We provided the draft report to the FBI, EOUSA, and the Criminal Division for their comments, which are included in Appendix III. Our analysis of each component’s response is provided in Sections A through C of this appendix. In Section D of this appendix, we summarize the status of each recommendation and discuss the actions necessary to close the recommendation.

A. FBI Response

The FBI concurred generally but not entirely with our findings, and it agreed with all of our recommendations.

The FBI stated that it agreed fully with the OIG that it must collect and maintain accurate statistics. The FBI’s response then discussed its reorganization and restructuring since the September 11, 2001, terrorist attacks. It stated that many of the apparent weaknesses in statistical reporting occurred during this period and were an outgrowth of the FBI’s reorganization and restructuring. The FBI also provided an extensive discussion of new procedures it is implementing to ensure the accuracy of its reporting of terrorism-related convictions. We believe that, if fully implemented and followed, the FBI’s procedures can help improve the accuracy of its future statistics.

However, the FBI objected to certain aspects of our findings. In its discussion of the number of terrorism-related statistics, the FBI stated that it agreed that while a case must have a terrorism nexus to be properly coded as terrorism-related, the ultimate charge or conviction need not be for a terrorism offense to properly be included as a terrorism statistic.

We agree with that point and in our report stated that cases could have a terrorism link and be properly counted as a terrorism conviction even if the ultimate charge was not for a terrorism offense. As we note on page 19 of the report, we counted convictions as being terrorism-related as long as the FBI provided to us any evidence that the case had a terrorism link, regardless of the ultimate charge.

However, for many cases the FBI could not provide us with any such evidence of a link to terrorism. For example, pages 18-19 of the report provide examples of four cases included in the FBI’s terrorism-related convictions for which the FBI could not provide support for any link to terrorism.

The FBI’s response also described the work of Joint Terrorism Task Forces (JTTFs) and asserted that all cases worked by a JTTF are “by their nature terrorism related.” The FBI stated that it therefore disagrees that it should “recode investigations that the JTTFs handle based on the type of criminal charges filed at the conclusion of the case.”

The FBI misconstrues our finding. We do not believe that the case should be recoded based solely on the ultimate criminal charge, and we recognize that cases that have a link to terrorism might not result in terrorism charges. On the other hand, cases that the JTTFs pursue, based on an initial lead, might ultimately find no link to terrorism even if the JTTF finds other criminal conduct during the course of its investigation. For example, the JTTF may investigate a terrorism lead and subsequently clear the defendant of any connection to terrorism, but the JTTF might uncover evidence that the defendant committed some unrelated criminal conduct such as marriage fraud. Contrary to the implication of the FBI’s argument, the resulting conviction should not be considered a terrorism-related conviction solely because the JTTF handled the case.

We agree with the FBI that, after weighing the alternatives, a false statement or immigration case may present the best outcome to disrupt a terrorist plot rather than pursuing strict terrorism-related charges. But the FBI should be able to demonstrate a terrorism link in the cases it lists as “terrorism-related convictions,” regardless of whether the JTTF worked the case. In short, just because the case was worked by a JTTF does not mean that all its convictions necessarily contain a link to terrorism.⁶⁶

In its response, the FBI also stated that it had reviewed the 65 convictions we identified in the report as not having a demonstrated link to terrorism and agreed that 30 of the convictions were improperly reported. However, the FBI stated that it believes the remaining 35 cases were properly coded as terrorism-related. While the FBI did not provide sufficient detail in its response for us to reconcile its analysis of the 35 cases to our audit work, it appears that 29 of the 35 cases pertain to narcotics convictions of white supremacist gang members. The FBI stated that these investigations “focused on fully identifying and disrupting the members of this organization whose declared mission was to train and prepare for future violent race conflicts.” We are aware the FBI considers that domestic terrorism groups can include white supremacists, black separatists, animal rights/environmental extremists, anarchists, anti-abortion extremists, militias, and other anti-government extremists. While we agree that these groups can engage in terrorism, we also believe that these groups can engage in criminal activities such as drug trafficking that are unrelated to terrorism.

