5. The report of 1,300 gotaways at Campo
INS Headquarters Intelligence produces the SW Border Weekly, a summary of statistics and events of note along the entire Southwest border. It is distributed throughout INS and to the Attorney General and other DOJ officials. The July 24, 1996, issue (prepared on July 26, 1996) reported an estimate from Campo that, on July 24, there were 1,300 gotaways but only 107 apprehensions.150 At the Congressional hearings, Bonner claimed that management officials, including Williams, had responded to this "realistic" report by "trying to discover who had leaked the information."
Estimating gotaways was a difficult problem for the Border Patrol. The number of apprehensions was easily measured, but there was no good way to determine how many undocumented aliens successfully made it across the border in any given area. Scope reports told only of people actually observed. Drag roads in some areas indicated how many people had crossed a particular area, but they could not always be used because the terrain was too difficult to construct a road through the area. Estimating gotaways at Campo was particularly difficult because of the station's large area, difficult terrain, and shortage of manpower. Having a sense of the number of gotaways was very important, however, for determining where the weaknesses were along the border, where more manpower was needed, and whether particular deployments were effective.
As part of the increased reporting requirements of the Spring Plan initiative that began in January 1996, Sector Intel asked each station to provide an estimate of its effectiveness (the percentage of undocumented aliens entering the area who were apprehended), even if it was a "guesstimate." It did not give the stations any guidance on how to make such estimates. As a result, each station left decisions about how to calculate these figures to the agents assigned to the station intelligence function, none of whom had training in statistical methods. As these positions rotated frequently, there was no standard methodology for calculating effectiveness even within a station. The results were generally well-intentioned estimates that did not necessarily accurately reflect actual conditions.
This was certainly true at Campo. During 1996, six different agents were responsible for preparing the station's intelligence reports and hence for estimating gotaways. The OIG found that each agent devised his own method. Several said they used sign cut reports, agent observations, citizen calls, and other data. One said he simply assumed the station was 10 to 15 percent effective and calculated gotaways based on the number of apprehensions. All conceded that they plainly guessed as to the number of aliens who crossed the station's large unpatrolled areas. One agent said he simply looked at what the agent before him had reported and copied those figures.
Although some witnesses believed the 1,300 gotaway figure that made its way to INS Headquarters was larger than usual, numbers of that magnitude had been reported by Campo for several months prior to the July 24 report. Indeed, there was a February 9, 1996, estimate of 1,500 entrants in one day. Between January and July 24, 1996, a third of the Campo daily reports we found reported 1,000 or more entrants. In June and July, the figure jumped to nearly 65 percent.
One Campo collateral intelligence agent recalled that, in April or May 1996, the PAIC had told him that Williams wanted to know how he calculated total entrants. This suggests that Williams had seen some of the higher figures and questioned them. The witness noted, however, that no one had suggested that his estimates were too high or pressured him to reduce his estimates. This inquiry may have been the reason why Williams allegedly later complained to Sector Intel personnel that the gotaway estimates were invalid. Thereafter, Sector Intel stopped reciting gotaway figures in its daily face-to-face briefings of Williams. Although these figures continued to be included in Sector Intel's reports, they were not directly brought to Williams' attention. They apparently drew no attention from the PAIC at Campo or from anyone else at Sector Headquarters either.
The facts surrounding the release of the 1,300 figure and the reaction to it are not substantially in dispute. Early in the morning of July 25, 1996, the Campo daily report was faxed to Sector Intel and reported 107 apprehensions and an estimate of 1,300 entrants. After this information was incorporated into the Sector Daily Report, an intelligence officer from INS Headquarters called Sector Intel as usual to obtain information for the SW Border Weekly. His notes indicate that he was told that Campo had unofficially estimated 1,300 entrants the previous day and 1,100 on July 22. Apparently, this was the first time this witness had heard estimates of this magnitude.151 He told us that, because the information surprised him, he sought more information from Bradshaw at Sector Intel. His report cited this figure but noted that it was an "informal" estimate and that, if the number of entrants were that high, one would expect higher apprehensions at Temecula and San Clemente.
