An Investigation of the Belated Production of Documents
in the Oklahoma City Bombing Case
March 19, 2002
Office of the Inspector General
ANALYSIS OF THE CAUSES FOR THE
BELATED PRODUCTION OF DISCOVERABLE ITEMS
The OIG investigated why more than 1,000 items that were within the FBI's possession never reached the defense before McVeigh's and Nichols' trials. In particular, we carefully investigated the question of whether the United States had intentionally withheld exculpatory material from the defense, something the defense asserted and a significant portion of the public appeared to believe.
In seeking to determine the cause of the belated production of the OKBOMB documents, we examined the FBI's explanations for the failure. In the days immediately following the revelation of the OKBOMB documents problem, the press reported that FBI officials were blaming the FBI's "outdated" computer system for the failure. Some FBI officials cited the transition from one computer system to another during the OKBOMB investigation as the chief cause of the problem. In addition, senior FBI officials also placed the blame for the mess squarely on the field offices. A press statement from OKBOMB Inspector in Charge Danny Defenbaugh emphasized that the OKBOMB Task Force had repeatedly instructed the field offices to send all investigative material to the Task Force. Defenbaugh also stated that the FBI's 2001 review of the field offices' OKBOMB material "was the first time the OKBOMB task force had seen these materials…." This statement was reiterated in FBI Director Louis Freeh's testimony to Congress during a May 16, 2001, hearing. Although at one point Freeh acknowledged that the Task Force may not have properly accounted for documents it received, Freeh's primary explanation was that the field offices had failed to perform in a variety of ways.
To the OIG, the field offices denied that they were responsible for the problem. For the most part, they insisted that they had sent OKBOMB material to the OKBOMB Task Force as required. In fact, they noted that they often sent the same material multiple times to the Task Force. The field offices blamed the OKBOMB Task Force for losing the items they had sent. We heard many times during the course of our investigation that field office personnel were incensed at being blamed for the belated production of the documents.
Significantly, and contrary to FBI Headquarters' initial claims, we found that both the OKBOMB Task Force and the field offices were responsible for the defense not receiving the material. Most of the breakdown was caused by mistakes by individuals, failures to follow FBI procedures, inconsistent interpretations of FBI procedures, lack of an adequate quality control system, and by the immense volume of material that put stress on a complicated document management system. As we describe more fully in the following sections, document flow in the field offices, in the Task Force, and between the field offices and the Task Force was complex. These complexities increased the opportunities for documents to be mislaid and lost.
With respect to intentional misconduct, we found no evidence that the FBI or the prosecutors intentionally withheld information they knew to be discoverable from the defense. In a very few instances, individual FBI case agents failed to submit items to the OKBOMB Task Force because they believed that the information was non-pertinent, and they erroneously assumed that the information was therefore non-discoverable. Although these items should have been disclosed to the defense, we do not believe that the incidents evidence anything other than the mistaken judgment of a few agents.
We found no particular pattern associated with the belated documents other than the date of creation. The items came from almost all the field offices, one Legat, and one office in FBI Headquarters. Hundreds of different agents, supervisors, and support personnel were involved. Items that later were determined to be belated came from different sets of files: the main Auxiliary Office file, subfiles, Elsur, and confidential informant materials were all involved. We did find that the vast majority of the belated documents had been created in the beginning of the OKBOMB investigation. As shown on the following page, in our analysis of 837 belated documents (the "first wave"), we found that 83 percent originated in April and May 1995.
Table 1: Date Belated Documents Created
|Period||Number of Documents||Percent of Total|
|Date Cannot Be Determined||23||2.7%|
The most plausible explanations for this singular finding are that a significant quantity of the investigation occurred in the months immediately following the explosion and because the investigation was unusually intense and fast-paced in those months. During this period, the focus was primarily on getting the information to the Task Force rather than document management in anticipation of discovery.
In Chapter Four we describe in depth the chronology of events leading to the discovery of the belated documents. We briefly describe a few of those events here, however, so the reader will understand some of the references that we make in this chapter. We refer to "first wave" and "second wave" materials. The first wave consists of materials that were sent from the field offices to Oklahoma City prior to May 11, 2001. The second wave consists of materials sent after May 11 when the belated documents problem became public knowledge and FBI Headquarters instructed the field offices to send all OKBOMB materials to Oklahoma City. We also refer to a May 12, 2001, EC from FBI Headquarters that instructed 46 field offices and 1 Legat to explain in writing why they had failed to send OKBOMB materials to the Task Force before the OKBOMB trials. We refer to the field offices' responses to the EC in this chapter and throughout this report. ZyIndex, the OKBOMB Task Force's full-text retrieval database system, was a critical part of the 2001 review project that led to the determination that documents had not been properly disclosed to the defense. We also used ZyIndex as part of our analysis of the belated documents, and we periodically refer to that system.
We begin our analysis with a discussion of the discovery agreement and the process used to compile discovery.
In all criminal trials, the government, which is prosecuting the defendants, must provide certain information to the defense prior to trial. In the federal system, discovery is governed by Rule 16 of the Federal Rules of Criminal Procedure; the Jencks Act, codified at 18 U.S.C. § 3500; and case law, the most significant of which is the Supreme Court decision of Brady v. Maryland, 373 U.S. 83 (1963). In general, these rules and court decisions require that the government provide the defense with all evidence that will be used at trial, the statements of individuals who will testify during the trial, and evidence in the possession of the government that is materially favorable to the defense, including information indicating that the defendant did not commit the crime charged and evidence that could be used by the defense to impeach the government's witnesses. In essence, these rules require the government to disclose some information about its case to the defense, but the defense is not entitled to review the government's entire case file.
Although the broad outlines of discovery are the same in all federal trials, actual discovery practices vary considerably depending on the practices of district courts and the United States Attorneys' Offices. In some areas, federal prosecutors allow the defense to view most of the case file while in other jurisdictions the discovery is limited.
The question of what constitutes exculpatory information required to be disclosed to the defense, commonly referred to as Brady information, is the source of most discovery problems. The main responsibility for reviewing the prosecution file and disclosing Brady information rests with the prosecutors. In certain instances, a judge may review material and make a ruling as to whether it needs to be disclosed to the defense. The government commits prosecutorial error if it fails to disclose information to which the defense is entitled, and the defense may allege that the government has violated its discovery obligations even after a conviction. In some instances, the error may be so egregious that the court determines that the defendant did not receive a fair trial, reverses the conviction, and grants the defendant a new trial.16
After the defendants' indictment in August 1995, the Task Force began preparing for the discovery process. The government agreed to provide the defense with all FD-302s and inserts, which went far beyond the usual discovery practice in federal criminal trials. Initially, the prosecutors only intended to provide FD-302s. After reviewing numerous inserts, however, they realized that many of the inserts memorialized substantive interviews and were substantively indistinguishable from the FD-302s. The government also agreed to provide the defense with a list of 1As, 1Bs, and 1Cs from which the defense could determine whether it wanted to see the actual physical item. Agents' notes and internal communications such as teletypes were not discoverable.
We were told by the participants that the prosecutors and the senior FBI officials on the Task Force agreed that the government should provide more discovery than was required by federal discovery rules. The prosecutors believed that broader discovery would reduce the number of discovery issues that would likely arise in a case of this magnitude. The prosecutors also told the OIG that they believed the trial judge would have ordered them to make expansive discovery because a substantial percentage of the government's file consisted of information pertaining to the possible complicity of other individuals. They also believed that it would show the government's desire to ensure a fair trial for defendants accused of committing such a heinous crime and facing the death penalty.
The discovery agreement was not set forth in writing. According to the prosecutors, the government was unable to get the defense to agree in writing to the terms of a reciprocal discovery agreement. Nonetheless, OKBOMB Task Force personnel understood that the government was obligated to provide to the defense all FD-302s and inserts and allow access to physical items, such as 1As and 1Bs.
The FBI primarily handled the logistics for compiling and disclosing the discoverable material. Beth Wilkinson, an attorney from Main Justice, handled most of the legal issues involved in the discovery process. Wilkinson told the OIG that she did not have to make many decisions about what would be turned over, since the government agreed to provide the defense with essentially everything. Linda Vernon, an FBI Financial Analyst with the Oklahoma City Division, assisted with the discovery process beginning in September 1995 and by April 1996 had assumed responsibility for organizing the process.
The material that was to be turned over in discovery was compiled by copying the serialized FD-302s and inserts that were contained in OKBOMB Task Force subfiles D and E. Vernon created a database to keep track of the items produced during discovery.
For the 1As, Vernon transferred the information identifying the 1As from ACS into a separate database. OKBOMB Task Force agents reviewed the database information to identify which 1As were discoverable. Printouts of the database were provided to the defense, and the defense could select the items it wished to view.
The government made its first large discovery production in January 1996 when it disclosed approximately 11,000 items. Thereafter, additional discovery material was provided on a "discovery day" once a month.
By August 1995 the Task Force realized that it was not receiving everything it needed from the field offices, and it tried to remedy the problem by sending out numerous teletypes instructing the field offices to provide all OKBOMB materials to the Task Force. These repeated requests for documents sought to address two problems. First, in some instances, the Task Force had no record of receiving any information from the field in response to a lead that the Task Force had sent. Second, the Task Force had received information (generally by teletype) but not in the form that it wanted (an FD-302 or insert).
The Task Force's requests to the field varied in specificity and clarity, but taken as a whole they expressed the point that all original FD-302s, inserts, and 1As were to be sent to the Task Force. Although we do not describe every one of the teletypes going to the field, we set forth many of the most relevant.17
A teletype from OKBOMB to all field offices dated August 11, 1995, stated:
For information of all divisions, on August 10, 1995, a federal indictment was returned against Timothy McVeigh, Terry Nichols and Michael Fortier. Discovery in this matter is tentatively set for the last week of August. It has come to the attention of the Oklahoma CP [Command Post] that there are still instances where leads previously set have not been covered, original copies of FD-302's have not been forwarded to Oklahoma CP, 1A and 1B material has not been forwarded to Oklahoma CP, and executed copies of subpoena's [sic] have not been returned to Oklahoma CP. All field divisions are requested to have all leads covered, original copies of reports, 1A and 1B material … to Oklahoma City CP no later than August 15, 1995.
On September 6, 1995, the Task Force instructed the field that all OKBOMB information was to be reported to the Task Force even if the field believed the information had no investigative importance.
Because of the need to get information to the Task Force quickly, field offices routinely sent information via teletype. In December 1995, the Task Force instructed the field offices that FD-302s and inserts had to be prepared for information that the field had reported by teletype. A December 11, 1995, teletype from the Task Force explained that it was reviewing all teletypes sent by the field to ensure that all interviews reported in teletypes corresponded to an FD-302 or insert. The Task Force reported that 450 leads had been sent to the field requesting FD-302s or inserts. It estimated that as many as 2,000 to 3,000 interviews may not have been properly reported by FD-302 or insert.
On January 15, 1996, the Task Force reminded field offices that all investigative results must be reported by FD-302 or insert: "It is absolutely imperative that results of investigation, reported via FD-302 or insert, be provided to the OKBOMB CP in the most expeditious manner possible." This teletype was sent to the personal attention of each SAC. The Task Force also instructed the field to send in all originals, 14 copies, and a disk. In a teletype dated February 23, 1996, the Task Force noted that 1500 leads had been sent out requesting FD-302s or inserts. The Task Force also pointedly reminded the field that FBI policy required that the interviews be memorialized on either FD-302s or inserts.
An October 31, 1996, teletype sent to the personal attention of each SAC requested that each SAC ensure that all FD-302s and inserts had been sent to the Task Force because the documents had to be turned over to defense counsel.