When reviewing files of FBI cases it listed as terrorism convictions, we looked for any demonstrated link to terrorism and did not consider membership in a white supremacist group alone sufficient to establish such a link. As with the JTTF cases, we believe the FBI should be able to demonstrate a terrorism link in such cases, and if it cannot point to any evidence to indicate that such a link exists, the FBI should not code the case as terrorism-related.

In addition, according to the FBI’s response, 2 of the 35 disputed cases involved threats to the public and were appropriately counted as domestic terrorism cases. One of the cases involved an individual who sent a federal judge letters that stated they were contaminated with Anthrax. Contrary to the FBI’s belief, we considered this case to have a terrorism link and did not question how it was coded. However, the other case involved an individual who telephoned local police and reported that 30 vials of Yersinia Pestis bacteria, the infectious agent of bubonic plague, could not be located. When the FBI questioned the individual, he admitted that he had accidentally destroyed the vials and made up the story in an effort to account for the vials. Because the case agent could provide no evidence that this incident was linked to terrorism, we do not believe it was properly reported as a terrorism-related conviction.

The FBI’s response did not provide an explanation for the remaining four cases disputed by the OIG as having a link to terrorism.

In the remaining sections of its response, the FBI acknowledged other inaccurate FBI statistics but described how it had strengthened controls relating to the statistics.

The status of the recommendations related to FBI statistics is presented in Section D of this appendix.

B. EOUSA Response

In its comments on the report, EOUSA agreed that the report raised important issues regarding what constitutes an “anti-terrorism case” and that EOUSA must provide the clearest possible statistical picture of the terrorism and anti-terrorism work being done by United States Attorneys’ Offices. Toward that end, EOUSA agreed to rename its anti-terrorism program category-code and “modify and clarify its definition, in order to eliminate any misunderstanding regarding its meaning.” EOUSA also acknowledged that several of its statistics were inaccurate such as statistic 5 on defendants prosecuted in FY 2002, and statistic 6 on defendants found guilty in FY 2002, and it agreed to review its internal controls over terrorism statistics to determine what improvements can be made.

However, EOUSA objected to the OIG’s findings regarding several of its terrorism statistics, and EOUSA also raised concerns regarding the methodologies we used in reaching our conclusions. We present below our analysis of EOUSA’s comments.

EOUSA noted that the OIG interpreted EOUSA’s anti-terrorism program category code definition to require that defendants in anti-terrorism cases have an identifiable link to terrorist activity. EOUSA claimed that the OIG’s interpretation would not capture a “much broader group of proactive cases that have been affirmatively and intentionally brought to deter and prevent terrorism, particularly in the areas of critical infrastructure vulnerability, regardless of whether the defendant has any links to terrorist activity.” In support of its argument, EOUSA pointed to Operation Tarmac cases, as well as all cases worked by JTTFs, as examples of anti-terrorism cases. It also argued that the OIG “deems EOUSA’s statistics ‘inaccurate’ in large measure because they [the OIG] exclude Operation Tarmac cases, and cases from similar initiatives around the country, from the anti-terrorism definition.”

First, as a factual matter EOUSA’s argument is incorrect that its statistics were “in large measure” inaccurate because of the dispute over inclusion of cases like Operation Tarmac. After receiving EOUSA’s response, we reviewed how many of the disputed cases in the anti-terrorism statistics related to Operation Tarmac or similar operations. In fact, we determined that even if such cases were counted as anti-terrorism cases, EOUSA’s statistics would still be significantly inaccurate. We analyzed 9 of the 11 EOUSA statistics which were reported 16 times. Our analysis showed that, even giving credit for Operation Tarmac and similar cases, as well as all cases worked by JTTFs, 8 of the 9 EOUSA statistics remained significantly inaccurate. Only one changed from significantly overstated to overstated by a minor amount. For the 14 times the other 8 statistics were reported, the percentage error changed but it did not change the statistic from a significant error to a minor error. Thus, notwithstanding this disputed interpretation, EOUSA’s statistics remained significantly inaccurate.

Second, we disagree with EOUSA’s argument that all convictions in cases like Operation Tarmac were properly counted as anti-terrorism convictions given EOUSA’s stated definition of the anti-terrorism program category. Moreover, we believe that if EOUSA wanted to count such cases it should have made its definitions and descriptions of the type of cases that it was counting as anti-terrorism case transparent. We believe that a fair reading of EOUSA’s definition of the anti-terrorism program category would not indicate that such cases should be included.

