VI. Allegations Regarding Drag Roads
Drag roads are one tool that stations use to estimate how many aliens crossed the border at a particular point. Allegations regarding drag roads were raised in the initial OIA interviews. The OIG investigated further and heard complaints that management had manipulated drag road statistics in an effort to conceal the number of gotaways at Imperial Beach, Chula Vista, and Campo.206 To investigate these claims, the OIG interviewed 200 persons at the various stations and at Sector Headquarters, including 39 agents who had themselves dragged roads.
A. Background on drag roads
Drag roads are used to measure pedestrian traffic crossing a particular location. A dirt road surface is "dragged" by a vehicle pulling objects behind it, usually old tires. The dragged objects smooth the loose surface of the road, eliminating footprints and other imprints on the dirt surface. After a time, an agent returns and counts the footprints (the "sign") visible on the road surface. This process gives him the "drag count." He then drags the road to prepare it for the next count.
The skill of determining the drag count is called "sign-cutting." Some agents are particularly good at this and can estimate how much time has elapsed since the marks were made and distinguish between numerous prints to sort out how many persons have crossed. Skilled sign-cutters are familiar with different shoe markings and can often differentiate between prints made by Border Patrol agents or other legitimate traffic and those made by illegal traffic. Some agents have qualified as experts in court due to their exceptional skill at sign-cutting.
It is only relatively recently that drag roads have been used on a regular basis in the San Diego Sector. In the Sector's westernmost stations, the drag roads are some distance from the border and are used solely to measure how much traffic has made it past the primary levels of agent deployment. In Campo, the dragging is done on the dirt road adjacent to the border. The drag count information is relayed to backup teams that then try to apprehend the illegal traffic.
Where drag roads are in official use, the agents dragging the roads complete reports that indicate the count since the last drag. Weather and manpower permitting, the drag road is dragged at the beginning and end of each shift. This gives a count for traffic during the shift and traffic during shift change. This information is used in deploying manpower. In Campo, the road is continually dragged to locate and intercept traffic as quickly as possible.
B. Allegations that drag road statistics are intentionally inaccurate
Some agents at Imperial Beach alleged that the drag road statistics on the midnight shift were intentionally inaccurate. They claimed that this was accomplished by erasing the drag road shortly before the day shift's count to prevent the day shift from making an accurate count and/or by using untrained new agents to conduct the midnight shift count. Agents also claimed that agents who reported an accurate or high count of gotaways were removed from drag road duty. To resolve these claims, we interviewed 78 people at Imperial Beach, including 26 responsible for dragging roads during Operation Gatekeeper. We also reviewed hundreds of reports of drag road counts covering all three shifts.
Road dragging became an official duty at Imperial Beach in late 1995. In early 1996, an agent on the day shift frequently reported high numbers of gotaways when he conducted the first drag of the morning, which implied that large numbers of aliens had successfully crossed through Imperial Beach on the midnight shift.207 Other agents, particularly some on the midnight shift, challenged his figures. At the same time, the drag reports from the midnight shift were frequently low. The dispute became very heated in mid-1996 when one agent's intelligence report challenged the accuracy of the day shift's figures. The day shift road draggers perceived that they were being called "liars" and were very upset. Out of this dispute arose a series of claims of wrongdoing in relation to the drag roads and the drag reports.
The first claim was that the midnight shift had intentionally dragged the road just prior to the day shift so as to erase the evidence of crossings during the midnight shift and preclude an accurate morning count. Initially, when the day shift draggers conducted their drags at the beginning of their shift, they would find numerous footprints and no evidence that the road had been dragged for hours. There came a time, however, when they began encountering significantly fewer footprints, suggesting that the road had been dragged just an hour or so prior to their arrival. The day shift draggers believed the midnight shift was dragging the road to erase evidence of the number of gotaways during its watch.
We found that the midnight shift did indeed deviate from normal procedure: drags were supposed to be done twice a shift, as early as possible at the beginning, to count the crossings during the shift change, and as late as possible near the end, to count the crossings during the shift. For several months, however, the midnight shift either conducted only one drag or did its second drag long before the shift was finished. When we questioned midnight supervisors about this, they claimed that they had lacked the manpower to conduct the drag. As a result the midnight drag counts tended to undercount traffic crossing during the midnight shift. No data was lost, however, because the crossings after the midnight shift's drag would be picked up by the day shift, which regularly conducted its first cut near the beginning of its shift.