On November 13, 1996, Judge Richard Matsch, the trial judge in the McVeigh and Nichols cases, held an extensive hearing to address complaints by the defense about discovery issues.18 Judge Matsch ruled that FD-302s, inserts, lab reports, and other miscellaneous reports must be disclosed to defense counsel by December 15, 1996. The Task Force informed the field offices by teletype dated November 13, 1996, that any investigation requested by the Task Force before November 1, 1996, that was not completed and reported to defense counsel by December 15, 1996, could not be used during trial. The Task Force instructed the field offices and the FBI laboratory to complete and send to Oklahoma City any outstanding investigative work and documentation by December 1, 1996. During the discovery hearing the prosecutors advised the court that all documents and items of evidence had been disclosed.
Yet, on November 14, a defense attorney alleged to the court that the government had not disclosed surveillance logs. When the OKBOMB Task Force checked with the Phoenix Division, the Task Force learned that Phoenix had not sent the surveillance logs to the Task Force. This incident led to a November 15, 1996, teletype from the office of the FBI Director to all FBI field offices and Legal Attachés stating:
It will be the responsibility of all SACs and Legal Attaches to ensure any FD-302, insert, memo, EC, 1A, 1B, 1C, Elsur recording or document, bulky evidence log/document, photo, photo log, surveillance log, polygraph report, and any other item, not specifically described which pertains to the OKBOMB investigation, be forwarded to the OKBOMB command post in Denver by COB 11/29/96.
It will further be the responsibility of each SAC and Legal Attache to confirm by telephone, then follow-up with a teletype, advising the OKBOMB Command Post what items are being forwarded, or that their specific division/Legat has no pertinent items remaining under their care and control.
The teletype also stated that the heads of the offices were to check various components within their offices to "make absolutely sure no OKBOMB related material, not previously forwarded to OKBOMB, remains within their divisions." (We show this teletype at Exhibit 9 at p. A-25.)
Despite the Task Force's attempts, as set forth above, to gather OKBOMB material from the field offices, organize it, and disclose it to the defense in a timely fashion before trial, numerous documents were first disclosed to the defense in May 2001, just prior to McVeigh's scheduled execution. Any analysis of the belated documents should include a description of the number of documents at issue. We found, however, that it was impossible to identify the exact number of OKBOMB documents in the possession of the FBI from 1995 to 1997 that were not disclosed to the defense until 2001.
In May 2001 the government provided the defense with 1,033 documents and physical items that the government conceded should have been disclosed to the defense pretrial but which had not been disclosed. Initially, on May 9, 2001, 715 items were disclosed to the defense.19 These items constituted the materials sent to the Oklahoma City Field Office by the field offices and Legats prior to May 9, 2001. The Denver and Baltimore Divisions submitted materials to Oklahoma City a few days after the disclosure of the 715 items, and their materials added another 122 items for a total of 837 "first wave" items. The field office "second wave" consisted of 196 additional items. These were the items found in the files sent by the field offices in May 2001 after the belated documents problem became public. It also includes 39 items found when the OKBOMB files and the materials kept in the storage warehouse were reviewed. In sum, by the end of May 2001, 1,033 items, consisting of over 4,000 pages, had been disclosed to defense counsel.
Table 2: Belated Documents by Type and Wave
|Document Type||Wave One||Wave Two||Total|
|Insert or 302||583||189||772|
Table 3: Belated Documents by Office
|Division||First Wave||Second Wave||Total|
|Domestic Terrorism Operations||15||15|
|Salt Lake City||18||18|
Table 4: Belated Documents for the 13 Field Offices Visited by the OIG
|Division||First Wave||Second Wave||Total|
Only three field offices - Jacksonville, Newark, and Springfield - did not possess in their files discoverable documents that had not been disclosed. Two of the three (Jacksonville and Springfield), however, reported to the OIG that they had destroyed some portion of their OKBOMB files prior to sending the file to Oklahoma City.
We found that the 1,033 number overstated the problem by at least 38 documents because some documents that were included in the list had in fact been disclosed to the defense before trial or were duplicates of other belated documents. The actual number of belated documents is probably even smaller due to other factors that we discuss in Exhibit 10 at p. A-31.
On the other hand, the 1,033 does not reflect other documents that possibly should have been included on the list. Thirteen field offices destroyed documents following the December 20, 2000, EC from the FBI archivist that authorized field offices to destroy OKBOMB documents, and two other field offices destroyed documents prior to the archival process starting. As we discuss further in Chapter Six, if the documents had been analyzed rather than destroyed, the number of belated documents might have increased.
Accordingly, although the FBI and prosecutors disclosed 1,033 items to the defense in May 2001, the OIG investigation determined that the actual universe of discoverable items that should have been disclosed to the defense could not be determined with complete accuracy.
The items disclosed reflect a wide variety of information. Some of the documents memorialize interviews while others reflect different activity performed by agents, such as conducting records checks or reviewing magazine articles. Over 100 documents concern individuals who were reported to resemble John Doe #2.20 Approximately 30 items mentioned McVeigh, Nichols, or Fortier in some manner, including interviews of their family members. Physical items that were disclosed include photographs, printouts of records checks, and drivers' licenses.
The government asserted that much, although not all, of the information contained in the belated documents had been disclosed to the defendants before their respective trials through other means. For example, the government asserted that information obtained from an interview of an individual who knew McVeigh was substantially provided through another interview of the same individual that had been disclosed to the defense. All of the courts that reviewed McVeigh's claims regarding the belated documents concluded that the contents of the belated documents did not negate his guilt. (See, for example, Order, June 6, 2001, Judge Richard Matsch, United States District Court, "Whatever may in time [be] disclosed about possible involvement of others in this bombing, it will not change the fact that Timothy McVeigh was the instrument of death and destruction.")
In interviews with the OIG and in their written explanations to FBI Headquarters, the field offices asserted that they had followed procedures, that they had sent all materials to the OKBOMB Task Force, and that the belated production was the fault of the Task Force. We did not agree entirely with the field offices' assessment of their performance. The evidence shows that the field offices must shoulder some, although not all, of the responsibility for the belated production of the documents.
In the discussion to follow, we first discuss the FBI field offices' explanations as to why they believed documents were not disclosed to the defense. These explanations were provided in the field offices' responses to an FBI Headquarters' EC dated May 12, 2001, OIG interviews in 13 field offices, and responses to the OIG survey that was sent to 43 field offices.
We then discuss the evidence that leads us to conclude that the field offices failed to send some items to the OKBOMB Task Force as required. The last part of this section on field offices discusses the causes for the field offices' failures. Although we could not identify the definitive cause with respect to most documents, we identified a number of factors that we believe contributed to the problem.
In determining the reasons why documents had not been disclosed to the defense timely, we carefully considered the written explanations provided by the field offices to the FBI and to the OIG, as well as information we learned during our interviews of field office personnel. The field offices generally offered explanations that touched on two themes:
Many field offices stated that they could not provide any explanation because they no longer had access to their OKBOMB files, they were unable to access ACS, or ACS did not contain the needed information.21 A number of field offices did attempt to interview individuals who had been involved in processing OKBOMB documents. Some of the offices were successful in this effort, but others were hampered by retirements and transfers of staff.
In a few instances, field offices provided a specific explanation for the failure to transmit the documents previously. For example, the Baltimore Division stated that a copying error had caused its failure. In a very few instances, the field office acknowledged or implied that an agent made or may have made a specific decision not to send the material to the OKBOMB Task Force because the agent believed that the material was not pertinent to the OKBOMB investigation.
Some offices argued, however, that the problem was not their failure to send items to the OKBOMB Task Force in the 1995 to 1997 time frame, but rather that the item should not have been disclosed to the defense in 2001 and not counted as a belated document. For example, the San Francisco Division stated that one of its belated documents was really a draft that would never have been disclosed during discovery and therefore should not have been counted as a belated document.
The vast majority of the field offices said they believed they had sent everything to the OKBOMB Task Force before the defendants' trials as they had been required to do. One field office strongly responded to FBI Headquarters by rejecting a statement made in a communication sent from FBI Headquarters to the field offices: "SAC [ ] notes that HQ EC received Sunday 5/13/2001 … states '… never forwarded to Oklahoma City until recently' is entirely inaccurate."
Despite the field offices' strong assertions that they timely sent all OKBOMB materials to the Task Force, we believe the evidence shows that some material that should have been sent to the Task Force remained in the possession of the field offices.22
In determining whether an item had been sent to the OKBOMB Task Force, two factors were most probative:
We considered whether a belated document contained indicia indicating that it was more likely an original or more likely a copy. If it appeared to be an original item, the OIG believed that more likely than not, the field office had not sent the document to the OKBOMB Task Force. We reached this conclusion because FBI policy required originals to be sent to the Office of Origin and the field offices appeared to generally follow that policy. The presence of an original in an Auxiliary Office file indicated to us that procedures had broken down.
In determining whether a document was an original, we first looked to see whether the document contained markings consistent with the document being a copy, that is, a block stamp or index markings. If the document did not have those markings, we looked for agent initials in original ink. Only if original ink initials were present did we consider the document to be presumptively an original. With respect to the first wave FD-302s and inserts, we found that 305 (52 percent) have indicia indicating they are originals.
We examined this issue more closely with respect to the 13 field offices that we visited. We showed the field offices the items that we believed were originals.23 In a few cases, the offices agreed that the items appeared to be originals. They conceded that the items should have been sent to the OKBOMB Task Force, and they stated that they had no explanation as to why the items had not been sent. In other cases, the offices asserted that the documents, while appearing to be originals, could actually have been copies because in some instances agents placed original initials on document copies.24 Some offices also asserted that even if the document was an original, they nonetheless might have sent a copy to the OKBOMB Task Force. We acknowledge that for any particular document this speculation of what could have happened might be true.25 However, since we had no evidence to support these theories with respect to most of the documents, we considered a document that bore the indicia of an original to mean that more likely than not the field office had failed to send the item to the OKBOMB Task Force.
We also looked for airtels referencing the belated items to determine whether they had been sent to the Task Force before the defendants' trials. When sending items to the OKBOMB Task Force, the field offices generally sent an attached airtel, essentially a cover memorandum, describing the items being sent.26 In some cases, either the field office or the OIG found airtels sent from a field office to the OKBOMB Task Force that referenced documents that later were construed as being belated documents. In those instances, we agreed with the field office that the item probably had been sent.27 Of the 305 original items, we found airtels or teletypes referencing 4 of them and stating that the field office was sending the document to the Task Force. For these 4 items then, either the field office sent a copy to the Task Force and kept the original, or it had failed to send the document even though it sent an airtel saying the document was attached. Because we could find no evidence indicating that the remaining 301 had ever been transmitted to the Task Force, we believe that at least 301 FD-302s and inserts were originals that had never been sent to the Task Force and remained in the field offices' possession until 2001.
The OIG interviewed agents, supervisors, and administrative personnel in 13 field offices, including former employees. We also reviewed the explanations provided by the field offices in response to the OIG survey and Headquarters' May 12, 2001, EC. While we were very successful obtaining information regarding the field offices' actions in 2000 and 2001 (see Chapter Six), we were less successful in determining why material was not sent to the OKBOMB Task Force in 1995-1997. Many of the key participants are no longer with the FBI and even those that we were able to interview understandably have little recollection of their actions with respect to particular documents.
Accordingly, with regard to most of the individual documents, we could not identify for certain the specific reasons why they were not sent to the Task Force. However, we did identify a number of factors that likely contributed to the problem, including the use of teletypes, the involvement of Resident Agencies, and difficulties with whether non-pertinent information should be sent to the Task Force.
The one theme that emerged from our analysis of these factors is that the flow of paper in the FBI is complicated and various OKBOMB procedures added to that complexity. Breakdowns could have occurred easily because of the number of employees involved in handling OKBOMB documents. Various investigating agents, each office's OKBOMB case agent (and over time there may have been more than one case agent for any particular field office), squad supervisors, agents and supervisors in Resident Agencies, secretaries, rotors, and other support personnel all had responsibilities for ensuring that the paper flowed properly. Generally, we found that while individuals knew what to do, they occasionally made mistakes, such as believing that they were processing a copy rather than an original or mistakenly assuming that someone else had sent a document to the Task Force. Adding to the complexity in this case was the fact that paper often traveled back and forth between the field offices and the Task Force, thereby increasing the opportunities for documents to be mislaid or lost.