EOUSA’s definition of the anti-terrorism program category states that the category: (1) includes any matter or case where the underlying purpose or object of the investigation is anti-terrorism related, (2) is meant to capture activity intended to prevent or disrupt potential or actual terrorist threats where the offense conduct is not a federal crime of terrorism, and (3) includes any matter where evidence or information exists in any form reasonably relating the case to terrorism or the prevention of terrorism. The definition goes on to provide examples of cases that should be included under this category, such as immigration fraud, but also sates that the subject or target must be “reasonably be linked to terrorist activity.” Thus, in Operation Tarmac, which resulted in immigration charges, the definition suggest that the subjects must also reasonably be linked to terrorism. As EOUSA acknowledges, most of those subjects in Operation Tarmac cases were not. Moreover, none of EOUSA’s anti-terrorism statistics reported to Congress in the President’s budgets and in its statistical reports included an explanation of what EOUSA meant by “anti-terrorism” or whether the cases reported were actually linked to terrorist activity.

We recognize that efforts like Operation Tarmac may be intended to deter potential terrorists, as well as a wide range of other criminal activity. However, we believe that including all Operation Tarmac cases under its anti-terrorism category – without explanation – does not clearly provide full information to Congress and the public about EOUSA’s statistics. Rather, simply suggesting that because “hundreds of these cases were brought around the country specifically to deter potential terrorists from infiltrating regional airports” does not, in our mind, justify labeling all these cases as “anti-terrorism” absent a more identifiable link to terrorist activity. Instead, EOUSA has an obligation to clearly articulate the types of cases it is including in these categories.

We also note that after making these arguments about our analysis of its statistics, EOUSA stated that “Notwithstanding, in an effort to ensure greater clarity about the purpose and scope of the anti-terrorism category code, EOUSA shall rename the code and will, after a prompt but thorough internal discussion of the issue, modify the definition to make it more transparent.” We agree that this corrective action is needed, and we believe EOUSA’s intended action is appropriate and responsible.

Third, to support its argument that the OIG is mistaken in its position not to support inclusion of all Operation Tarmac-type cases as anti-terrorism cases, EOUSA cites a January 2003 report by the Government Accountability Office (GAO) that allegedly “reviewed EOUSA’s terrorism statistics and confirmed that Operation-Tarmac-type cases were properly coded as anti-terrorism cases.” EOUSA further claims that “OIG’s interpretation of the anti-terrorism code therefore differs not only from EOUSA’s and that of the USAOs, but with the GAO’s interpretation as well.”

In fact, the GAO report cited by EOUSA did not confirm that Operation Tarmac-type cases were properly coded as anti-terrorism cases. Rather, the GAO report concluded that the Department of Justice did not have sufficient management oversight and internal controls in place to ensure the accuracy and reliability of terrorism-related convictions.

However, upon receipt of EOUSA’s response, we contacted GAO managers regarding the report language quoted by EOUSA. The GAO managers stated that the GAO report should not be construed as an endorsement of EOUSA’s classification of such cases as “anti-terrorism.” GAO managers said their report neither confirmed that Operation-Tarmac-type cases were properly coded as anti-terrorism cases or endorsed EOUSA’s use of the anti-terrorism code.

Fourth, similar to the FBI’s response, EOUSA claims that any case investigated by a JTTF “regardless of whether the defendant has verifiable links to terrorist activity, is going to be, by definition, part of a proactive effort to prevent terrorism because that is what the JTTF does.” As noted in our response to the FBI’s comments, we disagree with this argument. An investigative lead may be pursued by the JTTF but the outcome of the investigation may clear the defendant of any connection to terrorism while finding other criminal activity. We believe it to be inaccurate to include all such convictions as anti-terrorism simply because a JTTF pursued the investigation rather than other investigators.

Fifth, EOUSA disputed “certain aspects of the methodology used in the OIG review,” and argued that it was inaccurate or misleading to report that EOUSA and USAO statistics were “unsupported” because some of the statistics were under-reported rather than over-reported. EOUSA stated that because 3 of the 11 statistics and 4 of the 19 subcategories were understated rather than overstated, these statistics therefore were “supported.”