Although a consequence of the midnight shift's practice would have been to increase the number of gotaways recorded by the day shift,208 there would not have been any effect on the total number of gotaways. Any claim attributing such a motivation to the midnight shift therefore makes no sense. As long as a drag is done before the signs of people crossing disappear, the crossings will be counted in a subsequent drag, whenever it is. Moreover, because the day shift attributed the higher count to the midnight shift, there was no significant effect on each shift's statistics.
We find no reason to attribute the midnight shift's failure to do a second drag to any effort to hide information. That shift was significantly busier than the day shift and frequently did not have available the skilled personnel needed to conduct a drag. The drag road was merely an intelligence tool, not a means of capturing aliens. When faced with the choice between placing experienced agents in the field to apprehend illegal aliens or using them to gather intelligence, the midnight shift elected to keep the experienced agents in the field. There is nothing in this decision that evidences wrongdoing.
Ironically, when the midnight shift received additional resources and began conducting two drags, with the second near the end of the shift (as proscribed by the program), this too was viewed with suspicion by the day shift, which claimed that it was done to erase any evidence of crossings before the day shift's first drag. There is no reason to believe that the midnight shift started adhering to proper procedures because it wanted to reduce gotaway counts. The traffic no longer picked up by the day shift's count was simply picked up by the midnight's shift's count; the total number of crossers remained the same.
The second claim of impropriety by day shift members was that the midnight personnel had intentionally undercounted crossings. When the road is dragged, tires are pulled along the center of the drag road. The "shoulders" of the road are not dragged, and the footprints on either side of the area dragged are thus not erased. These witnesses claimed that they regularly saw more footprints along the undragged edges of the road than were reflected in the midnight shift's count. They took this as evidence that the midnight shift did not accurately count the number of crossings. However, the number of footprints on the side of the road would have been relevant to the midnight shift's count only were one able to tell when they had been made. While some sign-cutters can make very good estimates of the age of particular sign, conditions are not always sufficient to be that precise.
In the end, there is no basis to believe that the midnight shift agents were intentionally undercounting. None testified to having been pressured to undercount the sign. Indeed, at the heart of the day shift's complaints was simply a belief that the midnight shift was not sufficiently skilled. Moreover, there was also a dispute between the shifts as to what should be counted. Some members of the day shift believed that every print should be counted, regardless of its origin; one even counted the prints of birdwatchers he saw walk across the road.209 We also heard of instances where a day shift road dragger counted the footprints on the edges as well as the prints across the center. Because the drag road runs east and west, the illegal traffic must cross the road to travel north. Thus, the footprints at the edges are those of people who crossed the road prior to the last time it was dragged. These individuals should have been counted by previous sign-cutters. By counting these prints in subsequent shifts, the agents were double counting prints from previous shifts. In contrast, we found members of the midnight shift who made the effort to distinguish between legitimate and illegitimate traffic. Different agents make different judgments. One Union official, for example, said he only counts the sign if he is absolutely certain that it is from a separate individual; he noted that distinguishing among sign is difficult.
Further evidence that the alleged undercounting by the midnight shift was not a result of any effort to conceal the traffic levels can be found in the responses by management to the shift's drag counts. In a number of electronic messages to Station management in February 1996, a midnight supervisor expressed his skepticism about the very low drag figures. In one, he noted that the low number of gotaways reported were "a little hard to believe because the "90s" (series of sensors) were so active. . . ." Following another low count the next day, he wrote that the shift needed to find a good sign-cutter because it had lost the one that had previously given it an "accurate count." Similarly, messages from the PAIC to Sector Headquarters commented on the low numbers on the midnight shift. In one, the PAIC noted that he did not believe the low figures and needed to find out who was doing the cutting on that shift.
One of the day shift draggers testified that during a time when the midnight shift was reporting few gotaways, the ACPA and an FOS approached him and asked about the disparity in the midnight and day shift drag road reports. He told them that the midnight report was not accurate. A short time later, the midnight shift began to conduct the second drag at the end of the shift. Although the witness giving this account saw something suspicious about this sequence of events, the evidence, when viewed as a whole, shows management was genuinely concerned with the problem on the midnight shift and endeavored to correct it.210
The only evidence we found of intentional understatement of gotaways was the testimony of No. 1922. He testified: "I'm not a good drag person. I can't tell you exactly how many got away, you know? I just more or less `guesstimate'." He later noted that, because high drag counts lead supervisors to inquire about why there were so many gotaways, "I don't want to screw my buddies, so I'm going to write down zero gotaways. It's easier and quicker. Zero gotaways and save my friend's ass."