We note that many of the employees that we interviewed expressed surprise that documents they drafted or handled wound up as part of the belated documents. They emphasized that they had diligently tried to follow procedures and had put great effort into ensuring that all OKBOMB investigative tasks were performed correctly and timely.
One significant factor contributing to the belated documents was the use of teletypes to transmit information to the OKBOMB Task Force. Teletypes were used frequently in the early stages of the investigation when the primary goal was to get information to the OKBOMB Task Force as quickly as possible. When responding to an OKBOMB lead, the field office would summarize the interview and send it to the Task Force using a teletype. However, FBI policy and the discovery procedures used by the Task Force required that investigative activity be recorded on an FD-302 or insert. As a result, while information on the teletypes was provided to the Task Force, it was not in a form that was used for discovery and, therefore, the information in the teletypes was not provided to the defense. We found 171 belated FD-302s and inserts where the OKBOMB Task Force received a teletype with the information but not the corresponding FD-302, insert, or 1A.
In December 1995 the OKBOMB Task Force realized that it had a problem with the teletypes and tried to remedy it. The OKBOMB Task Force sent a lead (lead number 13725) to the field offices instructing them that the information contained in teletypes had to be converted to FD-302s or inserts.
A review of teletypes forwarded to Oklahoma City Command Post has revealed several instances where interviews have been reported.
A check through Rapid Start system and serials within the Oklahoma City Command Post has failed to locate a copy of an FD-302 or insert to correspond with the reported interview. It is believed in many cases the interview was reported by teletype, but that no FD-302 or insert was generated. It is necessary that all interviews be reported in FD-302 or insert format.
Further, the Task Force specifically identified with respect to each field office the teletypes at issue. Eventually, the Task Force sent over 1500 leads to accomplish this project.
Yet, many field offices did not provide the requested FD-302s or inserts.28 For the 171 mentioned above, the FD-302s or inserts had been created, as shown by the fact that they were later found in the Auxiliary Office file and determined to be belated. Accordingly, these were not situations where the field office failed, for whatever reason, to move the information from the teletype to the FD-302. Of the 171 belated documents that have corresponding teletypes, 56 have indicia indicating they are originals and have no associated airtels showing the document was sent to the Task Force.
We do not know whether the field offices were ignoring the requests, disagreed with the Task Force regarding the necessity for the documents, or did not fulfill the request because of the crush of work or for some other reason. One of the OKBOMB supervisors, Mark White, recalled an incident where he called a field office supervisor because the OKBOMB Task Force was having trouble getting an FD-302. White said the supervisor resisted sending the FD-302 because, in the supervisor's view, the information was not "testimonial." White said he eventually had to read the FBI's operations manual to the supervisor to prove that there was no requirement that information be "testimonial" before it was placed on an FD-302. White also said that he believed that some offices probably thought they had put information on FD-302s and transmitted them when they actually had not.
Resident Agencies generated many of the documents that were determined to have not been disclosed to the defense.29 The involvement of Resident Agencies added additional steps, increased the number of individuals involved in processing documents, and furthered the complexity of getting documents and other items to the Task Force.
Generally, the "Headquarters City" was responsible for handling the paperwork of the Resident Agencies.30 An agent in a Resident Agency would draft the FD-302, usually a supervisor in the Headquarters City would review it, and the Headquarters City rotor would process the paper. The Resident Agency's files regarding pending cases were kept on the rotor in the Headquarters City, although Resident Agency agents frequently kept "working copies" in the Resident Agency. However, seven offices reported in response to the OIG's survey that the Resident Agencies sent OKBOMB material directly to the OKBOMB Task Force without going through the Headquarters City.
The process broke down in some instances. In a few offices, we received contradictory answers regarding Resident Agency procedures, indicating that confusion over who was doing what might have caused some documents to be filed in the Auxiliary Office files rather than sent to the Task Force. The Boston OKBOMB case agent told us that the Resident Agencies sent materials directly to the OKBOMB Task Force. Yet, an agent who had drafted one of the belated documents and who had worked in a Boston Division Resident Agency told us that the Boston Division Resident Agencies sent OKBOMB materials to Boston. As we discuss in Section V B 6 of this chapter, the belated document found in the Boston files was likely an original that was mistakenly treated by the rotor as a copy and filed in the Auxiliary Office files. In addition to Boston, our interviews with personnel in Los Angeles and New Orleans also revealed inconsistencies about whether the Resident Agency or the Headquarters City was responsible for sending documents to the OKBOMB Task Force.31
Many agents speculated that the rotors somehow erred and failed to forward the document to the Task Force. And indeed, although we do not know exactly how the process broke down with respect to each document, one possible explanation is that a rotor mistakenly assumed that what the rotor had received was a copy and that the original had been sent to the OKBOMB Task Force. Accordingly, the rotor might have incorrectly filed the item in the office's OKBOMB files without sending it on to the Task Force. One field office supervisor suggested that the problem may lie with the Resident Agency sending the material to the wrong place in the Headquarters City. In his field office the Headquarters City case agent was responsible for sending material to the Task Force; therefore, if the material went directly to the rotor, it would be filed but not sent to the Task Force because it never got to the Headquarters City case agent.
We found instances among the belated documents where the document in question had been generated in response to a lead generated by an Auxiliary Office rather than the OKBOMB Task Force. Although most leads came from the OKBOMB Task Force, some leads were generated by Auxiliary Offices, particularly other command posts. The Indianapolis-Detroit documents discussed in Section V B 7 are an example of this problem. Detroit sent the lead to Indianapolis. Indianapolis responded and sent the original FD-302 to Detroit. Detroit, however, failed to send it on to the Task Force. Because we found so few examples involving this chain of events, we believe that field office personnel generally understood how to handle documents associated with Auxiliary Office leads.
Confidential informant files added another set of complexities to the document production process. Over 10 percent (108 documents) of the belated documents recorded agent contact with confidential informants. For documents relating to information provided by a confidential source, the original insert or FD-302 contained the source's name and was placed in the secure confidential informant files. A copy with an identifying source number rather than the source's name should have been placed in the field office's OKBOMB file and another copy with the identifying source number sent to the OKBOMB Task Force. The agent responsible for drafting the document was usually responsible for indicating on the FD-209 what files the document should be placed in, and the agent, the rotor, or the confidential files support employee was usually responsible for ensuring that the copies were distributed to the proper locations.32
If the process broke down and the insert did not get to the Task Force, then the error would almost certainly be missed. Indeed, the office would have a hard time even locating material related to OKBOMB in its confidential informant section absent a file-by-file search. The computer system used for tracking confidential informant files has severe limitations (see footnote 85) and is extremely difficult to search.
In the prior examples, agents intended to send the documents to the OKBOMB Task Force but because of inadvertent mistakes, the documents were not sent. We did find that in some instances agents made conscious decisions not to send material to the Task Force. In such instances, agents perceived information they had gathered to be non-pertinent or insignificant to the OKBOMB investigation, and they believed the document did not need to be sent.
For example, the Charlotte Division stated that its OKBOMB case agent had determined that a statement referencing the bombing that was written on a postcard was a political statement and did not meet the standards for sending to Oklahoma City.33 The Albany Division OKBOMB case agent did not provide a letter from an inmate referencing the Oklahoma City bombing because he did not believe the letter was connected to the OKBOMB investigation. (We discuss this document further in Section V B 5 of this chapter.) In a written explanation to the OIG, the Miami Division stated that it had not provided the contents of several 1As (such as photographs and investigative notes) to the Task Force because Miami personnel deemed the documents to have not been pertinent to the OKBOMB investigation.
Other field offices implied that they intentionally retained documents because personnel did not believe the documents were pertinent to the OKBOMB investigation. Milwaukee gave as one of two possibilities the explanation that the documents may not have been given to the OKBOMB Task Force because they contained only "negative" information (for example, that no information was obtained after a name had been checked through an informational database). Omaha implied that its belated documents may not have been sent because of the irrelevance of the information contained in them. Similarly, the Mobile Division noted that the November 15, 1996, Director's EC instructed field offices to send all "pertinent" information. Mobile stated that while its belated documents may have been inadvertently overlooked, it also was possible that the items were not sent because they were non-pertinent. In addition, we found written contemporaneous references in teletypes from the OKBOMB Task Force and ECs from the field indicating that in 1995 some field offices were sending only what they deemed as pertinent information.
We did not attempt to determine whether these field offices correctly assessed that the documents were not pertinent to the OKBOMB investigation because the Task Force instructed the field offices to send even non-pertinent information to it. In a teletype dated August 17, 1995, the OKBOMB Task Force wrote to the field offices:
Oklahoma City CP has received telephone calls from various divisions questioning the need to send in 1A and 1B material which those divisions do not deem important. Additionally, it has been revealed some divisions have not produced FD-302s or inserts reflecting interviews in which the information was deemed unimportant by that division.
To clarify this situation, all field offices are requested to send all 1A and 1B material to Oklahoma City CP. Additionally, if an interview was conducted, an FD-302 or insert must be generated and sent to Oklahoma City CP as soon as possible.
On September 6, 1995, the OKBOMB Task Force reiterated this request:
Oklahoma City Command Post has received telcalls from certain divisions concerning the following situation: during the course of the OKBOMB investigation field divisions received information directly at that division, mostly through telephone calls from local citizens. These divisions conducted investigations into these matters without notifying Oklahoma City Command Post. Additionally, no lead control number was obtained. These field divisions determined these investigations were of no investigative importance to the OKBOMB matter and therefore have not sent copies of these reports to Oklahoma City Command Post.
For information, all investigation in captioned matter requires control numbers and all reports must be forwarded to Oklahoma City Command Post. Should divisions have reports from unproductive investigations that were not previously reported to Oklahoma City, these divisions must contact Oklahoma City Command Post, obtain lead control numbers, indicate those numbers in their reports and forward these reports to Oklahoma City.
Yet, the explanations provided by the field offices show that, despite the Task Force's specific instructions, some offices and agents still used their discretion to determine whether to send certain items to the OKBOMB Task Force. Charlotte's, Albany's, and Miami's explanations specifically acknowledged that personnel made decisions not to send items to the OKBOMB Task Force. In addition, the survey responses from other field offices indicated that there could have been other belated documents that were intentionally not sent to the OKBOMB Task Force because an agent did not believe it was sufficiently significant.
We acknowledge, however, that the issue is a complicated one, particularly with respect to the second wave documents because they were coming from places other than the Auxiliary Office OKBOMB file. Because the field offices were being extremely broad in their judgment of what constituted OKBOMB material, some of the belated documents include items that do not have an OKBOMB case number, and their relationship to the OKBOMB investigation is hard to discern. Accordingly, with respect to a few of the belated items, it is not entirely clear to us today whether the problem was that the items should have been sent to the Task Force and disclosed to the defense or whether they should not have been disclosed to the defense in 2001 and not counted as belated items.
Even if field offices sent a teletype but not an FD-302, put the original FD-302 in the file instead of sending it to the Task Force, or sent the original to another field office instead of the Task Force, most of these documents would have been sent to the Task Force anyway if the field offices had responded properly to multiple document requests from the Task Force. The evidence showed that many field offices did not follow the Task Force's instructions to search their files and to ensure that all leads had been documented properly and sent to the Task Force.
The written record showed that throughout 1996 the Task Force continually reminded the field offices to send material. In addition to the instructions we set out previously in this chapter, Section I D, a July 7, 1996, teletype stated:
Due to discovery concerns, it is necessary that all interviews be reported in FD-302 or insert format. In some instances, previous requests to place interview results into FD-302 or insert format have elicited no response from the field, thereby necessitating an additional communication restating the request.