In our view, a statistic reported to Congress and to the public by the Department of Justice is unsupported if the Department cannot provide support for the accuracy of that number, regardless of whether it is under-reported or over-reported. However, in order to avoid a non-productive disagreement about the meaning of the word “supported” in this context, the OIG has decided to change the wording in the report to state that various statistics are “inaccurate” rather than “unsupported.” However, the OIG’s ultimate conclusion remains that in all 11 of the terrorism statistics we examined, EOUSA did not report its numbers accurately. Congress, Department managers, and the public need accurate terrorism statistics in order to assess the Department’s work, and Department statistics that are either under- or overstated are unacceptable.

Sixth, when discussing our methodology EOUSA made various arguments that the OIG did not identify the criteria used to question the reporting of certain statistics or that we did not provide sufficient details of the transactions questioned for EOUSA to determine if the transactions were accurately reported or not. For example, EOUSA stated that we did not identify in the report the criteria we used to judge whether cases reported as anti-terrorism cases were linked to terrorism. EOUSA is incorrect. The report explains that our judgment of whether such a terrorism link existed was based on whether or not the USAOs could provide any evidence that tied the subjects to terrorist activity. We accepted as evidence of a terrorism link virtually any written or verbally-provided indication of a terrorism connection. Although we asked for written documentation, in practice we accepted USAO and EOUSA officials’ verbal explanations of the terrorism links.

Moreover, after we provided the draft report to EOUSA and discussed the report with EOUSA officials at the audit closeout meeting, we provided EOUSA with comprehensive lists showing every case we questioned for each statistic we reviewed. The listings identified whether the case was questioned for lack of a link to terrorism, lack of documentation to show the case was reported in the proper period, or both reasons.

In addition, EOUSA stated that we did not explain why we did not accept that cases referred to the USAOs by the JTTFs were properly coded as anti-terrorism cases. EOUSA further stated that we determined that 13 of 21 sampled JTTF cases did not have an appropriate terrorism link, but that we did not identify the basis we used to make this judgment. Again, EOUSA is incorrect. We asked USAO and EOUSA officials for documentation of such a terrorism link, and we accepted verbal explanations if they were offered. For the 13 sampled cases, no terrorism link was provided in writing or verbally by the USAOs or EOUSA other than the explanation that these cases were JTTF matters and therefore by their nature constituted terrorism-prevention cases. As explained previously, we do not accept this logic.

EOUSA also claimed that we failed to share with it case documentation we received from the USAOs to support findings that EOUSA reported cases in the proper year. This statement is also inaccurate. For each statistic we reviewed, we provided EOUSA a comprehensive listing showing whether the case was questioned because of a lack of a terrorism link, lack of documentation to show the case was appropriately included in the year reported, or both reasons. EOUSA staff did not provide support to prove that the cases we questioned were reported in the proper year.

Finally, EOUSA stated that the OIG failed to provide it with the information it needed to identify individual cases discussed in the report and therefore it was impossible for EOUSA to reply to the examples on a case-by-case basis. EOUSA’s statement is incorrect. As previously noted, for every statistic we reviewed we provided EOUSA with a comprehensive listing showing the case and the reason it was questioned. Consequently, we believe that EOUSA had the information – as well as the time it needed – to research each case and determine whether our conclusions were accurate.

The status of the recommendations related to EOUSA statistics is presented in Section D of this appendix.

C. Criminal Division Response

The Criminal Division also disagreed with our findings, although it concurred with each of our recommendations.

First, the Criminal Division argued that because it understated six of the nine statistics we reviewed, its numbers were fully “supported.” As we discussed in the analysis of EOUSA’s response, we disagree with this argument. We believe it is unacceptable for the Criminal Division of the Department of Justice to provide inaccurate statistics describing its work, regardless of whether the numbers are under- or over-reported. However, as discussed previously, in order to avoid a non-productive disagreement about the accuracy of the word “supported” in this context, the OIG has changed its wording in the report to “inaccurate” statistics rather than “unsupported” statistics.