On July 26 a draft of the SW Border Weekly was sent to Robert Bach, Executive Associate Commissioner for Policy and Planning, and his staff who regularly reviewed such drafts before they were distributed outside of INS Headquarters. Thereafter, one of Bach's staff sent an electronic message to Williams saying that Bach wanted more information on the methodology for collecting this information, such as who collected the information, whether it was from sign-cutting or actually observed aliens, and whether Williams had any idea why the number was so high. The final report, distributed on July 29, contained the 1,300 figure. Thereafter, in a July 30 telephone conversation with the staffer, Williams said he did not know where the INS Headquarters had gotten its information. He noted that the station's gotaway figures were reported by an agent with collateral intelligence responsibility and that he did not review or edit the station reports. Williams also suggested that the number was too high, noting that different agents may report the same group or simply overestimate the amount of traffic.
Williams testified that he had first learned of the 1,300 figure from the Sector Daily Report, which he would have received on July 25, a day before the message from INS Headquarters. Williams said he had not believed it possible to count that many people on a road, but if that many were crossing in one place something had to be done about it. To find out how valid the information was, he called ACPA Ken Stitt (who had oversight responsibility for Campo). Stitt later told him that the PAIC of Campo had been too busy to review the reports and that the statistics were "guesstimates" of areas that had not been checked for three or four days. When Williams discussed the methodology with the PAIC himself, the PAIC said the figure was based on actual counts, scope observations, and sensor activity. Williams later noted that the magnitude of reported traffic did not surprise him, but the exact figure of 1,300 did because he thought it would be hard to count that number. He observed that "Not much attention was paid to the East County because they expected increased traffic."
Williams declared that Bonner's claim that he had tried to find out who leaked the information was "nonsense." His questions, he avowed, were about how the gotaway statistics had been determined, not how INS had obtained them. He also said he never gave Campo any instructions on how to count gotaways and never said they could not estimate activity in unpatrolled areas. He noted that when he saw gotaway estimates fall dramatically after July 26, 1996, he had asked the PAIC at Campo whether those figures were accurate because he thought they should be higher.
Williams may have been confused about where he first learned of the 1,300 figure. Although he would have received the Daily Report a day before the message from INS Headquarters, many previous Daily Reports had contained similar estimates of entrants at Campo without drawing his, or anyone else's, attention. There was nothing in the July 24, 1996, Daily Report's reporting of the daily entrant figures that differed in any material way from previous reports. In contrast, the SW Border Weekly devoted two paragraphs to the 1,300 figure and made it the lead item for the San Diego Sector. While DCPA Harold Beasley corroborated Williams' account by testifying that he saw the 1,300 figure in the Daily Report and flagged it for Williams, Beasley offered no explanation for what allegedly drew his attention to the figure. Unless someone tipped them off that the figure had been reported to INS Headquarters, it would have been an extraordinary coincidence that Beasley or Williams first noticed these figures a day before Headquarters questioned these figures. Given that the intelligence officer at INS Headquarters reported having spoken to Bradshaw about the figure, Bradshaw may have in fact given Sector management advance notice of the report.
