inaccurate primarily because the reported statistics included threats that were counted multiple times. In addition, according to an FBI official, the total threats tracked during FY 2003 and FY 2004 did not include about 60 percent of the threats tracked by FBI field offices, the FBI’s Counterterrorism Watch Unit, or the FBI’s International Terrorism Operations Sections.

• The number of threat assessments issued was inaccurately reported because the FBI inadvertently reported some assessments issued during FY 2003 as being issued during FY 2004.

• The number of Presidential Terrorist Threat Reports issued was overstated because the spreadsheet database appears to have been queried by calendar year or for another time period instead of by fiscal year, and the request for the statistic was worded in a way that could produce different answers.

EOUSA and USAO Terrorism Statistics

As shown in the following chart, EOUSA and the USAOs did not accurately report the 11 EOUSA and USAO statistics we reviewed.

OIG Analysis of Terrorism-Related Statistics Reported by EOUSA [Image Not Available Electronically]

Source: EOUSA and USAO budgets, annual statistical reports, and Legal Information Office Network System (LIONS) data; OIG interviews with EOUSA and USAO staff; and analysis of documents provided by EOUSA and USAO staff

We found that most of the inaccurate statistics occurred because the USAOs coded the statistical data as terrorism or anti-terrorism related but did not support that a case showed any reasonable link to terrorist activity.⁸ EOUSA’s definition of the anti-terrorism program activity is not clear regarding the link to terrorism. The definition indicates the anti-terrorism program activity is meant to capture activity related to prevention or disruption of terrorist threats where the conduct is not obviously a crime of terrorism. However, the definition’s two examples indicate that the subject, target, or defendant must be reasonably linked to a terrorist activity to record the case under the anti-terrorism program activity. Taken as a whole, we believe this definition establishes that a case or defendant must have some identifiable link to terrorism to be categorized as being “anti-terrorism.”

EOUSA disagrees with our interpretation of the definition of the anti-terrorism program activity. EOUSA and USAO officials told us they believe they correctly reported defendants under the anti-terrorism program activity because almost all the defendants reported under this program activity were arrested as the result of either operations carried out to prevent terrorism or through Joint Terrorism Task Force (JTTF) activities.⁹ For example, Operation Tarmac was a worksite enforcement operation launched in November 2001 at the nation’s airports. During this operation, Department and other federal agents went into regional airports and checked the immigration papers of airport workers. The agents then arrested any individuals who used falsified documents, such as social security numbers, drivers’ licenses, and other identification documents, to gain employment. EOUSA officials told us they believe these defendants are properly coded under the anti-terrorism program activity. We do not agree that law enforcement efforts such as these should be counted as “anti-terrorism” unless the subject or target is reasonably linked to terrorist activity.

We acknowledge that some law enforcement operations and the JTTF’s focus originate in concerns regarding terrorism. In fact, with the Department’s current top priority being to prevent, disrupt, and defeat terrorist operations, much of the Department’s current law enforcement effort originates in terrorism concerns. We also believe that EOUSA could fairly and accurately report to the Department, the Congress, and the public the successes of these operations without inaccurately implying that all of the resulting cases are terrorism related. We are concerned that EOUSA’s view of the anti-terrorism category permits criminal cases arising from virtually any federal law enforcement effort, including immigration violations or border enforcement activities, to be categorized as anti-terrorism regardless of the actual circumstances. In our review of the statistics reported by EOUSA, we looked for and accepted any stated terrorism linkage. However, we found many cases involving offenses such as immigration violations, standard document fraud, or drug trafficking, where the subject or target showed no link at all to terrorist activity. Therefore, in accordance with EOUSA’s anti-terrorism definition, we did not accept those cases as having support for coding in the anti-terrorism category.

In general, we found that EOUSA and the USAOs had not established effective internal control procedures for verifying the accuracy of the Legal Information Office Network System (LIONS) data to ensure statistics reported based on LIONS data were accurate.¹⁰ More specific reasons for the inaccurate statistics included:

• The statistics on terrorism convictions were inaccurately reported because the USAOs categorized the cases against the defendants under the anti-terrorism program activity when the case was filed but did not change the categorization based upon further investigation or based on the actual evidence found or offenses for which the defendants were convicted.