On October 3, 1996, the OKBOMB Task Force told the field:
Steven [sic] Jones, counsel for OKBOMB subject, Timothy McVeigh has filed a large number of motions alleging government prosecutors are withholding evidence and FD-302/inserts which could be critical to the defense of his client. In furtherance of insuring all, repeat, all FD-302s and inserts have been turned over during discovery, as ordered by trial judge Richard Matsch, we have found 1As with agent investigative notes where no FD-302 was generated. [Inspector in Charge] OKBOMB would like to remind each division that it is imperative to the successful prosecution of this important case that all investigation be recorded as required in either an FD-302 or insert as applicable to prevent further defense allegations.
(Emphasis added.) And, as previously discussed, the Director's November 15, 1996, teletype not only sternly directed the SACs to ensure that all OKBOMB material had been sent to the Task Force, it also specified the numerous locations where field office personnel were to look.
Yet, many field office personnel that we interviewed could not recall receiving the teletypes from the Task Force, nor could they recall conducting any type of officewide search for material. In addition, 29 field offices responded to our survey question about what they did in response to the Director's November 15, 1996, EC by stating that they did not know what was done because they currently lacked access to their files.34
Because most field office personnel believed they had sent everything in to the Task Force, we believe it is likely that in many field offices no further search was conducted in response to the subsequent teletypes or at best only a superficial search. For example, Los Angeles acknowledged that it failed to send 18 1Bs (consisting of audio and video tapes) to the OKBOMB Task Force as required. Los Angeles, like the other field offices, had received a September 16, 1995, request from the OKBOMB Task Force for all 1B evidence and had responded by stating that they had no such evidence. The Los Angeles personnel to whom we spoke could not explain why Los Angeles responded that it had no evidence when, in fact, it did. They also could not explain why Los Angeles incorrectly responded to the Director's November 15, 1996, EC by stating that all documents and evidence had been submitted to the Denver Command Post.
The fact that over 300 original items were found in field office files is compelling proof that the field offices did not adequately respond to the Task Force's and the FBI Director's instructions to provide original items to the Task Force. The presence of so many originals in field offices' files indicates that the field offices failed to conduct searches despite the instructions, conducted inadequate searches, mistakenly assumed a copy of an original item had been sent to the Task Force previously, or mistakenly concluded that the original did not have to be sent.
In addition to missing items that might have been discovered if the field office had performed appropriate searches in 1995-1996, the field offices' failure to respond properly to Task Force requests had other consequences. For one thing, the Task Force had to expend valuable time and effort repeating requests that it had made previously and reminding field offices to follow instructions that had been sent previously. For example, on September 16, 1995, the Task Force requested that field offices send all evidence in their possession to the Task Force or respond to the Task Force that no evidence was being stored in the field office. On October 25, 1995, the Task Force sent a teletype to 11 field offices stating that no response had been received pursuant to the September 16 teletype and requesting a response. On November 11, 1995, the Task Force sent a teletype to three field offices, including Baltimore, stating that those field offices had not responded to the previous two teletypes. Neither the Baltimore OKBOMB case agent nor the squad supervisor recalled the teletypes, and they had no explanation as to why Baltimore repeatedly failed to respond to Task Force requests. Even the communication from the office of the FBI Director did not generate timely responses. A December 14, 1996, teletype to the field stated that the OKBOMB Task Force was still waiting for responses from some offices to the Director's November 15, 1996, directive requiring the heads of all offices to certify that all original FD-302s, evidence, 1As, logs, and other documents had been forwarded to the Task Force.35
As another example, eight months after the Task Force's request that non-pertinent information be sent as soon as possible, Dallas sent an EC dated May 7, 1996, stating that it was forwarding the non-pertinent results of investigation conducted by the Dallas Division. Attached to the EC were the original and 10 copies of 44 FD-302s and inserts. Apparently explaining the delay in responding, a Dallas agent wrote in an undated memorandum that although the OKBOMB Task Force had requested all non-pertinent information, "No specific dates or priority was provided," although the agent also acknowledged in the memorandum that "the priority was implied."36
The field offices' failures to follow instructions also increased the complexity of the Task Force's document management process. An August 1, 1995, teletype reminded field offices that it was essential to place "lead numbers" on all communications, FD-302s, and inserts.37 Yet, by teletype dated September 1, 1995, the Task Force had to repeat the instructions because some field offices were not complying:
Although previous teletypes have been sent to all field divisions advising lead control numbers need to appear on all communications, including FD-302's and inserts, Oklahoma City CP continues to receive a number of FD-302's without this information.
Because of the voluminous amount of reports generated, hundreds of man hours have been spent identifying lead numbers for these FD-302's to insure leads have been covered. There are still many FD-302's which can not be matched to any specific lead control number.
Due to this situation, all communications, including FD-302's and inserts, which do not contain lead control numbers … will immediately be returned to that field division to have the lead control number added.
In addition, the Task Force received FD-302s and inserts that had not been initialed by the agent, requiring the Task Force to send the documents back to the field office for initialing. Sending documents back and forth increased the chances that documents would be lost.38
During our investigation, we learned of other incidents where field offices had failed to send material timely to the OKBOMB Task Force. These incidents involved documents that were turned over to the defense in 1996 before McVeigh's trial but past the discovery deadline set by the court. We discuss these incidents because they show that the "belated documents" were not the only items that were delayed in getting to the Task Force.
In December 1996 the OKBOMB Task Force received three FD-302s after the discovery deadline of December 15, 1996, had passed. Two of the FD-302s were from the Anchorage Division and one was from the Seattle Division. Prosecutor Beth Wilkinson requested a written explanation from the FBI.
In a letter to Wilkinson dated January 8, 1997, Defenbaugh explained that as part of the "OKBOMB FBI's continued quality control efforts,"39 they had discovered on December 9, 1996, that Anchorage had interviewed two individuals on September 7, 1996, and reported the information by teletype the same day. Anchorage, however, had failed to draft FD-302s or inserts.40 The OKBOMB Task Force instructed Anchorage to draft FD-302s of the interviews and send them immediately. Defenbaugh wrote in his letter that although Anchorage complied, the FD-302s were not received by the Task Force until December 19, 1996. In explaining the Seattle problem, Defenbaugh wrote that on December 12, 1996, the OKBOMB Task Force discovered that Seattle had failed to draft an FD-302 for an interview it had conducted on May 25, 1995. According to Defenbaugh, "[a]lthough Seattle forwarded copies of the FD-302 almost immediately, the FD-302 did not arrive at the [command post] until [December 19, 1996]." The persons whom we interviewed could not recall why it took ten days in the Anchorage matter and seven days in the Seattle matter for the Task Force to receive documents that were needed urgently.
While we found significant evidence that field offices failed to send documents to the Task Force in a timely manner, it is also true that the OKBOMB Task Force bears some responsibility for the failure to get the materials to the defense before trial. The volume of material coming into the Task Force made it likely that some items would not be included in the discovery process, particularly given the FBI's computer systems used for handling these documents. But it does not appear that Task Force supervisors recognized the deficiencies in their own document processing. Consequently, they gave little consideration as to whether any measures could or should be taken to plug the gaps in the Task Force's document management system.
We concluded that as many as 48 percent of the belated FD-302s and inserts disclosed in the first wave could have been sent timely to the Task Force by the field offices. We reached this conclusion based upon: (1) instances where we located the belated document in a Task Force file other than the D or E subfiles; (2) airtels sent by the field offices stating that they were sending the belated document; and (3) a presumption that the original had been sent to the Task Force if only a copy was found in the field office files.
The strongest evidence supporting the field offices' claims that they had sent a particular belated document to the Task Force was to find the actual document somewhere in the Task Force files. The OIG found nine of the belated documents in OKBOMB files other than the D or E subfiles. In other words, the field office had sent the document to the Task Force, which in turn had filed it somewhere other than the D or E subfile, and it therefore had not been disclosed to the defense. Because FD-302s and inserts were compiled for discovery only from the D and E subfiles, if an FD-302 or insert was misfiled, the document was not disclosed.
In addition, as part of the second wave review in May 2001, the FBI found 39 items that had not been disclosed to the defense in the OKBOMB Main File, the warehouse storage facility, and the Oklahoma City confidential informant files. Although these were called "Oklahoma City Division belated documents," in reality the documents illustrated that field offices had sent FD-302s and inserts to the Task Force but that the documents had not made it to the defense.
Table 5 sets forth the documents and where they were located.
Table 5: Belated Documents Sent by Field Office to Task Force (TF) in 1995-1997
|Bates Stamp||Document Type||Originating Field Office||Location Where Discovered|
|376||Insert||Boston||TF Main File|
|537||1A||Buffalo||Contents of 1A were found attached to a 302 in the TF Sub D file|
|561||1A||Buffalo||Contents of 1A were found attached to an EC in the TF main file|
|785||302||Detroit||In ZyIndex unserialized|
|1314||302||Los Angeles||TF Main File|
|1704||302||Kansas City||In ZyIndex unserialized|
|2623||Insert||Sacramento||In ZyIndex unserialized|
|3788||Insert||Oklahoma City||Informant (137) Files|
|3976||Insert||Louisville||TF Main File, attached to serial #504|
|3978||Insert||Oklahoma City||TF Main File, Sub S4-2|
|3979||Insert||Phoenix||TF Main File, Sub S-43|
|3980||Insert||Dallas||TF Main File, serial 330|
|3982||302||Dallas||TF Main File, serial 334|
|3983||302||Charlotte||TF Main File, Sub O-492|
|3984||302||Oklahoma City||TF Main File, Sub F-942|
|3987||302||Oklahoma City||TF Main File, Sub S-1401|
|3991||302||Charlotte||TF Main File, Sub O-232|
|3992||302||Philadelphia||TF Main File, Sub O-119|
|3993||302||Philadelphia||TF Main File, Sub O-117|
|3994||302||Philadelphia||TF Main File, Sub O-118|
|3995||302||Philadelphia||TF Main File, serial 1686|
|3996||302||Atlanta||TF Main File, Sub O-642|
|3999||302||Philadelphia||TF Main File, Sub O-734|
|4013||Insert||Boston||TF Main File, Sub UU-250|
|4021||302||St. Louis||TF Main File, Sub L-324, second 302|
|4022||302||St. Louis||TF Main File, Sub L-321|
As previously discussed in Section III B 2 of this chapter, the OIG concluded that an airtel that described a belated document warranted a strong presumption that the field office had sent the document to the Task Force. While many field offices asserted that airtels would show that they had sent the belated documents to the Task Force, in most cases they did not provide the airtels. The field offices stated that because their files were in Oklahoma City and because they could not get access to ACS, they could not find the appropriate airtels.41
We did not just rely on the field offices, however. We conducted our own search for evidence that the documents had been sent to the OKBOMB Task Force as the field offices claimed. We searched extensively in ZyIndex for airtel records that would correspond to the belated documents. We found airtels for 58 documents. The majority of those airtels related to documents sent by Buffalo and Detroit.
The OIG concluded that when the belated document consisted of only a copy that had been found in the field office files rather than an original, it was reasonable to presume that the original had been sent to the Task Force. This presumption is warranted, we believe, even though the original document was not found in the Task Force files during any of the searches conducted in 2001 and there was no reference to the document in ZyIndex or other databases.
A number of factors lead us to believe that it is more likely that the original was sent by the field office to the Task Force and misplaced after it arrived rather than that the original was misfiled or lost in the field office's OKBOMB file and never sent. These factors include: the volume of documents the OKBOMB Task Force was attempting to process relative to those in the field offices; the fact that document processing at the Task Force was different and more complex than the standard protocols the field office staff followed; and the fact that a substantial portion of the document processing in Oklahoma City was being handled by personnel from other field offices who were stationed at the Task Force on temporary duty. We examine these factors in the discussion in the following section.