Regardless of how Williams first learned of the 1,300 figure, the important point is the nature of his reaction. On this score, other witnesses corroborated Williams' own recollections. DCPA Beasley said Williams asked him to contact ACPA Jesus Martinez to make sure the information was correct before it went to the Attorney General. ACPA Martinez said he recalled receiving a telephone call from Williams asking whether the 1,300 figure was accurate. He recalled Williams' characterizing the figure as "out of the ordinary" and that he and Williams had agreed that it would be hard for agents to count that many footprints. Martinez also said he asked Bradshaw to verify the figure, stressing that, if the information was correct, it could be reported but it should be verified before release. Martinez avowed that nobody had cared about who had given the figure to Headquarters; they simply wanted the figures verified. On July 31, Martinez said he received a note from the DCPA attached to a copy of the INS Headquarters report: "We need to ensure that this type of information, if true, is coordinated thru [sic] you before it is provided to [the person from INS Headquarters]. He calls Intelligence, speaks to someone, and then writes his report I guess." [emphasis in original] Martinez said he responded by sending Bradshaw an electronic message saying "in essence," that if the information were true, it could be given to INS Headquarters, but it had to be valid information. He said he reiterated this point when he later spoke to Bradshaw personally, but he also said he told him that everyone from the collateral intelligence agent on up needed to see when numbers were "out of kilter with everything else" and question the validity.
Stitt testified that Williams had asked him where the 1,300 figure came from and whether it was possible to count that many footprints. He said he recalled that Williams did not "jump all over him" but just wanted to know how they were arriving at numbers of gotaways. Stitt said he had then called the PAIC at Campo and asked him to check into the figure. He said his only concern was that reported figures be substantiated and not merely reflect someone's "gut feel."
That Williams' primary concern was with whether the 1,300 figure was supported, and not the mere fact that Headquarters had received a report indicating a large number of gotaways, is consistent with the fact that the large amount of alien traffic in the East County was no secret. Citizen groups had complained to a wide range of officials in Washington about the amount of traffic in that area, and there had been considerable media coverage as well. Internal INS documents demonstrate that Washington was quite aware of the problems encountered in East County and the lack of sufficient numbers of agents to manage the situation.
The PAIC, APAIC, and agents involved at Campo all agreed that no one from Sector Headquarters had conducted an inquiry into how the figure was forwarded to INS Headquarters, and that no one had demanded that gotaway estimates be reduced. They said they were told that merely guessing at figures was inappropriate. While one agent believed this requirement made the estimates less accurate because there were large areas of Campo that were not patrolled and for which no good estimates were available, he conceded that he did not believe that the station was being pressured to lower estimates of entrants or that the change was done to make Gatekeeper look better. Others concurred that the higher estimates were probably closer to the actual number of entrants than the subsequent lower figures, but conceded there was no evidence to support the higher "gut feel" figures and no evidence that the lower figures were intended to deceive anyone. Indeed, several agents told of recent efforts to obtain more precise estimates of alien traffic, including a system of drag roads along the entire Campo border that is continually monitored. No one contended that anyone had been pressured to change these newly supported figures, which in mid-1997 were averaging around 400 entrants per day.
Only one witness imputed any improper motive to Williams' reaction to the 1,300 figure. A witness from Sector Intel said that Bradshaw had told him that Williams was "very upset" about the gotaway figure from Campo and had wanted to know where the figure had come from. The witness also said Bradshaw allegedly had told Williams the figure did not come from Sector Intel. The witness claimed that, on August 2, Martinez sent an electronic message to Sector Intel saying something about "being careful what they said" to the intelligence officer at INS Headquarters. The witness said that the Headquarters intelligence officer had later told him that he had been "grilled" about where the information had come from and, according to the witness, had said he got it from Sector Intelligence, but said he could not recall from whom.
It is notable that the only person to suggest that Williams was upset about the report was someone without first-hand knowledge. Moreover, he was the only person at Sector Intel who claimed that Martinez had essentially suggested that they withhold negative information from the intelligence officer at INS Headquarters. This claim is inconsistent not merely with Martinez's own testimony but with the DCPA's note that prompted Martinez's electronic message.152
Based on all the evidence, we conclude that Bonner's claims that Williams did not want anyone to know that there were 1,300 gotaways and that he wanted to find the source of "the leak" are unsubstantiated. The evidence indicates that Williams was concerned that guesses rather than documented figures were being reported, and that he therefore ordered that future figures be documented. This was an appropriate request. DOJ management had long been unhappy with the quality of reporting out of INS and the Border Patrol, often complaining that data was not adequately researched and verified. Indeed, then Deputy Attorney General Gorelick testified that she had asked INS to prepare the SW Border Weekly as one means of bringing some order to its information gathering.
