Return to the USDOJ/OIG Home Page
Return to CUSA Report Table of Contents

V. Criminal History Checking Procedures

  1. Introduction

Of the allegations made concerning the CUSA program, none received as much attention either within or outside the Department of Justice as those concerning INS 'mishandling of naturalization applicants' criminal history record checks did. The possibility that INS had inappropriately naturalized persons who were ineligible to become citizens because of their criminal histories was a cause of considerable alarm. As a result, INS and the Department first concentrated on these allegations. To answer questions from Congress about whether persons with disqualifying criminal records were improperly granted citizenship, INS began a review of all CUSA naturalization cases. In late November 1996, Justice Management Division hired KPMG Peat Marwick LLP, an independent consulting firm, to "oversee, validate, and independently report on" INS' review. The first review focused on how many persons naturalized despite disqualifying criminal history checks or despite not having had records checked at all.

By this time, the deficiencies in INS' criminal history checking procedures had been exposed. One weakness in INS policy was of immediate concern to the Attorney General: that INS adjudicators were trained to assume that an applicant for naturalization had no criminal record if no criminal history report had arrived within a designated amount of time since the applicant's fingerprint card had been sent to the FBI. This long-standing policy, known as the "presumptive policy," allowed adjudicators to assume that an applicant had no criminal record merely on the basis of having failed to receive—by the time of the naturalization interview—any criminal history report. This procedure was clearly more vulnerable to errors than a policy requiring an actual response on each applicant's criminal history. The Attorney General directed INS to abandon the presumptive policy in November 1996 in favor of a policy that required definitive record checks for each applicant.

When Congress learned in March 1997 the preliminary results of the KPMG-supervised review of CUSA naturalization cases, the criticism of INS' criminal history checking procedures grew even more intense. The preliminary results of the review showed that 179,524 of the 1,049,872 approved CUSA naturalization cases (17 percent of the naturalized population) did not have a definitive criminal history check conducted by the FBI,259 and that 10,800 persons (1 percent) had been arrested for at least one felony offense.

Members of Congress reacted with concern not just about the extent of the processing errors indicated by the audit results, but also about what the audit data revealed concerning the reliability of information previously provided to Congress by INS officials in the fall of 1996. Commissioner Meissner had written to Senator Alan Simpson in October 1996 that INS had received preliminary indications that "only a few dozen individuals out of the more than 1.3 million naturalization applicants processed" had been "wrongly naturalized." Executive Associate Commissioner Alexander Aleinikoff had offered the same assessment in his October 1996 statement to the Subcommittee on Immigration of the Senate Judiciary Committee and in an interview broadcast on the radio. In light of the extensive problems indicated by the KPMG-supervised audit, Members of Congress questioned whether INS officials had originally misled Congress by understating the extent of the fingerprint processing problems.

In response to the allegations concerning the errors in criminal history checking procedures, Commissioner Meissner and other INS officials repeatedly pointed to the ill-conceived presumptive policy as the primary source of the problems. In response to the concern that INS officials had misled Congress in their original reports on criminal history checking procedures during CUSA, Commissioner Meissner insisted that the information provided to Congress in September and October 1996 had been provided in "good faith" based on information obtained by querying INS field offices as opposed to a systematic audit of naturalization cases. By March 1997, officials at INS Headquarters conceded that the case review revealed a serious systemic flaw in their criminal history checking procedures, but at the same time continued to maintain that they had made every effort to safeguard the system—indeed, to improve it—and to respond accurately to congressional inquiries.