Outside of this independent decision by a single agent to turn in intentionally inaccurate reports, we found no evidence supporting the day shift's belief that supervisors on the midnight shift intentionally tried to hide evidence of gotaways on their shift.
C. Allegation that figures on drag reports were altered by supervisors
The only allegation we received regarding altered drag reports was at Imperial Beach. A witness testified that another agent had told him that, after reporting a large number of gotaways from the drag road one night, the FOS had told him that the numbers were "not right." The witness said he heard that the supervisor told the agent to change the numbers on the report, but he had refused. The FOS then allegedly said, "Fine, I'll take care of it, give it to me." The witness said the other agent had checked his report the next day and saw that the figure had been significantly reduced. According to the witness, the agent later told him: "If anyone looks into Gatekeeper, I am going to have this guy by the balls because I have a copy. He is going to be in trouble because he changed a government document." The witness said the agent was waving a piece of paper in his hand when saying this.
The agent identified by the witness described the incident differently. He testified that once when he had reported 70 gotaways, the FOS had indicated that the number was too high. The agent ascribed the amount of traffic to the fact that the night had been foggy. The agent inferred that the FOS did not want him to report that many gotaways because the number would have to be explained to someone "up the line." The agent said he told the FOS, "If you want to change it, you change it." The agent kept a copy of his report, so he could compare it to the final report the next day. When he checked and found that the number had not been changed, he threw his copy away. He could not recall the date when this occurred but placed it sometime in the spring of 1996. He said it was the only time he had been questioned about a report, and noted that it was out of character for this FOS to question him. He noted that he had complained out loud when the FOS had spoken to him, but he denied having told anyone that the figure had actually been changed. He noted he might have indicated that if the report was changed, he would have proof, and waved his copy of the report around. He added that, if anyone claimed that he had said the report had actually been changed, they had misunderstood him.
The supervisor involved testified that, one morning the agent had reported a higher than usual number of gotaways (35, he recalled). When he questioned whether the count was accurate, the agent had been "adamant" that the report was correct. He noted that the agent had always been a "hard worker" and that that day was "highly unusual." The supervisor avowed that he had not suggested that the agent change the report, noting that he "would never do that."
Another supervisor on the same shift recalled a night when an agent - the same agent, he thought, who was involved in the incident just described - reported 100 gotaways.211 The supervisor said he and the other supervisor had already driven the drag road and made a rough estimate of the number of gotaways and had not come up with a figure close to the one reported by the agent. Because even the scope operators had not reported that number, the FOS could not understand how there could be so many gotaways. Although the FOS wanted to go back and count again, it was late and did not do so. He believed the FOS spoke to the agent about the figure, but the figure was reported as the agent submitted it.
Although without an exact date we could not look for a specific report, we found several drag reports with 70 or more gotaways reported. This suggests that supervisors did not in fact consistently prevent the submission of reports with high numbers. Several supervisors testified that when drag reports had very high figures they had gone out with the agent to confirm the counts. Each time, they had been able to confirm the reported figures. No one testified that his own report had been altered and no one produced any purportedly altered reports.212 In sum, based on the available evidence, we do not find any misconduct connected with this incident.
The only other claim of improper supervisor interference with drag reports came from No. 1922. He alleged that one evening, while assigned to the intelligence officer/drag road position, an agent had been ordered to give breaks to agents on the line. No. 1922 said this prevented the agent from dragging the roads, and he indicated as much in his intelligence report. No. 1922 also said that several days later, a supervisor (No. 1692) had complained that he had gotten in trouble for the report and demanded that the agent enter a figure to make it appear as if the drag had been conducted. No. 1922 testified that the agent claimed that the supervisor had "sort of implied" that he should erase his explanation and record that the drag report was completed and that there were no getaways. No. 1922 said that the supervisor did not say exactly that, " but that's what he meant by it." No. 1922 testified that the agent said he refused to change the report.
The supervisor involved (No. 1692) disputed this scenario. He said that, on one occasion, the agent's first-line supervisor had asked the agent to give breaks to agents on the line. No. 1692 heard from another supervisor that the agent had reported that these breaks had prevented the drag from being done that day. No. 1692 testified that he believed that the agent had actually dragged the road because on the Daily Intelligence Report there were numbers for the drag. He said he also heard that the Station Intelligence Officer had told the APAIC what happened. When the APAIC asked the agent why he had written what he had, the agent had replied, "to get back at [my first-line supervisor]."213 No. 1692 noted that, although he had asked the agent on several occasions to change the narrative and fix spelling and grammar errors on his intelligence reports, he had never asked the agent to alter any numbers.