In summary, we found conclusive evidence that 48 belated documents had been sent timely by the field offices because we found the documents in the Task Force files. Airtels provide strong evidence that 54 belated documents were sent timely by the field offices to the Task Force. We also concluded that 218 FD-302s and inserts that were copies (and that had no corresponding airtels) were more likely than not timely sent by the field offices to the Task Force.42
Although OKBOMB Task Force personnel told the OIG they were frustrated from trying to get material from the field offices, there were some indicators that the Task Force was having its own document management problems that were frustrating field office personnel. For example, many field office personnel told us that they received repeated telephone calls from the Task Force asking for documents that the field office personnel said they knew had been sent to the Task Force. They said that the Task Force would ask them for the same document two or three times.
In addition to telephone calls, field offices sometimes noted in writing that it had previously sent information to the Task Force. For example, a Detroit EC transmitting 10 copies of an insert stated, "Detroit does not have the original insert reflecting this interview inasmuch as it was previously furnished to [OKBOMB Task Force]." A Dallas EC stated:
[Witness was] interviewed on 04/25/1995. That interview was reported on Dallas teletype dated 05/03/1995. Later when requested by OCCP [Oklahoma City Command Post] under LCN 13725II-05 (not 13725II-05/06, as stated in OCCP-Denver's 07/05/1996 teletype), Dallas advised OCCP that this lead had already been covered; however, to assist OCCP, Dallas forwarded 10 copies of the Insert and the diskette version via Dallas EC dated 02/06/1996. Dallas also sent a teletype dated 07/25/1996 advising OCCP of the coverage of this lead. The instant EC is the fourth communication to OCCP to advise that this lead has been covered, and is the second to send photocopies of the original Insert.
Task Force personnel acknowledged that the field complained about having to send documents repeatedly. They said that rather than arguing about whether a document had been sent, they would simply ask the field to resend it.
Although it did not relate to a belated document, we also found reference in writing indicating that the Task Force had "lost" a document for some period of time. The document was ultimately found and disclosed but after the discovery deadline had passed. Defenbaugh's January 8, 1997, letter to prosecutor Wilkinson stated that the Phoenix Division had included an FD-302 in a box of materials that it sent to the Task Force in November 1996. As Defenbaugh wrote:
Normally, the inserts and FD-302s would be forwarded to the Discovery team for copying, and the Rotor to be serialized. In this particular instance, it is unknown what happened to [this] FD-302. On 1/5/1997, the FD-302 … appeared in the "in" box at the Rotor's desk. The Rotor had been previously cleared, so there is no possibility the report had been sitting in the Rotor area. The … FD-302 did arrive at OKBOMB [Command Post] prior to 12/1/1996, was misplaced at the [Command Post] and failed to make the 12/15/1996, deadline.
Much like the field offices, OKBOMB personnel could not identify the reason why a particular document was not disclosed to the defense. One explanation is that massive quantities of documents were coming into the Task Force and some of the documents were misplaced. The OIG found that at least 20 documents were serialized (meaning a Task Force rotor had processed the document) but filed in locations other than the D or E subfiles from which discovery was made.
We discuss in the following sections issues that we believe were contributing factors.
The process used to move paper around the Task Force was cumbersome and the opportunities to have a document go awry were numerous. Although the filing system used by the Task Force was, in a broad sense, the same as that used in other FBI cases, the volume of material required extra steps and extra personnel. For example, instead of a single rotor completing the entire process, the process was compartmentalized into small steps that were handled by several personnel, most of whom were on temporary duty. In addition, because each person handled only a very small part of the process, it was more difficult to determine if a mistake had been made during one of the steps.
As previously discussed, all documents received at the Task Force had to be serialized into FOIMS and later ACS. First, field offices were required to send an original, multiple copies, and a disk for each of the hundreds of thousands of documents generated in OKBOMB. These documents were sent to the Oklahoma City Field Office and were transported daily to the Task Force. According to the lead OKBOMB rotor, she reviewed all incoming mail and then sorted it. First, the disks were separated from the documents and routed to the ZyIndex team. Next, she routed copies to agents who determined to which subfiles the documents would be serialized. Once the rotor received a document back from the agents with filing instructions, the rotor sorted the documents into particular filing cabinets in order for other rotors on temporary duty to serialize and file.
The original cover communication (usually an airtel) was separated from the original FD-302 or insert and copies and was placed in a filing cabinet to be serialized to the OKBOMB Main File. The original FD-302 was placed in another filing cabinet for serializing to the FD-302 subfile. Copies of the FD-302 also might be placed in other file cabinet drawers, such as the "McVeigh" subfile drawer, for serialization. The serialization was handled by several rotors on temporary duty who were assigned a particular subfile to handle. According to the lead rotor, for many weeks they had three shifts a day running the operation.
This filing system could break down in many ways. First, the Task Force did not have a routine policy of checking to ensure that items the field office said were being sent had actually arrived at the Task Force. Also, as the description above shows, documents easily could be accidentally lost or placed in the wrong filing cabinet drawer, and the error would not be noticed. Because discovery was created from particular subfiles, the system depended on the rotors correctly serializing every FD-302 and insert. A mistake meant that the document would be placed in the wrong file and therefore not be disclosed.
Field agents sending documents directly to OKBOMB Task Force agents also may have contributed to the problem. We were told that field agents sometimes faxed documents to a particular agent at the Task Force. It is possible that the Task Force agent believed the field would send the original through regular channels while the field considered the document as having been sent.
During the course of the investigation, the Task Force used hundreds of FBI personnel on temporary duty (TDY). These personnel came from the field offices for limited tours, in some cases as little as two weeks. While OKBOMB personnel spoke highly of the efforts of many TDY personnel and appreciated the sacrifices of the field offices and the individuals involved, we also were told that some of the TDY personnel were incapable of performing the work and some failed to perform their tasks competently. Indeed, personnel from one of the field offices admitted to the OIG that they had removed a support employee from working the OKBOMB file in their field office because the employee had performed poorly. Yet, they then sent that employee to work TDY at the OKBOMB Task Force.
We note that even for those individuals who were trying to do a good job, problems developed because field offices vary in their document management procedures. Accordingly, the TDY personnel were receiving on-the-job training in the middle of a crucial case. The OKBOMB personnel also told us that the rapid turnover created problems for the supervisory personnel because they constantly had to train new personnel. In addition, most of the administrative personnel were unfamiliar with ZyIndex.
In the immediate aftermath of the disclosure of the belated documents problem, the media reported that FBI officials blamed the FBI's computer systems for causing the problem. Some officials specifically cited the transition from FOIMS to ACS as the cause of the difficulty.
The evidence shows that the FBI's computer system did not cause the Task Force to fail to turn over discoverable items to the defense, but the inadequate computer systems failed to assist the Task Force in realizing that they had a problem. Because of the limitations in FOIMS, the Task Force began the investigation without adequate quality control measures in place. They relied to some extent on manual quality control measures but these too had their limitations.
As discussed previously, the computer system in place at the time of the OKBOMB investigation was FOIMS. Documents were serialized individually by field offices and then a second time when they reached the Task Force. The two serial numbers were not cross-referenced, however. Consequently, the Task Force did not know what documents had been generated in the field until they received a hard copy of the document. As a result, the Task Force had no easy method of determining whether it was missing documents created by the field.
Indeed, items could be lost in the mail, and no one would know it. The Task Force would be unaware that something was on its way and therefore would not realize when it did not arrive. The field offices also had no way of knowing that the OKBOMB Task Force did not receive items because they had no system for acknowledging receipt of an item.
Quality control measures could have been devised to compensate for the Task Force's inability to easily determine what documents the field offices had generated. One method would have been to cross-reference the OKBOMB serial numbers and the local serial numbers generated in each field office by FOIMS. By tracking the local serial numbers, the Task Force could have determined if they were missing documents because the local serial numbers would be sequentially incomplete. This method would have had to have been instituted at the beginning of the case, however. Another method that could have been initiated later would have been to determine all the local serial numbers in FOIMS and then use ZyIndex to determine whether those documents were in the OKBOMB files. Either of these methods would have made it possible, albeit in a time consuming manner, to trace whether the OKBOMB Task Force was in possession of the field offices' documents.43
Neither the OKBOMB Task Force nor FBI Headquarters instituted such a quality control system. Task Force personnel told the OIG that they did not believe that they had a serious problem getting documents from the field as opposed to isolated instances of documents not arriving at the Task Force. Although the prosecutors acknowledged being aware of some issues, such as the Phoenix surveillance logs that were not disclosed, they also believed the problems were isolated and did not reflect broader failings.
On October 16, 1995, the FBI initiated a major change to its automation system when the entire agency moved from FOIMS to ACS. ACS was designed as a comprehensive and integrated system that would automate case management and allow for complete access to investigative information throughout the FBI. At the time of the transition, the data on every item listed in FOIMS was automatically transferred into ACS. Although there were many differences between the two systems, the two most notable for OKBOMB were the addition of full-text retrieval and universal serialization.
The full-text retrieval capacity was an important upgrade because it allowed greater and more convenient access to documents and information. In ACS, any FBI-generated document, such as FD-302s, inserts, and ECs, could be uploaded into ACS, and employees throughout the FBI could review the document using the computer system. Despite this capacity, the OKBOMB Task Force continued to use its stand-alone program, ZyIndex, which also had full-text retrieval capabilities.44
The addition of universal serialization also was an important upgrade. When a document is entered into ACS, it is assigned a universal serial number. In other words, in any given case, there is only one serial number 1 rather than a serial number 1 in every field office that participates in a particular case. If, for example, Buffalo uploads a document on a case into ACS, the system automatically gives the document a serial number, for example, serial 100; if Atlanta serializes the next document, the document is serial 101. If such a system had been in use in 1995-1997, the OKBOMB Task Force would easily have been able to recognize if it was missing serial 100 in its hard copy files because there would have been a skip in the numerical sequence.45
Universal serialization would have substantially improved the Task Force's ability to track all documents, particularly documents generated by the field offices. The Task Force, however, did not effectively utilize this feature because the Task Force prohibited the field offices from serializing or uploading documents into ACS. Instead, the Task Force primarily relied on the system it had used before ACS, which was to reserialize the documents once they arrived at the Task Force.
Some field agents recalled confusion during the OKBOMB investigation as to how the field was supposed to use ACS for processing OKBOMB documents. After examining numerous teletypes concerning this issue, we agree that there was cause for substantial confusion in the field about using ACS for OKBOMB documents. The following series of communications shows that the OKBOMB Task Force was reacting to and trying to fix various problems as they arose, but when they did other problems often developed.
When ACS initially came on line, the Task Force prohibited the field from uploading documents into ACS and serializing them. A teletype dated October 13, 1995, informed the field offices that:
Only the … OKBOMB Task Force… will serialize and index information into the ACS…. All divisions should continue to maintain their own paper file…. [C]ontrary to other pending investigative files, … OKBOMB will not be uploading any case documents to the electronic case file in ACS. There will be no full text retrieval for OKBOMB….
Defenbaugh told the OIG that he made the decision to prohibit the field from uploading documents into ACS. He said that he had learned from other major cases, such as the UNIBOMB case, that the Task Force had to maintain "command and control" over the information. Defenbaugh said that his decision upset a lot of field offices. In addition, because documents were being leaked to the press, Defenbaugh restricted the field's access to OKBOMB documents through ACS's full-text retrieval capacity.
However, one month later on November 18, 1995, the Task Force changed the instructions and permitted serialization and uploading.
Effective immediately all offices may start serializing in the OKBOMB file number 174-OC-56120. Uploading is not mandatory, but if you wish to do so you may.
On December 28, 1995, the Task Force changed its instructions and told the field to serialize but not to upload:
Due to the large volume of documents, Oklahoma City CP does not have the capacity at present to download documents and requests documents not be uploaded into ACS. It is also imperative documents be serialized into the correct subfiles, as set forth….