We examined INS' fingerprint checking process in detail, considering both its performance during CUSA as well as earlier warnings about and responses to process weaknesses. We also examined INS' response to congressional inquiries about criminal history checking procedures. We found that contrary to the assertions made by INS officials at congressional hearings in late 1996 and early 1997, INS had consistently failed to timely or adequately respond to known systemic weaknesses, including the presumptive policy. We also found that INS Headquarters' answers to congressional inquiries concerning criminal history checking procedures during CUSA were replete with mistakes that could have been avoided had INS officials paid sufficient attention to the information then available from the Field. In sum, we found that INS' responses to the congressional inquiries were flawed by the same kind of inattention to detail and to the seriousness of the flaws in the criminal history checking procedures that had marred processing integrity throughout the CUSA program. Although there is no evidence that INS witnesses deliberately gave inaccurate or incomplete information to Congress, there is similarly no evidence that those witnesses made efforts to fully understand the flaws in INS' procedures before offering a "defense" of the work they had done during CUSA. In this chapter, we discuss the evidence that supports these conclusions.

In the discussion that follows, we first provide background information on the criminal history checking procedures employed by INS before and during the CUSA program. We then discuss the origin and definition of the presumptive policy and the admissions INS has made since CUSA about the weaknesses of that policy. Our report on criminal history checking procedures then picks up where INS left off the discussion in March 1997, with the problems, other than the presumptive policy, that contributed to the breakdown of INS' criminal history checking procedures during fiscal year 1996.

After describing aspects of the process that were already discernibly flawed before INS launched CUSA, we turn our attention to the recommendations made to INS by both the OIG and the General Accounting Office (GAO) to remedy the weaknesses in its procedures. We detail INS' inadequate response to those recommendations in the years preceding CUSA. The record shows that INS failed to take action because it downplayed the importance of criminal history checks in naturalization processing integrity.

Once this history is detailed, we examine what happened during fiscal year 1996. We describe how the ambitious data-entry project in Laguna Niguel, California (described above in the overview and A-files chapters of this report), was undertaken with a single-minded focus on naturalization production and without regard for the ramifications such high-volume processing would have on INS' ability to check applicants' criminal history. We also discuss how strategies ostensibly designed to improve fingerprint processing, including the transition to Direct Mail and the establishment of the Fingerprint Clearance Coordination Center, were hastily planned and poorly implemented because of the same overriding focus on increasing production. In the end, these strategies not only failed to improve INS' procedures, but also worked to exacerbate the problems. Our report then describes INS' failure to timely adjust its procedures when it learned about the FBI's increased processing times later in the CUSA season.

We next describe the weaknesses in INS' biographical check, or "bio-check," a procedure used by INS to determine whether an applicant is being investigated by any other federal agency. This status would not ordinarily be revealed during the course of a routine fingerprint check unless the federal investigation had resulted in an arrest. We conclude that INS' bio-check procedures were flawed and contributed to the lack of processing integrity during CUSA.

Finally, we examine the events of August, September, and October 1996, a time period during which many of the flaws in INS' fingerprint processing procedures came to public light. It was also during these three months that INS officials made representations to Congress about the processing of applicant criminal histories that were later called into question. We describe what information INS Headquarters officials learned about the mishandling of rap sheets260 in the Field and when they learned it. By juxtaposing this with what Headquarters officials were simultaneously reporting to Congress concerning fingerprint processing, we show that there was indeed cause for congressional concern about the reliability of reports from INS. Instead of providing Congress with an accurate picture based on available information, INS dramatically understated the nature and extent of the problems that existed in its criminal history checking procedures.

  1. Background on criminal history checks and the presumptive policy

    1. Purpose and description of criminal history checks

All naturalization applicants must possess "good moral character" to be eligible for citizenship. As discussed above in our chapter on interviews and adjudications, specific acts or conduct by an applicant will preclude a finding of good moral character, almost all of which pertain to criminal activity. The regulations list specific crimes such as a murder conviction at any time, as well as categories of criminal behavior, such as a violation of any U.S. or foreign drug law during the 5-year period immediately prior to the filing of the application for naturalization (except when the violation was a single offense for possession of 30 grams or less of marijuana).