No. 1692's testimony is consistent with the documents and other testimony we received. No. 1922 did not provide the exact date of this incident but placed it in October 1996. Upon reviewing the Station's daily intelligence reports from September 1, 1996, through the end of October 1996, we found only two days on which the agent whose report was allegedly changed was the shift intelligence officer. Information we received relating to the second date established that the incident involving the drag report had to have happened on the first date, 11 days earlier in October. In the daily intelligence report, where drag information is recorded by the shift intelligence officer, the box on the form where drag results are normally reported is captioned "FAILURES TO YIELD/GET AWAYS/RETURNED TO MEXICO." For the particular shift involved on the day in question, this box merely says "none." Another shift reported the same thing. This document corroborates No. 1692's account to the effect that he had thought the drag had been completed. When no drag was conducted, the drag report usually was left blank or specifically indicated that there was no drag done. And drag reports frequently used "0" or "none" in the box to indicate that no evidence of gotaways had been found on a particular shift.
No. 1922's account makes no sense. Although he claimed that No. 1692 had "implied" that the drag report should be changed several days after it was written, the report would have been filed the day it was written and its data sent to Sector. Changing it would not have accomplished anything.
Finally, No. 1922 gave two inconsistent accounts of the incident, and his testimony was demonstrably false in several details. For example, he claimed that after this incident the agent involved continued to be the shift intelligence officer and "would pretty much routinely every day be called in to the office and be told, `Don't write this. Do it this way. Don't write that if it doesn't look good.'" Yet the station intelligence reports and the G-481s (the station's daily agent assignment records) show that the agent served as intelligence officer for only three days in the three months surrounding this alleged incident and never consecutively. In one interview, the agent involved claimed the supervisor's reaction to the drag report occurred "a couple of days later"; in the other, it was "a few weeks later." Initially, No. 1922 merely claimed that the supervisor had "implied" the agent should change the report. In the second interview, No. 1922 claimed the supervisor had specifically said "I want you to change it" and that the agent's "character was in jeopardy" if he did not change the report. In No. 1922's initial interview, this reference to "character in jeopardy" arose in relation to a different incident.
Based on all of these factors, we find No. 1922's claim that a supervisor asked the agent to falsify a report not credible.
D. Allegations that drag roads were abandoned when high numbers of gotaways were reported
A Union official alleged that the drag roads at the Chula Vista Station had been abandoned because high numbers of gotaways had been reported. A supervisor at Campo testified that he believed that station had not conducted road drags because it wanted to hide the number of gotaways. To investigate these two claims we interviewed 50 agents at Chula Vista and 18 at Campo, and we reviewed numerous drag road reports from both stations.
Chula Vista instituted a drag road program in late August 1996 purely for intelligence reasons - to measure where and how much alien traffic was successfully penetrating the Station. Agents were sent to the Campo Station for sign-cut training. A study was conducted at Chula Vista to determine where the best locations were to conduct the drag, and a team of agents was assigned to drag the roads on each shift.
The Union official could offer no evidence to support his belief that the drag road program was halted because it was regularly reporting high numbers of gotaways. The supervisors at Chula Vista and the agents who conducted the road drag agreed that the drag road program had been halted, but they all attributed the decision to a lack of manpower. This explanation is consistent with the fact that Chula Vista had manpower shortages because of a combination of increased traffic through its area and an expansion of its area of responsibility. With such resource shortages, it certainly would have made sense to abandon an intelligence tool in favor of the primary Border Patrol functions of deterrence and apprehension.
The drag reports from September 1996 through December 1996 support the manpower shortage explanation. From the program's inception, the drag counts were relatively high - frequently registering over 50 gotaways and on occasion over 100 on a single shift. But the program still continued for several more months. Moreover, on various days, the reports note that a manpower shortage had prevented the drag from being done. Furthermore, when the drag was finally suspended, it was not reporting large numbers of gotaways. During the program's last several weeks, the reports indicate that it was not conducted because of rainy weather. We also find it significant that, although the drag program was discontinued while we were conducting our investigation, we received no complaints about this action. When agents felt that something improper had occurred during the investigation, they were quick to tell us.
Because there was no evidence supporting the claim that the Chula Vista drag road program had been halted because it was reporting large numbers of gotaways, and substantial evidence contradicted the claim, we find this allegation unsubstantiated.
The complaining supervisor at Campo said he had told the PAIC that an accurate gotaway count would require agents to drag the entire border twice a day seven days a week. He attributed the Station's failure to do this to an effort "to avoid the embarrassment that would result from the number of gotaways that would be reported." His claim is contradicted by the facts.