By teletype dated February 23, 1996, the instructions were changed again. The field offices were now required to serialize and upload:
Due to the discovery process, effective immediately, all field offices are requested to upload all FD-302's and inserts into the ACS system. All FD-302's are to be serialized into 174A-OC-56120-D. All inserts are to be serialized into 174A-OC-56120-E.
And on November 16, 1996, the instructions changed one last time:
It is as previously requested imperative that [Auxiliary Office] does not upload/serialize documentation into ACS. Serializing and uploading will be conducted by [the OKBOMB Task Force].46
The OIG was told by the lead OKBOMB rotor that the field offices were prohibited from serializing documents in 1996 because they often uploaded documents incorrectly (for example, into the wrong subfile), which caused substantial problems during the discovery process. None of the participants to whom we spoke could recall why Defenbaugh's initial decision to prohibit the field offices from serializing and uploading documents into ACS was changed in December 1995 to allow it.
While prohibiting the field offices from serializing solved one discovery problem (the problems created by uploading into the wrong subfile), it impeded the resolution of another. The lack of universal serialization meant that the OKBOMB Task Force continued to be without the ability to comprehensively track documents generated by the field offices. We did not find evidence that this consequence was discussed or considered when the decision was made to prohibit use of the universal serialization feature. We were told that when the decision was made, the Task Force was not aware that the field offices were not sending all of their documents to the Task Force as they were instructed to do. Although the Task Force could be criticized for failing to take advantage of such an important quality control measure, because the vast majority of the belated documents came from the FOIMS period, ACS's universal serialization feature would not have significantly improved the system for most of the belated documents.
We also note that the transfer to a new computer system in the middle of this mammoth investigation created other technical difficulties, although these difficulties do not seem to have affected or caused the belated documents problem. For example, ACS had trouble accepting OKBOMB's subfiling system, and it also created duplicate 1A numbers during the transition from FOIMS.47 FBI Headquarters had to develop specific solutions for these problems.
The primary automated system used by the OKBOMB Task Force to manage the investigation was the Rapid Start Information Management System (Rapid Start). Rapid Start is used to track leads and lead responses. Rapid Start was developed in 1992 to provide information management services in crisis situations, special events, or major cases. Immediately after the bombing a Rapid Start team was sent from FBI Headquarters, and the system was in place and running by April 20, 1995. In addition to a core group in FBI Headquarters, the FBI also trains numerous personnel throughout the field who can be deployed in a crisis to organize Rapid Start.
Leads, whether they are reported by telephone, complaint, or investigation, are recorded on a "lead sheet." Certain information about each lead is recorded into Rapid Start, such as the office to whom the lead is assigned, the date assigned, and the task assigned. A "lead number" is assigned automatically by Rapid Start to assist in tracking the lead, and the computer system is updated when responses arrive. The system can be used to generate reports, such as the number of outstanding leads for a particular field office. The team known colloquially as "Lead World" was responsible for setting the leads, tracking them, and reviewing and organizing the responses.
Rapid Start is not a document management system, however. It was designed to ensure that the persons making the decisions about the direction of the investigation had the necessary information to make those decisions. It was not designed to track whether documents were created in field offices, whether documents had been sent to the Task Force, or whether documents had been generated correctly. Rapid Start was not linked to FOIMS, ACS, or ZyIndex; therefore, a document could be listed in Rapid Start but not in the other databases and therefore not make it to the subfiles used to compile discovery.48
Even for tracking leads, Rapid Start had its limitations. In addition to being used by the OKBOMB Task Force, Rapid Start was used in many field offices that were generating significant investigative leads, particularly other FBI command posts. These Rapid Start systems in the field offices, however, were not linked to the Rapid Start system used by the Task Force. This meant that if a field office, such as Detroit, generated its own leads, the leads would not be tracked automatically by the Task Force. Accordingly, the Task Force would not know what locally generated leads the field offices had failed to pursue or if they had failed to submit documentation after completing the lead. Although the Task Force instructed the field offices to obtain a Task Force lead number for all leads, including locally generated leads, it had no way to ensure that the field offices complied.
In addition, although Rapid Start was designed to help track outstanding leads, the Task Force relied heavily on a manual lead tracking system that utilized binders.49 When the lead was set, the lead control sheet was filed in a binder for outstanding leads. When the response was received, it was attached to the lead control sheet and the package was filed in a different set of binders.
Task Force personnel attempted to use Rapid Start in some ways as a document management tool by initiating several quality control projects to assist them in obtaining documents they believed were missing. One such project was designed to ensure that all agent interview notes that the Task Force received in 1A envelopes were matched to a corresponding FD-302. Another was Project 13725. The OKBOMB Task Force initiated the project to ensure that the field converted the information in teletypes into FD-302s and inserts (13725 was the lead number under which the Task Force's instructions were sent). OKBOMB Task Force personnel conducted a labor intensive review to manually match teletypes with FD-302s and inserts. If no FD-302 or insert was found, the OKBOMB Task Force sent a lead to the field office instructing the field office to send the FD-302 or insert. (See Chapter Three, Section I D.)
These projects were prone to human error, however. Some of these errors did have an impact on the belated documents issue because some of the belated documents might have been located before the defendants' trials if the quality control projects had been more successful. For example, the Task Force sent over a thousand leads to field offices specifically identifying teletypes that needed to be turned into FD-302s or inserts. Yet, we found that approximately half of Atlanta's belated documents were FD-302s that contained information that had been sent previously by teletype but we did not find any specific request by the Task Force for those FD-302s or inserts - which was the point of project 13725. Although the field offices had the primary responsibility for ensuring that information was reported in the proper format, the error could have been caught with a more effective quality control project.
Another automated system component used by the OKBOMB Task Force was ZyIndex. FBI Headquarters recognized the need to be able to electronically search and retrieve the text of documents, which was not possible under FOIMS. ZyIndex was used instead. ZyIndex is not, however, a document management tool. It is an off-the-shelf software application that indexes words and phrases to allow for electronic retrieval of documents. For example, ZyIndex allows an investigator to quickly retrieve all documents that mention a particular person's name. It did not track what was generated in the field offices, whether documents had been sent to or received by the Task Force, or how many documents were being generated. In addition, ZyIndex had no impact on the discovery process. Discovery production of documents was made based on the hard copies of documents that were in the FD-302 and inserts subfiles.
Early in the process, documents were sent from the field offices to the OKBOMB Task Force and were retyped into ZyIndex. By May 1, 1995, the field offices were instructed to load all FBI generated documents, such as FD-302s and inserts, onto floppy disks and send the disks along with the hard copies to the Task Force. The documents contained on the disks were then uploaded into ZyIndex along with the OKBOMB serial number and descriptive information.
However, because of the enormous volume of material that was being sent to the Task Force, the OKBOMB rotors who were responsible for serializing the documents quickly became inundated and fell behind in the serialization process. In order not to delay the entire process, the OKBOMB administrative supervisors altered the process so that the disks were sent to the ZyIndex team even though the documents had not been serialized. Consequently, documents were uploaded into ZyIndex without serial numbers.
Although this kept documents flowing into ZyIndex, it also increased the complexity of the process because the serial numbers had to be entered into the computer separately. A quality control measure was instituted to attempt to solve the problem. After the serialized hard copy documents were filed into volumes, personnel took one volume at a time and tried to match each hard copy document with a document in ZyIndex. If the document was found in ZyIndex, the serial number was entered into the database. If the document was not found, it was retyped into ZyIndex. After each document in a volume of FD-302s and inserts was matched to an entry in ZyIndex, a notation was written on the outside of the volume - "Quality Control Rapid Start" - along with the date of the review.
The ZyIndex team implemented other quality control measures. Since the ZyIndex team received only disks and not the hard copies, the ZyIndex team could upload documents into ZyIndex that the Task Force had never received in hard copy. After a few weeks of uploading into ZyIndex, the ZyIndex supervisor realized that there continued to be documents in ZyIndex that had no serial numbers. Two possible explanations were that the Task Force either never received the hard copy or received and lost the hard copy. The ZyIndex team identified electronic files for which there was no serial number and provided those documents to the lead rotor. Rather than try to go through thousands of hard copies of documents to see if the document had in fact been serialized (because FOIMS was cumbersome and had no full text retrieval), field offices were contacted and agents were asked to send the document to the Task Force even if it had already been sent. This resulted in complaints from the field offices that the documents had already been sent and more work for the Task Force to serialize the documents again. It did, however, ensure that some documents that had been lost for whatever reason were received and serialized.50
A diagram of this process is shown on the following page.
[ Diagram describing the process of Zyindexing of OKBOMB Documents (FD-302s and Inserts)is not available electronically.]
We questioned FBI employees and former employees, analyzed circumstantial evidence, and investigated evidence the defense alleged showed that the government intentionally withheld exculpatory evidence. We concluded that the failure to provide timely all OKBOMB documents did not result from an intent to conceal exculpatory material.
We examined both direct and circumstantial evidence in reaching our conclusion. First, we received no direct evidence that any FBI or Task Force personnel intended to conceal exculpatory information. Second, the evidence showed that, for the most part, the failure to provide documents and other items to the defense was caused by mistakes on the part of various individuals. As previously discussed, we did find a few instances where agents acknowledged that they decided not to send material to the OKBOMB Task Force. These items included a letter from an inmate referring to the bombing and a political statement on a postcard sent to the FBI following the bombing. The agents who made the decisions not to send the material to the Task Force stated that they did so because they believed the items were completely irrelevant to the OKBOMB investigation. We believe that, at worst, those incidents reflected an erroneous view of the discovery agreement and the agents' failure to follow instructions from the Task Force. We do not believe the incidents showed any intentional decision to withhold exculpatory information from the defense.
Third, the belated documents did not contain a significant quantity or quality of previously unknown exculpatory information. We found that a significant portion of the belated documents concern utterly useless information and would not have been discoverable in most cases. For example:
Although the OIG did not analyze each document to make a judgment whether it met the definition of exculpatory material, we did rely heavily on the fact that in his Petition for a Stay of Execution, McVeigh only pointed to eight belated documents (Defense Exhibits 9-14, 16-1751 and 19) as being particularly significant. In addition, in many instances the information contained in the belated document had been disclosed in a different format.
Fourth, the fact that the government disclosed information pretrial regarding allegations that persons other than McVeigh and Nichols had bombed the Murrah Building is evidence that the government was willing to disclose potentially exculpatory information. Fifth, most of the belated documents were still maintained in the field offices' Auxiliary Office File where they could easily be found if anyone looked for them.
Because the defense identified eight of the belated documents as particularly significant and a strong indicator of the government's intentional misconduct, we examined those documents specifically to determine why they had not been produced. We interviewed the agents who drafted the documents, their supervisors, and in some cases the rotors who processed the documents. In one instance, we found that an agent had determined that the item was not pertinent to the investigation and failed to send it to the OKBOMB Task Force. In the other instances, the individuals responsible for the document believed they had acted appropriately and could not explain why the document had not been disclosed to the OKBOMB defendants.
In the discussion that follows, we identify the documents by the exhibit number given to the document in McVeigh's Petition for Stay of Execution.
Defense Exhibit 9 is a 2-page insert that documented an interview conducted on March 13, 1997, by the Salt Lake City Field Office. In the insert, a confidential informant provided the agent with secondhand information about McVeigh and other individuals that had been provided to the informant by a militia member. Defense counsel argued that the information in this insert would have provided a "critical lead" to exculpatory evidence. The insert was produced by Salt Lake City in May 2001 in the "second wave." The insert was not located in Salt Lake City's Auxiliary Office file and was found only after the office conducted a hand search through their confidential informant files.