Historically, as we described in our chapter on interviews and adjudications, INS conducted background investigations of applicants to obtain information necessary to assist the adjudicator in evaluating the applicant's character and to comply with its statutory obligation to investigate the applicant. INS investigators typically interviewed the applicant's neighbors and co-workers, required applicants to produce character witnesses, and queried FBI and other databases for information about an applicant's criminal history. Over the years, however, the huge demand for naturalization necessitated changes in INS' procedures, and the frequency of such wide-ranging background investigations became less commonplace.261 By 1991, INS generally did not conduct its own background investigations of applicants and instead relied almost exclusively on criminal history checks performed by the FBI and on bio-checks to obtain information bearing on the "good moral character" of the applicant.

The fingerprint check compared the applicant's fingerprints to those on file with the FBI to determine whether the applicant had any criminal arrests.262 The "bio-check" reviewed information maintained by the FBI and Central Intelligence Agency (CIA) concerning persons associated with a federal investigation, such as an organized crime or terrorism investigation, rather than information on arrests or convictions. The applicant's name and other biographical data were used by the FBI and CIA to conduct this research, and the information was submitted by INS, either electronically or manually, on Form G-325A.

In this chapter, we focus primarily on the fingerprint check because that was the criminal history checking procedure at the heart of the congressional allegations about processing errors during CUSA. No allegations were made concerning INS' bio-check procedures. However, during the course of our investigation we learned that INS' bio-check procedures were also compromised during CUSA. Accordingly, although we do not address the bio-check process in detail, later in this chapter we offer a description of how INS failed to maintain this important facet of naturalization processing integrity during CUSA.

    1. Processing by the FBI

      1. Introduction

The Criminal Justice Information Services (CJIS) Division of the FBI conducts fingerprint checks for naturalization applications. The CJIS Division maintains several databases of fingerprint data acquired from federal, state, and local law enforcement agencies across the United States. Although submissions are voluntary, the CJIS Division has the world's largest repository of criminal history record information with more than 132 million fingerprint cards in 1997, representing over 35 million people who had been arrested for crimes.263

Beginning in 1990, Congress authorized the FBI to charge a fee for fingerprint checks it provided to other federal agencies for non-criminal justice and non-law enforcement purposes. Effective January 1, 1990, the FBI began charging INS and all federal government agencies $14 for each fingerprint card submitted for civil applications, such as naturalization and other immigration benefit applications.264 INS is one of the FBI's largest customers for fingerprint checks for non-law enforcement purposes, paying well over $10 million each year from 1990-1995 for fingerprint checks on benefit applications. In FY 1995, when the user fee increased to $18 for each fingerprint check, INS paid the FBI approximately $26 million, $16 million of which was specifically for fingerprint checks on naturalization applicants. For FY 1996, INS paid the FBI $32.5 million to conduct fingerprint checks for benefit applications, with approximately $19.5 million spent for naturalization applications.265

The fingerprint process begins when applicants submit their fingerprints to INS on fingerprint cards (Form FD-258) along with their naturalization application.266 The FD-258 is stamped in advance with an "originating agency identifier" or "ORI" code indicating which INS office is submitting the fingerprint card. The form is then detached from the application by an INS applications clerk and mailed to the CJIS Division of the FBI.267 The process of separating the fingerprint card from the naturalization application was known as "stripping" the fingerprint cards.

      1. Conducting fingerprint checks

Before a fingerprint card was checked against the CJIS Division's repository of fingerprints, FBI technicians took a number of preliminary review steps to ensure that the fingerprint card had been properly prepared. Fingerprint cards would be rejected if pertinent biographic information such as the applicant's date of birth, sex, or name had been omitted from the upper portion or "masthead" of the fingerprint card. If so, the FBI sent the FD-258 back to the INS office that had submitted the fingerprint card, generally with a notation explaining what category of information was missing.268 This type of rejection is known as a "masthead reject." The FBI did not charge for cards rejected at this stage in the process and kept no record that it had ever received the card. The missing or incorrect information had to be supplied or corrected by an INS clerk who would resubmit the card to the FBI, or the INS district office had to obtain the information from the applicant and then resubmit the fingerprint card.