Second, the Criminal Division’s response stated that it takes issue with the OIG’s conclusion that the Criminal Division did not have an effective reporting system. It stated that “While we concede that the reported statistics have not been maintained according to audit standards and the supporting documentation was therefore not readily available, the fact that this data could be successfully reconstructed actually demonstrated that there were systems in place, and that systems could effectively provide the basis for the reported numbers.”

We disagree. We were not applying “audit standards” to the Criminal Division’s statistical reporting systems. However, it is clear that the Criminal Division had inadequate internal controls on its reporting systems, considering the fact that it took the Criminal Division two tries at reconstruction over a 6-month period to provide support for its numbers. As described beginning on page 70 of this report, the Criminal Division initially could not produce documentation for the statistics we sought to test. Staff of the Criminal Division requested and were given time to reconstruct the records. When the first reconstruction effort failed, the Criminal Division requested and we allowed time for a second reconstruction effort.

The statistics reported continued to be inaccurate even after the Criminal Division’s second attempt at reconstruction. While the numbers were understated after the second reconstruction, we believe that under-reporting numbers in official statistics also shows a troubling lack of internal controls on the Criminal Division’s reporting systems.

With respect to the possibility of over-reporting its terrorism statistics, the Criminal Division argued that “we do not believe, given our system, that the Division could unintentionally report a terrorism-related matter that had no basis in fact.” The Criminal Division then provided a lengthy discussion of the coordination and documentation used in terrorism cases. Yet, until it attempted to reconstruct support for its statistics, the Criminal Division had not monitored the terrorism-related statistics that we tested and in fact could not do so with the data available prior to the reconstructions. In addition, internal controls did not exist to validate the Criminal Division’s statistics. Under these circumstances, we believe that inaccurate reporting in both directions is possible.

In fact, the Criminal Division’s first reconstruction attempt indicated just the opposite. Specifically, the Criminal Division’s first reconstruction included cases that were not supported by the facts and documentation for seven of the nine times that the five statistics were reported. For example, for statistic 2 on the number of individuals convicted or pleaded guilty resulting from terrorism investigations from September 11, 2001, through February 3, 2005, the first reconstruction included three cases where the investigations had been vacated and therefore should not have been included on the list. When we met with Criminal Division officials to discuss these cases, the officials agreed that these cases should not have been included on the first reconstructed list and excluded the cases from the second reconstructed list. Had we accepted the Criminal Division’s first reconstructed list, then the Criminal Division would have included cases not supported by the facts, contrary to its assertion that this could not occur.

We also disagree with the Criminal Division’s reference to audit standards. We did not apply audit standards to the Criminal Division’s responsibility to provide accurate statistics. Rather, the auditing standards to which the Criminal Division refers are broad statements of auditors’ responsibilities and provide a framework for the conduct of audit work.⁶⁷ These auditing standards apply to work of government auditors, not the work of Criminal Division managers. However, Criminal Division managers (like other government managers) should be held to standards requiring that significant events be clearly documented, that such documentation be readily available for examination, and that statistics be monitored, validated, and accurate. The Criminal Division’s difficulty in reconstructing support for its numbers, and the ultimate inaccuracy of those numbers compared to what it reported, does not show compliance with the standards to which all managers should be held.

The Criminal Division also stated that the OIG report contains inaccurate and potentially misleading statements. We disagree with the Criminal Division and discuss below each of its concerns individually. First, the Criminal Division stated that “Repeated statements that the Criminal Division statistics were ‘inaccurate’ and ‘flawed’… will likely lead readers to infer that the statistics were inflated.” We carefully wrote our report to demonstrate both graphically and in the text the extent to which the Criminal Division inaccurately reported its statistics by either under-reporting or over-reporting. We believe that our description of the Criminal Division’s statistics as both inaccurate and flawed is correct.

Second, the Criminal Division noted that we identified broad causes for inaccurate statistics, including that the departmental components we reviewed: (1) could not provide support for the numbers reported for the statistics; (2) could not provide support for 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 provided evidence showed that some items counted in the statistic reported did not occur in the period reported. The Criminal Division argued that none of these explanations pertained to its statistics.