We note that the lack of supervision of intelligence production at all levels of management from the PAIC up to and including the CPA was inexcusable. If the stations were going to be required to estimate entrants - a calculation that was central to the question of how effective various aspects of the operation were - they needed some guidance on how to make such estimates, with some uniformity in methodology across the stations. Haphazard guesses at various stations based on approaches selected by individual agents who have had no training or guidance inevitably resulted in meaningless figures.
6. General claims that negative information was deleted from Sector Intelligence reports
In addition to these specific claims, several witnesses more generally alleged that Williams tried to prevent Sector Intel from reporting negative information.
One witness believed part of the PIO's job in editing reports to the WRD was to remove negative information from them. The witness conceded, however, that he had never compared the original version of any such report to its edited version. Williams testified that he had decided to have PIO review and edit the reports to the WRD and the Gatekeeper Updates because the drafts produced by Sector Intel were full of grammatical mistakes. Agents in the PIO agreed that their mission was solely to correct grammatical errors and improve the writing in the reports, and that they had no authority to delete items or change any statistics. Our review of several report drafts with PIO's editing marks on them corroborated the PIO agents' testimony and leads us to find the complaining witness's claim unsubstantiated.
One witness pointed to a May 30, 1996, memorandum from Williams as evidence that he was seeking to control all of the information coming out of the Sector in order to quash any efforts to "tell the truth" about Gatekeeper. The memorandum instructed that, when persons outside the Sector called for information, the stations were to forward the request to a "Sector Staff Officer" who would deal with the request himself or would clear the station to provide the information. Either way, the inquiry and response were to be documented in the Sector's files. The purpose of this procedure, the memorandum noted, was to ensure that the Sector provided "clear, consistent and accurate information."
Defending this policy, Williams said he liked to know what information was going out of the Sector. He said he was particularly concerned that the person who produced the SW Border Weekly had been calling the stations directly and receiving unverified information. Although he had permitted the chief of Sector Intel to freely provide information to INS Headquarters personnel, he had also instructed the agents in Sector Intel to document the data they provided outside the Sector. The DCPA also defended the policy.
Significantly, neither the witness who complained about this memorandum nor anyone else could name an instance when a Sector Staff Officer refused to allow data sought by someone outside the Sector to be released. And no one other than the complainant suggested that the policy was intended to hide the Sector's "dirty laundry." Although a centralized information system can be used to hide unfavorable data, there was no evidence that this one was either created or used for that purpose. Indeed, it addressed some critical deficiencies in the Sector's intelligence operations. With Gatekeeper came an exponential increase in requests for information from all quarters - requests that individual stations were not well-equipped to accurately respond to. Even the Sector Intel personnel who were highly critical of Williams and his policies testified that the reports from the stations frequently contained mathematical and other errors. Moreover, the stations were not always aware of the need to withhold certain sensitive information, like that relating to the investigation into a spate of shootings at agents at Imperial Beach. Given these considerations, a policy providing for a central monitoring of information requested and released was entirely justifiable.
Certain allegations related to a meeting that Williams had with Sector Intel in early 1996. One witness, whom we previously described as being extremely antagonistic towards Williams, took Williams' instructions that Sector Intel should "show we're doing good" as an order to say that Gatekeeper was working even though it was not. Another witness - the source of most of the specific claims that Williams wanted to suppress negative information - recalled that "the gist" of Williams' instructions was not to put anything negative in reports. He conceded, however, that Williams never said this directly, and could specifically remember only that Williams had said something like "make us look good."