If the masthead biographic information was complete, FBI technicians attempted to classify or categorize the fingerprints based on the loops, whorls, arches, and ridgelines of each print. Each fingerprint would receive a "score" concerning its classifiability. Some cards contained fingerprints that were smudged or were otherwise not legible enough to permit classification of all ten fingerprints.

The fingerprint cards then underwent a "name-check"—an automated search of the FBI's criminal history database using the applicant's name, descriptive data from the masthead of the fingerprint card, and the types of fingerprint classifications, to the extent that the fingerprints were of sufficient quality to determine classifications.269 Because of the possibility of aliases, misspelled names, multiple surnames used by some immigrant groups, and other errors or variations of names, the computer software was designed to search not only for exact spelling of names but also closely-related names and aliases. For example, the last name "Lin" could also be spelled "Len" or "Lyn," and the computer would look for these other names as possible matches.

A name-check search could yield many potential matches. The computer would rank the potential matches according to the number and type of criteria that matched. FBI technicians then compared the fingerprint card on file for the top two candidates identified by the computer with the fingerprints on the submitted card.270 If a match was identified (and verified by a second examiner), the criminal history report or "rap sheet" relating to the person whose fingerprint card was on file at the FBI was attached to the FD-258 and returned to the originating INS office.

If there was no match identified by the name-check process, fingerprint cards that had 10 classifiable fingerprints were submitted for an automated search of the FBI's fingerprint files. If the automated search resulted in an "indent" or match, an examiner verified the result and the criminal history report was forwarded to the submitting agency. If the automated search did not yield an "ident," the FBI destroyed the submitted fingerprint card in accordance with INS' presumptive policy that was established in 1982 and no other information was provided to INS.

The classifiability of the fingerprints on the submitted card affected not only the FBI's ability to use the automated reader, but also the potential success of the name-check process. First, the lack of classifiability would limit the FBI's ability to use the "type of classification" (or type of loops and whorls) criterion for the name-check process in selecting or ranking possible matches. Second, if the name-check process produced possible matches, the quality of the fingerprints on the submitted card determined the ability of an examiner to compare them to the fingerprint cards of the potential matches at the FBI. Although some FBI examiners had sufficient experience and skill to make a positive identification based on only one classifiable print on the submitted card, that was not always possible. If the fingerprints on the submitted card were not suitable for comparison to fingerprint cards on file at the FBI and identified as a result of the name-check, there was no method for confirming that the person whose fingerprints were on the submitted card was the same as the person whose fingerprints were on file at the FBI. In such cases, the unclassifiable card was returned to the submitting office with a notation of "no record based on name-check." The submission of a classifiable fingerprint card was thus the only method of ensuring a complete and reliable FBI criminal history check.271

The FBI charged INS for each fingerprint card that it analyzed through the "name-check" stage. However, if the fingerprints were deemed "unclassifiable," the FBI provided INS the opportunity to submit a new fingerprint card for the same applicant at no extra charge. In order to avoid being charged again for the same applicant, the FBI required that INS submit the original, unclassifiable card with the new card to be processed.

    1. Origin and elimination of the presumptive policy

INS' core operating principle with respect to the fingerprint process before and during CUSA was known as the "presumptive policy." Under this policy, INS assumed that an applicant did not have a criminal history if a rap sheet was not received within a certain number of days after the fingerprint card had been sent to the FBI. The designated waiting period was known as the "presumptive period."272 This policy was established after INS and FBI officials met in December 1981 to discuss ways to improve fingerprint processing between the two organizations. In January 1982, INS Acting Commissioner Alan Nelson wrote to the Director of the FBI and requested, among other things, that the FBI return only rap sheets to INS and that applicants' fingerprint cards that turned up no record of criminal activity be destroyed by the FBI for all naturalization, adjustment of status, and asylum cases. In February 1982, the FBI agreed to destroy rather than return to INS all fingerprint cards submitted by INS for which a search against the FBI's criminal history database resulted in no criminal record.