That is not correct. As noted above, the Criminal Division’s reconstructed documentation did not support the statistics it reported because they were inaccurately understated. Yet, even beyond the dispute about the term “support,” it is clear that the Criminal Division’s statistics included items that were included in the wrong period. On page 73 of the report we summarized the Criminal Division’s transactions that we tested which did not occur during the reporting period and therefore should not have been included in this reporting period. While these numbers are small and we characterize them as such in the report, this deficiency applies to the Criminal Division’s statistics as well as to the other Department components.

Third, the Criminal Division objected to our explanation regarding how it had to reconstruct the required support for its statistics, and it noted that it provided us with assistance in understanding the information. The Criminal Division concluded that “A full understanding of the audit and reconciliation process must credit the efforts by the Counterterrorism Section to supply the OIG with all the materials they requested and educate them on each criminal investigation and prosecution so that they could properly evaluate those materials in the context of the audit.” We agree that Criminal Division staff provided us with requested materials and worked with us to ensure our proper understanding of those materials. While we appreciate these efforts (and similar efforts of the FBI and EOUSA), we view these efforts to be part of the Department’s routine compliance with the information and assistance provisions of the Inspector General Act.⁶⁸ Such interactions are common and are crucial to properly evaluate the wide range of Departmental activities covered by our audits. That said, we believe that the difficult and time consuming reconstruction process required by the Criminal Division was a significant issue and is noteworthy of mention in the OIG’s report.

The status of the recommendations related to Criminal Division statistics is presented in Section D of this appendix.

D. Status of Recommendations

1. This recommendation was directed to the FBI, EOUSA, and Criminal Division to establish and document internal control procedures for gathering, verifying, and reporting terrorism-related statistics.

FBI: Resolved. The FBI generally agreed with the recommendation but noted concerns about our analysis of terrorism-related convictions. Those concerns are discussed in Section A of this appendix. However, the FBI stated that it has improved oversight at FBI Headquarters and enhanced policy guidance to strengthen its internal controls, which will limit the improper reporting of terrorism-related statistics in the future. This recommendation can be closed when we receive documentation of the FBI’s enhanced policy guidance to strengthen its internal controls for gathering, verifying, and reporting terrorism-related statistics.

EOUSA: Unresolved. EOUSA stated that it agrees with the recommendation but believes that it is already complying with it. EOUSA’s response reiterated the five tools that we explained in the report it uses to ensure that the LIONS data is correct. However, as we also explained in the report, the USAOs are only required to use one of the five tools – the United States Attorneys’ semi-annual self-certification that the LIONS data is complete and accurate. EOUSA only suggested that the USAOs use the other four tools (case certifications by events tool, Alcatraz Case Certification report, AUSA workload reports, and Alternate District Reporting Method). EOUSA’s response also provided no data to indicate the extent to which the USAOs use these suggested tools to ensure the accuracy of LIONS data. In addition, EOUSA stated that the primary cause of the differences we questioned is our interpretation of EOUSA’s definition for its anti-terrorism code, and the differences were not the result of failures of controls EOUSA has in place.

As discussed above, that is not correct. Some of the differences we questioned were related to the anti-terrorism code, but EOUSA’s statistics were still inaccurate apart from that issue. Moreover, we also found four statistics for which EOUSA stated that it had established controls but the controls were not documented. EOUSA’s response did not address this issue. Therefore, we do not agree that EOUSA has already complied with this recommendation. This recommendation can be closed when we receive documentation showing that EOUSA has established and documented internal control procedures for gathering, verifying, and reporting terrorism-related statistics.

Criminal Division: Unresolved. The Criminal Division concurred with the recommendation but stated that it had existing procedures for gathering, verifying, and reporting terrorism-related statistics in the form of the Counterterrorism Section’s Circular on Terrorism Statistics and in material discussing specific performance measures. According to the Criminal Division, these mechanisms were supplemented by gathering information through telephone and e-mail requests and responses to and from terrorism prosecutors in the USAOs. The Criminal Division stated that it believes there was nothing haphazard about its system, which resulted in daily reporting of terrorism litigation events by which Department leadership is kept current on such matters.