In contrast, another line agent who attended the same meeting recalled Williams saying something to the effect that he wanted to be able to show that Gatekeeper was working. Neither this witness, nor any others who attended this meeting, thought Williams' instructions inappropriate. Williams himself recalled that he had only one meeting with all of Sector Intel, and had merely asked the unit to show the progress made during Gatekeeper. His account is born out by notes of the meeting provided by one of the complaining witnesses, that indicate that Williams was seeking to "prove shift eastward," something that he plainly believed was in fact occurring. Given that one of the few points of agreement between critics of Gatekeeper and supporters is that the operation shifted traffic to the east, it is difficult to see Williams' request for information about that shift as an effort to conceal or distort information.
Although we found no evidence supporting claims that Williams attempted to suppress information, one Sector Intel document drew our attention even though no one complained about its contents. An October 13, 1995, memorandum to the stations outlining new reporting requirements noted, "These figures are for Border Patrol internal usage only. There is no need to `fudge' any data reported, including estimations." It went on to stress that estimates of effectiveness were for internal Sector use only, and that "Accuracy in such figures is needed." One possible interpretation of this memorandum is that figures not for the Border Patrol's use were supposed to be "fudged." We found no one, however, who suggested that this interpretation was correct. The person who drafted the document, a line agent in Sector Intel, said he was unaware that data given to anyone was "fudged," and that he had intended only to prevent anyone at the stations from wanting to provide data to make the station "look good." He had recognized that everyone wants to "put on the best face" and had tried to minimize that tendency in the stations' communications with Sector Intel. Others in Sector Intel did not recall seeing this memorandum before and contended that no one had ever suggested that tampering with data had occurred or should occur. Williams said he had never told anyone to "fudge" any information and he had no information that any station had fudged its data.
On this evidence, we find the memorandum more reflective of the disarray in - and insufficient supervision of - the Sector's intelligence operations than of any intentional effort to conceal information.
C. Conclusions regarding the suppression of negative information in Sector Intelligence reports
The OIG could not substantiate any of the specific allegations that Williams (or anyone else) ordered that information be excluded from Sector Intelligence reports because it reflected poorly on Gatekeeper. The OIG located numerous examples of reports containing information demonstrating that Gatekeeper still exhibited weaknesses. Moreover, even had the five alleged incidents occurred, they would hardly have constituted a pattern of suppressing information given that Sector Intel produces about 500 narrative reports a year. It should also be noted that, over the course of Gatekeeper, the number of reports increased as did efforts to obtain more, often negative, information through greater use of scopes, drag roads, and sensors. These efforts cut against any inference that there was a concerted effort to conceal negative indicators of Gatekeeper's performance.
Although there was no apparent intent to suppress information disclosing Gatekeeper's weaknesses, there was significant disagreement about when it was appropriate to report certain data. Sector management clearly erred on the side of caution, hesitating to report isolated incidents lest they be viewed as typical. Sector Intel staff (as well as agents at the respective stations), however, thought it appropriate to report isolated incidents or even rumors. Although it might be tempting to infer that management's caution was due to a desire to protect Gatekeeper, or that the staff practices reflected a desire to undermine Gatekeeper, either inference would be simplistic. Much of the disagreement centered on the role Sector Intel was expected to play. And matters were not helped when, instead of clearly communicating his concerns about unsubstantiated opinions being circulated outside the Sector, Williams resorted to the cryptic May 30, 1996, memorandum placing inadequately explained limits on Sector Intel's autonomy.
Prior to Gatekeeper, little attention was paid to what information went out of the Sector. With the advent of Gatekeeper, the demands for information from Sector Intel, DOJ management, and INS Headquarters increased exponentially but without any coordination or direction. Even when data was provided on a regular basis, such as for the SW Border Weekly, the process remained quite disorganized. Different levels of management requested similar data but in different formats, creating significant overlap and duplication of effort. Nor were the agents in Sector Intel given direction or training that would have enabled them to better cope with these new demands. No one reviewed their work to ensure that the analysis was sufficiently rigorous or objective. As a result, they relied on the information provided by the stations without evaluating its quality. Meanwhile, the gap between what Williams wanted in Sector Intel reports and what he actually received broadened, and job satisfaction in Sector Intel plummeted. The problem was exacerbated by the impression unit members had that Williams did not respect them or their work. Several witnesses described instances when he ridiculed Bradshaw in front of the unit. Additionally, his "robust" language on the draft memorandum to INS Headquarters was dismissive and degrading.