As specified in Commissioner Nelson's 1982 letter, this change in procedure was intended as a "system improvement." As explained to the OIG by INS Headquarters officials, INS' processing of fingerprint cards, including the filing of rap sheets and returned fingerprint cards indicating a negative response, had historically been viewed as demanding a great deal of clerical resources because the process was "paper driven." As the number of naturalization adjudications increased, so did the length of time it took INS to process cases, in part because of the time required for filing all of the FBI responses. Adoption of the presumptive policy was an attempt to address this issue. Because the majority of applicants did not have criminal records—according to the FBI, the figure has historically been approximately 90 percent —the decision to stop receiving fingerprint cards from the FBI indicating a negative response was intended to significantly decrease the filing burden on INS clerical staff.

The presumptive policy, intended to improve the process, was inherently flawed. Since the policy was based on the assumption that the fingerprint card would be processed within a specified timeframe—both at INS and at the FBI—any delays in this timetable could result in an applicant being naturalized without having his or her criminal history reviewed. As Associate Commissioner Crocetti acknowledged to the OIG, "we knew that there would be a percentage of cases, a risk, that would not get to the files in time" for the interview. He explained that the policy resulted from an effort by INS to balance the risk to the public against the obligation to process naturalization applications in a timely manner. As Commissioner Meissner explained to the OIG, most decisions in government are governed by a paradigm that assumes errors are going to occur "and the question is what is the range of those error rates and what are acceptable error rates." INS adopted this approach to fingerprint processing even though it never made a specific determination as to what an acceptable error rate would be.

The application of the presumptive policy as a "cost-effective" approach to naturalization eligibility came as a surprise, however, to the Department of Justice leadership. After the initial congressional hearings about the failings of CUSA, Assistant Attorney General for Administration Stephen Colgate told the OIG that he recalled arguing with EAC Aleinikoff about INS' view that the margin of error in naturalizing applicants with criminal records was insignificant in view of INS' high production rate. AAGA Colgate said he found this "business notion" unacceptable because the "product" was citizenship.

The Attorney General also found it unacceptable. She told the OIG that she had been dismayed to learn in the fall of 1996 that INS relied on an assumption that records had been checked and that no disqualifying information had been found. When she discovered in November 1996 that INS had not implemented steps to ensure that a definitive policy was in place, she ordered an end to the presumptive policy and the immediate implementation of a definitive record check.273

When viewed from outside INS, the inherent weakness of the presumptive policy seemed obvious. Not only did AAGA Colgate and the Attorney General recognize its vulnerability, so did Members of Congress, who likened it to the belief that "no news is good news." However, other factors contributed to INS' maintenance of this vulnerable policy for more than a decade and a half. According to INS managers, there was a common belief throughout INS that naturalization applicants were not a population of persons particularly likely to have a significant criminal history, so fingerprint checks rarely revealed conduct that would have an impact on the adjudication. INS officials told the OIG that it was also commonly believed that in those instances where an applicant had a criminal history, INS had other tools at its disposal—in particular, the naturalization interview—for uncovering that history, thus making the fingerprint check moot in many cases. These notions, as discussed more fully later in this chapter, increased the sense that fingerprint checks for every naturalization case were too costly compared to the rare benefit—the occasional criminal history for which INS had not been aware—to INS or to the public. These, too, were flawed notions that reflected misunderstandings by Commissioner Meissner and senior INS staff of the state of naturalization processing when CUSA began, but they subtly helped to extend the longevity of the presumptive policy until it was dismantled in the aftermath of CUSA.