• The number of defendants in cases filed under the anti-terrorism program activity was inaccurate because the number included defendants that the USAOs could not provide support for a terrorism link. In addition, the number reported included cases filed in a year other than the year reported or for which the USAOs could not provide documentation to show when the case was filed. These same problems contributed to the inaccurate statistics on terrorism convictions, terrorism-related convictions, defendants sentenced to prison, terrorism and anti-terrorism cases against defendants that were terminated, terrorism and anti-terrorism pending, defendants for which terrorism and anti-terrorism cases were terminated, and percentage of defendants whose cases were terminated that were convicted.

• The statistics reported for defendants prosecuted, defendants found guilty in FY 2002, and terrorism cases tried in 2001, were inaccurate because EOUSA provided a listing from its LIONS system that did not reconcile to the numbers reported and EOUSA could not explain the differences. EOUSA’s explanation for these differences is discussed on page 43.

In August 2006, we briefed EOUSA officials regarding the findings of this audit. The Legislative Counsel for EOUSA argued that the defendants in cases coded under the anti-terrorism code were not required to have a link to terrorism. The official stated that the cases focused on the prevention of terrorism activities in line with the Department’s number one goal of terrorism prevention.

Criminal Division Terrorism Statistics

As shown in the following chart, the Criminal Division under-reported most of its statistics, based on the support it eventually provided to us.

OIG Analysis of Terrorism-Related Statistics Reported by the Criminal Division [Image Not Available Electronically]

Source: Criminal Division budgets, OIG interviews with Criminal Division staff, and analysis of documents provided by Criminal Division staff

We determined that the statistics reported by the Criminal Division were inaccurate primarily because the database used to track the statistics was incomplete and not kept up-to-date. This database was incomplete because the Criminal Division had not established formalized procedures to: (1) instruct staff on what data should be reported in the database, and how and when the data should be reported; or (2) validate the accuracy of the information reported in the database.

Moreover, the Criminal Division provided varying sets of data to us regarding its statistics. Because the Criminal Division had not established effective procedures to gather and report accurate terrorism data, the Criminal Division had difficulty providing us a stable or reliable list of cases reported for each of the five statistics we reviewed. When we initially requested information to support the statistics we tested, a Criminal Division official told us that she could not readily provide lists to match the numbers reported because: (1) the database used to support the first four statistics tested was not up-to-date, and (2) the list the Criminal Division had for the last statistic tested did not include all the relevant cases.

As a result, in response to our initial findings about the inaccuracy of its statistics, the Criminal Division requested more time to update the database and reconstruct a list of terrorist financing cases, investigations, or matters that the Criminal Division participated in or coordinated that would support statistics reported. In March 2006, the Criminal Division’s Counterterrorism Section (CTS) provided the updated database and lists. We reviewed evidence in support of the March 2006 reconstructed data and found that the Criminal Division’s reconstructed data did not support the accuracy of the five statistics we tested for eight of the nine times the statistics were reported. Specifically, our analysis of the reconstructed data and supporting documentation showed that the statistics were overstated six times, understated two times, and reported accurately one time.

In August 2006, we briefed the Criminal Division on the results of our testing of the March 2006 reconstructed data. The Deputy Assistant Attorney General for the Criminal Division stated that while the Criminal Division needed to strengthen its controls to gather, report, and document accurate terrorism statistics, he believed the Criminal Division could provide the documentation needed to support all the statistics reported. In August 2006, the Criminal Division provided us another set of reconstructed lists to support the five statistics we tested.

After extensive analysis and discussions with the Criminal Division, we determined the August 2006 reconstructed data was also inaccurate for seven of the nine times the statistics were reported. The August 2006 reconstructed lists were inaccurate because either: (1) data from the March 2006 reconstructed lists was not included on the August 2006 reconstructed lists but should have been; or (2) relevant data was missing from both the August 2006 and March 2006 reconstructed lists. We reviewed these concerns with Criminal Division officials who agreed that the August 2006 reconstructed lists were incorrect based on the inaccuracies we identified.