As discussed above, we found the Criminal Division’s procedures for gathering, verifying, and reporting terrorism-related statistics resulted in inaccurate statistics. The Criminal Division also did not have procedures detailing the sources it would use to gather each statistic or the methodologies it would use to verify the accuracy of those statistics. As a result, the Criminal Division spent considerable resources during two attempts to reconstruct the support for the statistics reported, and in the end was unable to reconcile the numbers reported for each statistic. As a result, this recommendation can be closed when we receive documentation showing the Criminal Division has established and documented internal control procedures for gathering, verifying, and reporting terrorism-related statistics.

2. This recommendation was directed to the FBI, EOUSA, and Criminal Division to maintain documentation to identify the source of all terrorism-related statistics reported in official operational documents such as budget requests, performance plans, statistical reports, and others.

FBI: Resolved. The FBI stated that it will establish a process of maintaining supporting documentation identifying the original source of terrorism-related statistics and that its Counterterrorism and Finance Divisions will work together to establish a formal process to maintain supporting documentation and establish appropriate record retention policy. This recommendation can be closed when we receive documentation of the FBI’s process for maintaining documentation to identify the source of all terrorism-related statistics reported in official operational documents such as budget requests, performance plans, statistical reports, and others.

EOUSA: Unresolved. EOUSA responded that it already maintains documentation to identify the source of all terrorism-related statistics and refers to its response to the first recommendation. As noted in our report, we identified 16 statistics for which the source of the statistics could not be identified. Four of these 16 statistics were reported in Attorney General Testimony and appeared to be statistics that may have been collected by EOUSA or USAOs. Therefore, we believe that EOUSA needs to maintain documentation to identify the source of all statistics, even those reported in non-EOUSA documents but originating with EOUSA. This recommendation can be closed when we receive documentation showing EOUSA’s plans to maintain documentation to identify the source of all terrorism-related statistics reported in official operational documents such as budget requests, performance plans, statistical reports, and others.

Criminal Division: Resolved. The Criminal Division concurred with the recommendation and said it intends to maintain documentation in a more readily available manner to identify the source of the statistics it reports. This recommendation can be closed when we receive documentation showing how the Criminal Division plans to maintain documentation to identify the source of the terrorism-related statistics it reports.

3. This recommendation was directed to the FBI, EOUSA, and Criminal Division to maintain documentation of the procedures and systems used to gather or track the statistics reported.

FBI: Resolved. The FBI stated that its Counterterrorism and Finance Divisions will establish a process to maintain supporting documentation that will provide an audit trail on the systems utilized and any unique procedures followed to accumulate the data. This recommendation can be closed when we receive documentation showing the FBI’s process to maintain documentation of the procedures and systems used to gather or track the statistics it reports.

EOUSA: Unresolved. EOUSA responded that it already maintains documentation of the procedures and systems used to gather or track the statistics it reports and referred to its response to the first recommendation. As discussed in our analysis of EOUSA’s response to Recommendation 2, because EOUSA provides statistics for use in other documents or speeches, such as Attorney General Testimony, it should maintain documentation of the procedures and systems used to report such statistics. This recommendation can be closed when we receive documentation showing EOUSA’s plans to maintain documentation of the procedures and systems used to gather or track the statistics it reports.

Criminal Division: Resolved. The Criminal Division concurred with the recommendation and said it intends to maintain documentation of the procedures and systems used to gather or track the statistics it reports. This recommendation can be closed when we receive documentation showing how the Criminal Division plans to maintain documentation of the procedures and systems used to gather or track the statistics it reports.

4. This recommendation was directed to the FBI, EOUSA, and Criminal Division to maintain documentation of the methodologies and procedures used to verify the accuracy of the statistics reported.

FBI: Resolved. The FBI stated that documentation of the methodologies and procedures used to verify the accuracy of statistics rests with the entities that have program oversight of the systems that capture the data. The FBI further stated that much of the process for reviewing statistic accuracy centers on trend analyses conducted in the field and at FBI Headquarters and the FBI plans to continue this process. The FBI also stated that one internal control it uses to verify the accuracy of statistics is its inspection process which has identified shortcomings in various statistical accomplishments. While these controls are beneficial to improving the accuracy of reported statistics, the controls alone were not sufficient to prevent the improper reporting of statistics that we identified in this report. Therefore, we believe the FBI needs additional controls to verify the accuracy of its terrorism statistics. This recommendation can be closed when we receive documentation showing the FBI’s additional methodologies and procedures used to verify the accuracy of the statistics it reports.