Sector Intel must be an integral part of the Sector's operations and the new CPA must address these issues.153 We offer the following recommendations in this regard:
First, there must be open lines of communication. The CPA must clearly articulate what role he expects Sector Intel to play. Will the members of the unit be free to develop factual information on their own or are they to merely consolidate information produced by others? What authority will Sector Intel personnel have to report information outside the Sector? The CPA must clarify what types of information are appropriate for dissemination. Must it be verified? Are isolated instances reportable (with sufficient caveats to their potentially limited significance)? Or must the information have broader significance to be worthy of repetition? He must also ensure that members of the unit feel free to discuss these issues without fear of rebuke should they raise questions or issues.
Second, there must be a concerted effort to eliminate duplication of information in different formats. The unit needs to work with the various consumers of its output to consolidate and simplify the reporting of statistics. This would free up resources to engage in more substantive work. As a part of this effort, a uniform format should be devised for data that the stations transmit to Sector Intel. At the same time, an effort should be made to identify and obtain useful information, such as whether various directives have been complied with. One example of this is checkpoint information. The INS entered into an agreement with Congress that it would operate checkpoints on a 24-hour basis, and the Sector was given instructions to this effect. Believing the Sector had complied because no one told her otherwise, the Commissioner thereafter assured Congress that this directive had been carried out. Only later did she apparently learn that her assumption was incorrect.
Third, the CPA needs to become less concerned with the potential fallout from every piece of information sent to INS Headquarters and DOJ management. Once he establishes more reliable reporting structures, he should have more confidence that the recipients of Sector data will be able to distinguish between isolated incidents and trends. The 1,300 gotaway incident should give him some confidence in this regard. The figure was reported in the SW Border Weekly, but it was described as unverified and accompanied by several counterbalancing facts suggesting that it might be an overestimate. The fallout from INS Headquarters was limited to a question as to how the number was determined.
Finally, Sector Intel staff needs to recognize that their role is to serve the CPA and the Sector. The reports they produce are the CPA's reports, not their own or the Union's. If that is not a role they are willing to fulfill, so long as the reports are designed to convey balanced and accurate information, they should seek reassignment to a more suitable position.
It should be noted that over the course of the investigation there was some movement in the direction of these recommendations. The OIG was given demonstrations of new efforts to automate data collection from the stations and to simplify some reporting procedures. There is more progress to be made, however, and it will require significantly improved cooperation between the new CPA and Sector Intel staff.
150 Although the INS Headquarters report indicated that there were 1,300 gotaways, the estimate in the Campo report was actually that there were 1,300 entrants. Subtracting the 107 apprehensions leaves an estimate of 1,193 gotaways. Because everyone referred to 1,300 gotaways when describing this incident, we will too.
151 It is not clear why Sector Intel started reporting the high figures to Headquarters at this juncture while making no mention of similar or higher numbers on previous occasions. It is worth noting that this was just one week after Bonner's testimony before the State Assembly Committee that Campo was being overrun and that there were no agents to patrol it.
152 No one retained a copy of Martinez's electronic message. Although a witness from Sector Intel claimed it had been posted on an office door at Sector Intel, he said it had been taken down and he did not know what became of it.
153 During the investigation, Johnny Williams was promoted to the position of Western Regional Director. In May 1998 William Veal, the former DCPA under de la Viņa and CPA of the El Paso Sector, was selected to replace Williams as the CPA of the San Diego Sector. He is expected to assume this position in July 1998.