    1. INS blames the presumptive policy for its fingerprint processing errors during CUSA

In her prepared statement for a March 5, 1997, Joint Hearing before the Subcommittee on National Security, International Affairs, and Criminal Justice of the House Committee on Government Reform and Oversight and the Subcommittee on Immigration and Claims of the House Judiciary Committee (hereinafter March 1997 Joint Hearing), Commissioner Meissner described the "lessons learned" by INS concerning its fingerprint processing procedures. She pointed out that while INS had been relying on a presumptive policy of 60 days the FBI had been completing "73 percent of fingerprint checks within 30 days, 89 percent within 45 days, 94 percent within 60 days, and 98 percent of fingerprint cards within 90 days." She explained:

This policy created a significant vulnerability. But it was a vulnerability that was not apparent until INS eliminated backlogs and became timely in its processing because with backlogs, any "idents" not delivered in 60 or 120 days still had plenty of time to reach the file before a case would be adjudicated.

As INS added personnel and other resources, our processing times dropped, and by August of last year [1996], a few of our offices were completing applications that had been filed less than six months before. As a result, INS was adjudicating some naturalization applications while the FBI was still in the process of completing its background checks.

Commissioner Meissner noted that "despite the improvements to the process" made by INS during CUSA (a reference to the transition to Direct Mail for naturalization applications and the creation of the Fingerprint Clearance Coordination Center (FCCC)), reliance on the "outdated assumption" that INS would receive "idents" from the FBI within 60 days had been INS' mistake. In her interview with the OIG the Commissioner reiterated that the combination of reliance on a presumptive period and decreasing INS processing times was at the heart of the fingerprint processing errors INS made during fiscal year 1996.

Many other INS officials, including EAC Aleinikoff and Associate Commissioner Crocetti, offered this same analysis to the OIG. They insisted that because FBI processing times were increasing while INS processing times were decreasing, they failed to timely notice that criminal history reports were being returned late. The implication was that the primary error made by INS was an inadvertent consequence of the improvements INS had made in its naturalization processing.

Although we do not question the sincerity of these witnesses' assertions, the evidence shows that INS' many mistakes in fingerprint processing during CUSA cannot be attributed solely to the fact that INS followed a policy that, in the name of cost-effectiveness, took the risk that the FBI would fail to process a fingerprint card within the 60 days allowed by the INS processing schedule. After all, as Commissioner Meissner noted, the FBI processed 94 percent of the INS-submitted fingerprint cards within 60 days. If INS had been administering its presumptive policy properly, only six percent of its naturalization cases would have been vulnerable, an error rate that, while unacceptable, would have been a considerable improvement over INS' actual performance.274 The more troublesome role of INS' presumptive policy was that, by permitting the naturalization of an applicant regardless of what might have happened to his or her fingerprint card after its initial submission to INS, it permitted INS to persist in its presumption that all of the intermediate steps involved in a proper fingerprint check—whether those steps had to be taken by the FBI or by INS itself—were in fact being taken. The presumptive policy enabled INS to ignore all of the ways in which the agency itself—not the FBI—was failing to properly process applicants' fingerprint cards.

Thus, emphasizing the vulnerability of the presumptive policy to explain INS' processing failures is inaccurate. Our investigation found that INS' failure to timely or adequately check applicants' criminal histories during CUSA was attributable to its failure to manage fingerprint procedures appropriately even within the boundaries of the presumptive policy. INS failed to ensure that applicants submitted suitable fingerprint cards or that once the FBI rejected an unsuitable card a new one was submitted in its place. Even in times of long backlogs, INS failed to ensure that the Field did not schedule interviews of an applicant until 60 days after having sent the card to the FBI for processing. INS failed to ensure that criminal history reports were timely placed in applicants' A-files or otherwise made available to adjudicators in time for the applicant's naturalization interview. Had all of these fundamental steps been properly and timely followed, INS could have succeeded at checking the overwhelming majority of applicants' criminal history records even under its presumptive policy.

We begin our report where INS testimony in 1997 left off. Having conceded the vulnerability of a presumptive policy, we next address those issues which, in addition to the weaknesses inherent in a presumptive policy, contributed to the sub-standard quality of criminal background checks of naturalization applicants even before CUSA. INS' failure to address these weaknesses before launching the largest naturalization program in its history, combined with the missteps that occurred during the implementation of CUSA, resulted in fingerprint processing procedures that were extremely deficient and inadequate to prevent those with disqualifying criminal records from naturalizing.