EOUSA: Unresolved. EOUSA responded that it already maintains documentation of the methodologies and procedures used to verify the accuracy of the statistics it reports and refers to its response to the first recommendation. As discussed in our analysis of EOUSA’s response to Recommendation 1, EOUSA’s statistics were inaccurate even apart from the issue related to the definition of its anti-terrorism code. Moreover, we also found four statistics for which EOUSA stated that it had established controls but the controls were not documented. EOUSA’s response did not address this issue. Therefore, we do not agree that EOUSA has already complied with this recommendation and believe EOUSA should maintain documentation of the methodologies and procedures used to verify the accuracy of its statistics. This recommendation can be closed when we receive documentation showing EOUSA’s plans to maintain documentation of the methodologies and procedures used to verify the accuracy of the statistics it reports.

Criminal Division: Resolved. The Criminal Division concurred with the recommendation and intends to rely primarily on the methodologies and procedures set forth in the recently updated Circular on Terrorism Statistics. The Criminal Division will also use the additional systems described above as back-up for the international terrorism and terrorism-related cases database. This recommendation can be closed when we receive documentation showing how the Criminal Division plans to verify the accuracy of the statistics it reports.

5. This recommendation directed to the FBI, EOUSA, and Criminal Division to ensure that terrorism-related statistics are not reported unless evidence is maintained to support the statistics.

FBI: Resolved. The FBI agreed with this recommendation and stated that it is the basic premise of statistical reporting. The FBI stated that it will not report statistics which cannot be supported. This recommendation can be closed when we receive documentation showing how the FBI plans to ensure that terrorism-related statistics are not reported unless evidence is maintained to support the statistics.

EOUSA: Unresolved. EOUSA responded that this recommendation is the basic premise of any statistical reporting and that EOUSA fully agrees with it. However, as noted in the report, we found that EOUSA could not provide a listing from its LIONS system to match the numbers reported for 4 of its 11 statistics. Moreover, we also found four statistics for which EOUSA stated that it had established controls but the controls were not documented. However, EOUSA did not address this issue in its response. Therefore, we do not agree that EOUSA has adequately addressed this recommendation. This recommendation can be closed when we receive documentation showing EOUSA’s plans to ensure that terrorism-related statistics are not reported unless evidence is maintained to support the statistics.

Criminal Division: Resolved. The Criminal Division concurred with the recommendation and agrees that terrorism-related statistics should not be reported unless evidence is maintained to support the statistics. This recommendation can be closed when we receive documentation showing how the Criminal Division plans to maintain documentation of the procedures and systems used to gather or track the statistics it reports.

6. This recommendation was directed to EOUSA to establish and implement procedures to recode transactions in the LIONS system when investigations that began as terrorism-related investigations do not link the case to terrorist activity.

EOUSA: Unresolved. EOUSA agreed that a case should be recoded if it changes over the course of its investigation and it would be inaccurate to report the case under the code it was originally reported under. However, EOUSA commented that it does not see such situations arising very often. EOUSA also stated that cases brought as part of an operation to prevent terrorism are properly coded as anti-terrorism. However, EOUSA said it plans to modify and clarify its anti-terrorism code definition to eliminate any confusion as to its meaning.

As discussed above, we disagree with EOUSA’s analysis of this issue. Simply reporting the results of prevention operations as anti-terrorism in the Department’s annual budgets without explaining that the majority of the subjects had no ties to terrorist activity and were arrested for immigration violations or identity theft does not clearly and fairly characterize these operations. This recommendation can be closed when we receive documentation showing EOUSA has established and implemented procedures to clarify in the LIONS system the type of cases included in the anti-terrorism category, and to clarify what these statistics represent.

66. This issue is also discussed in the section of our analysis examining EOUSA’s comments.

67. General Accounting office, Government Auditing Standards, GAO-03-673G, June 2003.

68. See 5 U.S.C. §6 Authority of Inspector General; information and assistance from Federal agencies;