We initially obtained information from the Salt Lake City Chief Division Counsel (CDC) and the File Supervisor. The CDC explained that typically when an agent prepares an insert (or FD-302) involving a confidential informant, the original of the insert is filed in the informant's file located in the confidential files section of the office. A copy of the insert is placed in the office's investigative file for that case.52 For OKBOMB, an additional copy should have been sent to the Task Force. According to the CDC and the File Supervisor, based on the file numbers on the insert, the agent who drafted the insert did not make a copy of it for the Salt Lake City OKBOMB file. Given that a copy was not made for the Auxiliary Office file, the CDC concluded that a copy of the insert also was never sent to Oklahoma City.
The agent who prepared the insert was assigned to the Coeur D'Alene, Idaho, Resident Agency and was the case agent for an investigation involving the Aryan Nation, a white supremacist organization. The insert has only the Aryan Nation file number although the FD-209, a form used with confidential source information, identifies the file numbers for the OKBOMB and Aryan Nation investigations.
When we interviewed the agent who prepared the insert, he stated that he did not recall the insert but did recall the informant. The agent said that one copy of the insert should have been sent to the Aryan Nation investigative file and one copy to the OKBOMB Auxiliary Office file. He explained the paperflow as follows: He would prepare the insert, a typist would provide two copies to him for his initials, he would then give the inserts to a supervisor, who would pass them to a rotor, and the rotor was responsible for placing the inserts in the proper files. The agent speculated that the problem may have occurred because the Coeur D'Alene Resident Agency did not have a rotor and the documents were sent to Salt Lake City for processing.
The agent told the OIG that although he did not believe the information provided by the informant was credible, he would have sent the insert on to the OKBOMB Task Force because it was the Task Force's responsibility to determine the relevancy of the information.
Defense Exhibit 10 is a 2-paragraph insert dated April 24, 1995, from the Colorado Springs Resident Agency relaying information from a confidential informant that a prison inmate had information that someone other than McVeigh and Nichols was involved in the bombing. The former agent who drafted the insert told the OIG that he believed he would have followed normal procedure, which was to send the original to the confidential source file in Denver, the Headquarters City for the Denver Division, and a copy to the OKBOMB Auxiliary Office file, also located in Denver. He could not recall who in the office was responsible for sending information to the OKBOMB Task Force. He stated that he knew that "everything" was supposed to be sent to the OKBOMB Task Force, and he had no explanation for why this particular insert had not been sent.
Defense Exhibit 11 is a 1-page insert provided by the Kansas City Field Office. In the insert, a confidential informant provided information on October 10, 1997, about individuals other than McVeigh and Nichols who the informant described "as persons behind the scenes of the Oklahoma City bombing." The defense stated that it never received this information during the trial.53 Kansas City located this document as part of the second wave after receiving a telephone call from FBI Headquarters on May 16, 2001, directing Kansas City to provide all informant serials bearing the Oklahoma City case number. The FD-209 contains the OKBOMB case number and a 266G-0 number, which is Kansas City's domestic terrorism miscellaneous file.54
According to the Kansas City Assistant SAC (ASAC), the original of the insert should have been maintained in the informant's file and a copy sent to the OKBOMB Task Force. In addition, a copy of the insert should have been maintained in the Kansas City OKBOMB Auxiliary Office file and the Kansas City 266G-0 file. Because a copy could not be located in either the Auxiliary Office file or the 266G-0 file, the ASAC concluded that copies were never created.
The special agent who prepared the insert explained to the OIG that his normal practice would have been to prepare the insert, put the original in his confidential informant's file, and send a copy directly to the OKBOMB Task Force along with a transmittal EC or airtel. When asked if he recalled sending this particular insert to the OKBOMB Task Force, the agent said that "part of me thinks I did" send the material. The agent said he only prepared one insert for the OKBOMB case. Although he said he typically would have sent the insert directly to the OKBOMB Task Force, the agent speculated that he could have sent the transmittal EC to the employee who maintained confidential informant files in Kansas City and asked her to send a copy of the insert along with the transmittal EC. The agent said it was possible that the administrative employee failed to send the insert to the Task Force.
The confidential files clerk told us that she did not recall this particular insert. She said that her normal practice was to take copies of the document to the rotor for the squads handling the cases listed on the FD-209. She said that she was not responsible for filing the documents or sending them to the Office of Origin.
This 2-page FD-302 from the Philadelphia, Pennsylvania, Field Office memorializes an interview conducted on June 14, 1996. The individual interviewed was associated with militia members but stated that he had no information about the bombing. The document could be a copy because it bears no agent initials. The agent who drafted the FD-302 said that he recalled the interview, and he recalled initialing the FD-302 because it had been returned to him for review and signing in an unusually expeditious manner. However, he did not specifically recall what happened after he signed the document. He could only speculate that he followed standard procedure, which was to return the signed original, copies, and a transmittal letter to his supervisor.
Defense Exhibit 13 is a 1A envelope that contained a letter dated April 18, 1995, from an inmate at the Federal Correctional Institution in Raybrook, New York. The letter was provided to the FBI on April 21, 1995. The 1A envelope and its contents appear to be originals. In the letter, the inmate stated: "No good news to report from here, except the federal building bombing in Oklahoma City. My people do good work, and it will continue until the United States government is destroyed…."55
The case agent who worked on OKBOMB for the Albany Field Office told the OIG that prison officials brought the letter to the attention of the FBI because the letter mentioned the Oklahoma City bombing. The case agent said that the inmate had psychological problems and that the letter had a total "disconnect" to the OKBOMB investigation. The agent stated that just because a "lunatic" made a comment that "his people do good work" and referred to the Oklahoma City bombing did not mean that the information was connected to the investigation of the bombing. He stated that the letter should not have been in the Albany Auxiliary Office OKBOMB file and, instead, should have been filed in the "zero file." The agent stated that he made the decision not to send it to Oklahoma City. He said that the FBI hires professionals who are supposed to be able to think and make judgment calls.
Defense Exhibit 14 consists of a 1-page FD-302 concerning an April 26, 1995, interview conducted by an agent from the Springfield, Massachusetts, Resident Agency. The document contains no markings other than the Boston serial number. We found another copy of the same document in the Auxiliary Office files sent in by Boston to Oklahoma City in 2001. That document had an original signature but also red index markings, which are usually found only on copies. The Boston Division OKBOMB case agent told the OIG that the Resident Agencies were responsible for sending OKBOMB material directly to the OKBOMB Task Force. He believed that both of the documents were copies even though one contained an original signature.
The agent who drafted the FD-302, however, told the OIG that for OKBOMB, the Resident Agencies in the Boston Division followed a different procedure than normal. He stated that the Resident Agency agents sent all OKBOMB material to Boston, which was responsible for sending the material on to the OKBOMB Task Force. He also stated that it was his practice to only sign the original. The agent theorized that the Boston rotor mistakenly placed index markings on the original and placed both the original and the copy in the Boston Auxiliary Office file without sending anything on to the OKBOMB Task Force. Despite numerous attempts, we were unable to contact the former Boston rotor.
Defense Exhibits 16 and 17 are duplicates - a 5-page FD-302 pertaining to an interview conducted on May 5, 1995, in Lafayette, Indiana. Exhibit 16, an uninitialed copy, was found in the materials sent by Indianapolis to Oklahoma City in 2001. Exhibit 17, which bears the agent's original inked initials, was found in the material sent from Detroit to Oklahoma City in 2001. In claiming that the government must have acted intentionally, defense counsel asserted that it was unlikely that the same document could be withheld by two different field offices without it being an intentional act.
The lead for this interview was likely sent by the Detroit Division rather than the OKBOMB Task Force, and it was directed to the Indianapolis Division. Our investigation suggests that the Indianapolis Division sent the original FD-302 to Detroit without sending a copy to the Task Force and that Detroit mistakenly filed the original in its Auxiliary Office files, also failing to send either the original or a copy of the FD-302 to the Task Force. Both offices apparently mistakenly assumed the other would send the document to the Task Force but took no action to verify their assumptions.
The Indianapolis agent who conducted the interview did not recall anything specific about the processing of the FD-302. Based upon his understanding that the FD-302 found in Detroit's files is probably the original, the agent theorized that he probably sent the original to Detroit because the lead was set by Detroit.
The Detroit OKBOMB case agent stated that, assuming Detroit set the lead, it would have been standard procedure for the Indianapolis Division to send the original FD-302 to Detroit, as the requesting office. He stated that unless Detroit had specifically requested that Indianapolis send either the original or a copy to the OKBOMB Task Force, the interviewing agent in the Indianapolis Division may not have done so on his own initiative. The Detroit case agent acknowledged that Detroit should have serialized a copy of the FD-302 into its files and sent the original on to the OKBOMB Task Force. He theorized, however, that it was possible that the Detroit rotor who received the FD-302 from Indianapolis assumed that it was an initialed copy intended for Detroit's Auxiliary Office files and that Indianapolis had sent the original to the Task Force. Thus, the rotor may have mistakenly serialized and filed the original into Detroit's Auxiliary Office files.56
Defense Exhibit 19 is a 6-page FD-302 that memorialized an interview conducted on May 11, 1995, in Los Angeles, California. The interview was of the girlfriend of an individual who the FBI at one time thought might be John Doe #2. The FD-302 appears to be an original and was found in the Los Angeles files sent to Oklahoma City in 2001.
A Los Angeles Division Associate SAC, an ASAC, and the Los Angeles OKBOMB case agent were all present for the interview. All distinctly remembered the interview but could offer little in the way of explanation about the disposition of the FD-302. The Associate SAC who prepared the document said that she would have forwarded it to the supervisor of the squad handling the OKBOMB matter. The squad supervisor told the OIG that he recalled an interview conducted by the Associate SAC and the ASAC, but he did not recall anything "out of the ordinary" about the FD-302. He said that he would have processed it like any other FD-302 with the original being sent to the OKBOMB Task Force with the requisite copies. He had no explanation for why it was not sent and remained in the Los Angeles Auxiliary Office files.
As part of their allegation that the government had intentionally failed to produce exculpatory information, the defense also pointed to information obtained from former FBI Special Agent Ricardo Ojeda, who during the OKBOMB investigation had been an agent in an Oklahoma City Division Resident Agency. In a May 2001 interview on the television show 60 Minutes II, Ojeda stated that he had drafted FD-302s that contained exculpatory information relating to the OKBOMB defendants and that he did not believe the government had disclosed these FD-302s.
The defense cited one of Ojeda's FD-302s in particular. This document was not one of the belated documents and had never been disclosed to the defense, even in 2001. The FD-302 concerned an April 14, 1997, interview Ojeda had conducted of a reporter. The reporter had previously interviewed an individual purporting to have knowledge of the activities of various militia members. The individual implied to the reporter that McVeigh and Nichols were either not involved in the bombing or at least had received substantial assistance from others. The defense asserted that this document was clearly exculpatory and had never been provided to them.
In response, the government claimed the FD-302 had not been disclosed because it had been written for a different case and that the case number on the FD-302 was not the OKBOMB case number. The government also asserted that the substance of the FD-302 had been disclosed to the defense in other documents and that the reporter who provided the information had worked closely with the defense attorneys.
Our investigation disclosed that the FD-302 was written in response to a lead generated in a different case, had a different case number, and therefore, in the normal course of document handling, would not have been placed in the OKBOMB files. Ojeda's FD-302, however, had made it to the OKBOMB Task Force in Denver because an agent recognized that it contained information relating to OKBOMB. But, because the document was not serialized and filed properly when it arrived in Denver, it was not disclosed to the defense.
Ojeda told the OIG that the lead to interview the reporter was self-generated as a result of information provided by an informant. However, the file records show that a lead dated April 14, 1997, from Tulsa Special Agent Peter Rickel directed Ojeda to interview the reporter.
The case number on the lead was 266A-OC-57917, a case involving bomb threats, possession of explosive components, and alleged members of militia groups. (The OKBOMB case number is 174A-OC-56120.) The file number on Ojeda's FD-302 is also 266A-OC-57917. Ojeda told the OIG that he did not know why the OKBOMB file number was not on the document. He said that although he had no specific recollection, he felt certain that he would have instructed the typist to add the OKBOMB file number on the FD-302.