Thus, during our audit, the Criminal Division made two attempts to reconstruct support for the terrorism-related statistics we reviewed. We carefully evaluated each attempted reconstruction and found each to be significantly flawed. At the end of this process, we determined that the Criminal Division understated the five statistics we tested eight of the nine times the statistics were reported. The extensive efforts required by the Criminal Division to reconstruct reported statistics demonstrates that the Division had no accurate basis for its reported numbers and could easily and unintentionally misreport in the opposite direction, absent the implementation of effective procedures for gathering and reporting statistics.

Conclusion and Recommendations

In general, we found that the Department components and the Department as a whole did not accurately report terrorism-related statistics. T he Department components lacked adequate internal controls for gathering, verifying, and reporting terrorism-related statistics.

To assist the Department and its components in improving the internal controls to ensure the accuracy of its reported terrorism-related statistics, we recommend that the FBI, Criminal Division, and EOUSA:
(1) establish and document the internal control procedures for gathering, verifying, and reporting terrorism-related statistics; (2) maintain documentation to identify the source of all terrorism-related statistics reported, (3) maintain documentation of the procedures and systems used to gather or track the statistics reported, (4) maintain documentation of the methodologies and procedures used to verify the accuracy of the statistics reported, and (5) ensure that terrorism-related statistics are not reported unless evidence is maintained to support the statistics. In addition, we recommend that EOUSA and the USAOs establish procedures to recode transactions in the LIONS system when investigations that began as terrorism investigations do not ultimately link the defendant to terrorist activity.

1. The Department received $3.6 billion in Fiscal Year (FY) 2006 for counterterrorism activities, an increase of almost 400 percent over the $737 million received in FY 2001.

2. While some terrorism-related statistics were reported only once, most were reported multiple times. Many of the statistics identified were also reported by more than one agency. We calculated the total times the statistic was reported by adding the number of times the statistic was reported by both the primary reporting agency and by other agencies.

3. They include the number of: (1) disrupted donors related to terrorist financing, (2) weapons of mass destruction cases initiated by the FBI and supported by the Hazardous Material Response Unit, and (3) number of subpoenas and search warrants issued to gather and cultivate detailed intelligence on terrorists in the United States. The statistics for which sources could not be determined are listed in Appendix II.

4. Informational purposes included statistics used in speeches, press releases, publications, and websites. Operational purposes included statistics used in budget requests, performance plans, and annual financial statements and statistical reports.

5. More details of our selection methodology are contained in Appendix I.

6. For those statistics where we reviewed a sample of items counted, the number supported could be less and the difference could be more based on a 100-percent review of transactions reported.

7. We considered the misreporting of a statistic as significant if the statistic was either overstated or understated by 10 percent or more.

8. The LIONS Manual states that the anti-terrorism program activity is meant to capture USAO activity intended to prevent or disrupt potential or actual terrorist threats where the offense conduct is not obviously a federal crime of terrorism. The manual provides that to the extent evidence or information exists, in any form, reasonably relating the case to terrorism or the prevention of terrorism, the matter should be considered anti-terrorism. The manual provides two examples that demonstrate that the subject, target, or defendant should be reasonably linked to a terrorist activity to be reported under the anti-terrorism program activity: (a) a case involving offenses such as immigration violations, document fraud, or drug trafficking where the subject or target is reasonably linked to terrorist activity; (b) a case of identity theft and document fraud where the defendant’s motivation is to obtain access to and damage sensitive government facilities.

9. The JTTFs are squads within the FBI’s field offices that focus primarily on addressing terrorism threats and preventing terrorist incidents.

10. The LIONS system is a database with on-line capabilities that permits the USAOs and EOUSA to compile, maintain, and track information relating to defendants, crimes, criminal charges, court events, and witnesses.