By the time we interviewed this supervisor, the Station had implemented a program in which the entire border road was dragged during each shift on a daily basis. This was part of the new deployment plan that accompanied the large influx of agents to Campo. Drag data not only was used to measure gotaways but was being relayed to "mobile interdiction teams" in backup positions, so they could track and apprehend the traffic.
Although there had been periods before this when the roads had not been regularly dragged, every witness on the subject, save for the complaining supervisor, attributed this to a lack of manpower and equipment. It should also be noted that drags could not simply be reinstituted as soon as the additional agents were deployed to Campo. Many of these agents were trainees who were just learning how to be a Border Patrol agent, and Campo - the largest station in the Sector, with remote and challenging terrain - is considered a particularly difficult station to learn. And even once regular dragging had been ordered, there may have been implementation problems having nothing to do with any desire to conceal data. Even the complaining supervisor commented that many agents were "too lazy to prepare reports," indicating that the failure to count getaways was not due to directives from managers.
Furthermore, the claim that management wanted to hide high gotaway figures is contradicted by the large numbers of gotaways consistently reported for Campo throughout 1996.214 It was not unusual for reports of several hundred "known" gotaways to be included in the Station's reports even during the slower fall months. Moreover, once the drag program was instituted and provided a more accurate count of undocumented aliens entering through Campo, the number of reported gotaways was lower than before the drag program began. This cuts against any inference that the failure to drag the border road earlier was an effort to hide high numbers of gotaways, which were in fact reported. No one, including Station and Sector management, claimed that Campo was under control at that time. Thus gotaway figures in the hundreds were still expected, and there was no apparent pressure to pretend that conditions were otherwise. In short, we found no evidence that at Campo road drags were intentionally left undone to hide evidence of gotaways.
206 One Union official also alleged that he had been told that drag road reports were being falsified at Brown Field. As there was no drag road program at Brown Field, and, consequently, no drag road reports to be falsified at Brown Field, he clearly was mistaken.
207 At the time the midnight shift was not conducting a drag at the end of its shift so the first drag during the morning shift was reporting traffic that may have crossed hours earlier.
208 Because of the location of the Imperial Beach drag road and the push to place manpower closer to the border, any crossings of the drag road were considered likely gotaways.
209 It makes little sense to include these prints in a count intended to estimate the number of illegal aliens who cross the road and get away. Including legitimate traffic subverts the intent of the intelligence gathering process.
210 This witness was quick to infer improper motives on the part of management. He claimed that two agents were taken off the drag assignment after turning in reports with high counts, allegedly because management did not want high figures. This inference is unwarranted. First, if management had wanted to get rid of anyone who handed in high numbers on the drag reports, the witness would have been the first to be removed. Second, both agents he spoke of continued to get drag road assignments long after the witness claimed they had been removed - even after the witness's interview. Finally, other agents who reported large numbers of gotaways were not removed from drag road duty.
211 This supervisor may be confusing two separate events. The same FOS recalled that, on another occasion, a different agent had reported 100 gotaways, which the FOS attributed to the fact that the drag had not been done for several days because of rain. He believed the agent counted all the footprints from the entire period instead of only the recent ones. The total figure was reported, however.
212 A Union official testified that an agent on the midnight shift had complained that he was told to change numbers on his drag report. He could not identify the agent, was not sure whether he was confusing this with the midnight/day shift conflict, and subsequently testified that he was confident that no falsification of reports was occurring at Imperial Beach. We found no midnight shift agents who complained that supervisors altered their reports. As the supervisors on midnights were clearly concerned about the numbers being too low, we believe the Union official was confused. Another agent relayed a third-hand claim that a particular agent had complained that his drag report had been altered. We could find no agent in the San Diego Sector by the name he gave us. The only person with a similar name was a Union official who did not conduct drags.
213 No. 1922 admitted that he wanted to retaliate against this particular supervisor because the supervisor kept asking him to do extra projects. He testified that, "being the kind of person I am," he told the supervisor he would not do these projects. He said the supervisor threatened him with discipline so he decided to retaliate. This clearly suggests that his account of these alleged events may be colored by his views toward this supervisor.
214 Save for several days that followed the dispute over a 1,300 gotaway figure in the summer of 1996, when the reported gotaway level dropped below 100, that was discussed at length in the chapter relating to Sector Intelligence, supra, at 178, the numbers of gotaways reported for Campo were extremely large and at least several times higher than those reported at any other station in the Sector.