We interviewed the individual who typed the FD-302. She had a vague recollection of the case but had no specific recollection of the FD-302. She stated that Ojeda would normally provide either verbal or written instructions regarding the case number or numbers that should be placed on the FD-302. Pursuant to her normal practice, if two numbers were provided, she would place them side by side on the number line of the FD-302. She said that she did not recall any supervisor requesting her to change or delete a file number on an FD-302. The typist added that if a supervisor wanted a number changed, the supervisor would normally tell the case agent who prepared the FD-302.
Tim Arney, the supervisor of the Muskogee Resident Agency and Ojeda's supervisor, told the OIG that he "generally recalled" the lead that was assigned to Ojeda. Arney said that the decision as to what file number to put on an FD-302 is determined by the case agent (in this case Rickel) and the case agent's supervisor. Arney said that he did not change any file numbers in this case and would not have changed any file number on an investigation that was requested by another office because he would not be familiar enough with the case to make that decision.
The Tulsa Resident Agency Supervisor, James Hawkins, initialed the block stamp on the copy of the FD-302. Hawkins had no memory of the FD-302 and had no knowledge of how it got to the OKBOMB Task Force. Hawkins told the OIG that it would be appropriate for the lead and the corresponding FD-302 to not have the OKBOMB case number because the lead was for a different case. He speculated that the FD-302 might have been sent to the OKBOMB Task Force because "someone probably saw" that the FD-302 contained references to OKBOMB.
On May 7, 1997, Oklahoma City Special Agent James Carlile forwarded Ojeda's FD-302 to the attention of "Denver Command Post SSA Mark White." The cover EC noted that the FD-302 was to be added to "binder number 3 regarding conspiracy videos." The OIG found Ojeda's FD-302 in that binder. Carlile, who is no longer with the FBI, told the OIG that he did not remember the circumstances surrounding this FD-302. He stated that Rickel was good about sending OKBOMB-related items to him and that he routinely forwarded anything that was in any way related to OKBOMB to the Task Force. Carlile said that he had no idea why the document was not given to the defense, but he speculated that the lack of an OKBOMB case number might have contributed to the problem.
Mark White, who was an OKBOMB Supervisory Special Agent, told the OIG that he put the FD-302 in the binder because of the faxed instructions from Carlile. He said that the information in the FD-302 was common knowledge, that the reporter had been working with the defense, and that the FD-302 had a different case number. He acknowledged that he likely read the FD-302 and saw the potentially exculpatory information contained in it. White reiterated that he filed the document only in the binder, rather than the FD-302 subfile, because those were his instructions from Carlile.
We do not believe that Carlile intended to dictate how the document was processed by the Task Force. White, who was a senior member of the OKBOMB Task Force, should have recognized that the document was discoverable and ensured that it was given to the discovery team. Although this was an error by White, the incorrect case number provides a stronger explanation for the error than the assertion that White intended to conceal exculpatory information from the defense.
During our interviews, we routinely asked FBI employees what they believed caused the belated documents problem. We received numerous suggestions, including the transition from FOIMS to ACS, confusion regarding the rules for retaining originals in the Auxiliary Offices, the volume of documents, chaotic conditions in the Task Force, and confusion over whether an item that appeared to be insignificant should be sent. The employees were correct in recognizing that many factors contributed to the failure to disclose discoverable materials to the defense. We did not isolate one cause, one office, or one individual as being primarily responsible for the belated documents. What we did find was that problems developed when the FBI tried to handle the OKBOMB case, a case of unprecedented scale for the FBI, with a complicated and ineffective document management system.
The first question that we set out to answer was why discoverable documents had not been disclosed to the defense. The evidence shows that both the Task Force and the field offices bear responsibility for the failure to properly disclose discoverable material. Although some field offices continued to insist that they had sent "everything" to the OKBOMB Task Force, we believe that most field offices had items in their possession that had not been sent to the Task Force. One of the primary causes for the failure was the use of teletypes to send information. When the field offices relayed information to the Task Force on a teletype, the field offices believed they had completed the lead; however, because of FBI policies and discovery requirements, the Task Force asked for the information on FD-302s and inserts. The resulting effort to get the teletype information converted to an FD-302 or insert increased the opportunities for documents to be mislaid.
But there were plenty of other opportunities for documents to be mislaid as well. Documents were being routed from Resident Agencies to Headquarters Cities, and they were being sent to FBI Headquarters and many other field offices in addition to the OKBOMB Task Force. A variety of personnel were responsible for processing OKBOMB documents. Generally, the more personnel that are involved, the more opportunities there are for problems to arise as each person believes someone else has sent the document to the right place when in fact no one has. While FBI policies on handling documents exist, field offices have their own practices regarding document processing. The FBI relies heavily on manual systems to determine whether documents have been sent, yet practices in individual field offices undercut the effectiveness of manual systems.57
Also, the FBI has a very complicated document management system that relies heavily on paper. The FBI requires agents to use numerous different types of formats - FD-302s, inserts, memoranda, ECs - and many different forms - FD-71s, FD-209s, and so on. There are also numerous locations for filing items associated with a case: documents go in one place, evidence another, and Elsur and confidential source material go in other locations. While there are valid reasons for separating these categories of items, it does involve more personnel, more forms, more procedures, and more opportunities for items to get lost. A substantial portion of the belated documents were confidential informant materials, which involve a different set of document processing procedures.
While acknowledging these contributing factors, the fact remains that over a third of the belated items appear to be "originals" that should not have remained in the possession of the field offices. The OKBOMB Task Force requested originals repeatedly, and adequate file reviews should have caught most of them. Even if the field believed the items were really copies masquerading as originals, they should have verified that with the OKBOMB Task Force. Therefore, individual agents and supervisors, both administrative and investigative, are responsible for failing to conduct adequate file reviews that would have revealed the presence of items that should have been sent to the Task Force. The failure is even greater given the number of requests made by the Task Force for the field offices to search their materials to find items that had not been sent.
The evidence also shows that documents were going astray at the Task Force as well as in the field. This is in contrast to the FBI's initial position regarding the matter and with the views of some OKBOMB Task Force personnel who stated to us that the field offices' failures to perform properly caused the problem. Again, we believe the root cause was unintentional errors by various individuals. Documents were placed in the wrong files, documents were mislaid, and some documents were simply lost, perhaps because they were sent to one of the numerous teams rather than being filed.
The Task Force document management system also was complicated and involved dozens of personnel, most of whom were on temporary duty and dealing with procedures that were unfamiliar to them. Because the discovery process relied exclusively on copying what was in specific subfiles, any misfiling meant that the item would not be disclosed.
Our investigation showed that Task Force managers did not recognize the systemic flaws in their own document management system. For example, Defenbaugh told us, "We tracked everything. We had to track everything." In fact the Task Force only tracked items once they were serialized by the Task Force and even then, documents could and did wind up in the wrong files with the result that discoverable documents were not disclosed timely to the defense.
Yet, we believe the investigative team could not be expected to devise foolproof ways to plug holes in its inadequate automated systems in the middle of a complex investigation. At the time of the OKBOMB investigation, the FBI as an institution was ill prepared to handle the document management component of such a large investigation. The computer system being used at the beginning of the investigation lacked essential document management components. As a result, the Task Force did not have adequate tools to manage the millions of documents coming from many different places. Although the Task Force attempted to implement quality control measures, the controls that were put in place were ad hoc and, given the volume and pace of documents generated, especially vulnerable to human error. Accordingly, even if the Task Force had recognized the flaws in its document management system, the solutions available to it at that point were limited.
We closely examined the evidence relating to whether the failure to disclose material was intentional. We believe the evidence shows that the failure to disclose material was not intentional. We found no direct evidence that government employees deliberately withheld exculpatory information from the defense. Furthermore, the circumstantial evidence supports such a finding, inasmuch as the information contained in the belated documents was insignificant or had been disclosed in a different form.
In a few isolated instances, agents decided not to send material to the OKBOMB Task Force because they believed the items were insignificant, irrelevant to the OKBOMB investigation, and not discoverable. Because the OKBOMB discovery agreement required disclosure of all FD-302s and inserts and because the OKBOMB Task Force had requested that all OKBOMB material be sent to it regardless of significance, the agents' decisions were not correct. Nonetheless, we believe this evidence only reflected a few instances of incorrect judgment, not an intent to withhold exculpatory information from the defendants.
Law enforcement agencies will inevitably make some mistakes when handling so many documents. Without quality control measures in place to catch the errors, something is bound to go wrong. This is particularly true given the massive quantities of materials coming into the Task Force. And, despite all of the complexity, the OKBOMB document management system lacked an effective quality control or tracking system. Without universal serialization, the OKBOMB Task Force could not determine without substantial effort whether it was missing anything or, indeed, what it had. This problem should be substantially, although not completely, remedied with ACS, which does have such a feature. However, the feature has to be used. As we discussed, with OKBOMB even after ACS came online, the universal serialization was discontinued because of Task Force concerns that field offices were not following instructions regarding serializing documents, causing havoc for the discovery process.
ACS was an important upgrade from the previous FBI computer system, but on the whole it is antiquated and inefficient. It is difficult to use, crashes frequently, and cannot be used as an effective case management tool. The FBI has traditionally dealt with its poor automation system mostly by developing other smaller systems to handle specific tasks. Currently, the FBI is developing a new comprehensive computer system - Trilogy. Because of the limits of this investigation, we did not study Trilogy in depth to determine whether it will make sufficient changes to become an effective and useful system. In the Recommendations section of this report, however, we discuss more specifically some of the issues that we observed that should be taken into consideration as the FBI develops its next generation of automation.
Although our investigation revealed numerous problems with the FBI's handling of the documents in this case, we believe the failings in this case need to be placed in context. The OKBOMB Task Force and the field offices were dealing with what, at that time, was the largest criminal investigation ever undertaken by a United States law enforcement agency. The FBI processed millions of documents and items of physical evidence, conducted thousands of interviews, and managed an investigation that involved thousands of investigators and support personnel from the FBI and other agencies. Our interviews showed immense effort by numerous FBI employees to conduct a successful and thorough investigation.
Furthermore, the nature of the belated documents should also be kept in context. Many of the documents at issue contain information that was disclosed to the defense in other documents prior to the defendants' trials. The belated documents also include considerable information that by any definition is insignificant and would not have aided the defense in its trial preparation. While this does not excuse the failings in this case, it provides context for our findings.
We only obtained limited information about what happened to the 1As in the intervening six years. When Buffalo personnel were boxing up material to send to Oklahoma City in March 2001, they included four 1A envelopes with enclosures. On top of the 1A envelopes was an undated "routing slip" from the OKBOMB case agent to the rotor. On the routing slip, the case agent had written, "Nancy - These literally 'came out of nowhere'! Could you please put them into the appropriate sub-file at your convenience? Thanks, Mike L." The rotor sent the materials to Oklahoma City personnel and attached a yellow post-it note dated March 12, 2001, to the routing slip with the following message: "Linda [Vernon] - the case agent just found these 1A's. Please handle. Thx Nancy." The case agent told us that he believed he had provided the 1As to the rotor in 1997 or 1998; the rotor stated that the agent provided the 1As to her as part of the March 2001 production. All four of these 1A envelopes were identified as belated documents.
For information of receiving offices, at present and since the onset of the OKBOMB investigation, … incoming/outgoing communications are loaded into ZyIndex by Rapid Start team members for search purposes. Lead control numbers are noted on each communication allowing for furtherance of quality control. This system of operation has proven to be extremely reliable, efficient and effective, as well as user friendly…. FBIHQ has mandated utilization of Automated Case Support (ACS) by all field offices, a system which presently proves to be grossly inadequate. However, executive management and all other personnel at the [OKBOMB Task Force] will fully support ACS….
to the USDOJ/OIG Home Page
Return to the Table of Contents