  1. INS failures to properly administer fingerprint policy and procedures that pre-dated CUSA

Prior to CUSA, INS did not have a comprehensive fingerprint policy for the naturalization program setting forth the basis for the presumptive policy, the length of the presumptive period, or the procedures to be followed in order to ensure timely processing of fingerprint cards and responses from the FBI. The vacuum created by the failure of INS Headquarters to provide guidance to the Field resulted in disparate practices. Some districts did not send fingerprint cards to the FBI at least 60 days before the naturalization interview. Many districts failed to ensure that "rejected" and "unclassifiable" cards were resubmitted, thereby resulting in the naturalization of applicants who never had their criminal histories checked by the FBI. Districts also failed to ensure that rap sheets received from the FBI were placed in the appropriate file before applicants were interviewed. The record shows that INS Headquarters had information about these disparate and problematic practices and yet did not take steps to remedy them by identifying proper procedures and disseminating instructions before launching CUSA.

    1. Failure to properly administer the presumptive period

Commissioner Meissner, in testifying before Congress about the presumptive policy during CUSA, referred to the presumptive period as being set by policy at 60 days. The record shows, however, that INS had no definitive 60-day policy or practice and that, as a result, some districts waited only 40 or 45 days for a fingerprint card to be checked by the FBI. In those offices, the risk that the adjudication would be completed before the applicant's criminal history had been checked was even greater than under the presumptive policy described by Commissioner Meissner.

      1. INS failed to clearly articulate the presumptive policy

During our review, we found that no definitive policy statement existed concerning the presumptive policy and the processing of naturalization applications. We found only two documents that addressed the issue of a presumptive period for certain background checks for applicants for various INS benefits. However, the two articulations of the presumptive period specified different time periods, and one of the documents was unclear as to whether it referred to the presumptive period for bio-checks (Form G-325A), for fingerprint checks, or for both. Although in an understated way, Commissioner Meissner acknowledged this lack of policy when she advised the Attorney General in a November 1996 memorandum summarizing the presumptive policy that, "INS' implementation of [the presumptive policy] is somewhat unclear."

INS' Operations Instructions (OIs) provided procedural guidance in connection with certain provisions of Title 8 of the Code of Federal Regulations. As its name suggests, these instructions were intended to be a resource for INS personnel who were seeking additional information about INS procedures, including naturalization procedures. However, they were regarded throughout the Field as out-of-date, incomplete, and rarely of much practical assistance.

A review of these Operations Instructions confirms that they would not provide employees with a clear articulation of the 60-day presumptive period in relation to fingerprint cards and naturalization applications. In fact, we found only one instruction that related to the presumptive policy in connection with naturalization applications, Instruction 105.10.275 This instruction, dating from 1982, was the sole articulation of the presumptive policy in any Operations Instruction, memorandum, or directive made available to the OIG as part of our investigation. The instruction applied to all fingerprint checks and bio-checks required in connection with any INS application or petition, not just applications for naturalization.276 However, Instruction 105.10 stated that an application could be processed on the assumption that the response from the FBI was negative if no response had been received within 40 days.

In her November 1996 memorandum to the Attorney General describing the fingerprint process, Commissioner Meissner implicitly acknowledged that INS had, at one point, a presumptive period of only 40 days. She explained that while the 40-day period formerly had been used for naturalization cases, it had "evolved" well before the 1990s to "a normal 60 day suspense." Her memorandum does not explain how the presumptive period "evolved" from 40 days to 60 days.

A review of the various memoranda issued by INS Headquarters to the Field concerning criminal history checking procedures (produced by INS in connection with congressional hearings and the OIG's document request) revealed only a single reference to the presumptive policy. This reference was different from both the description of the presumptive period contained in the Operations Instructions and the description offered by Commissioner Meissner in her 1996 memorandum to the Attorney General. A February 1989 memorandum from INS Headquarters to the Field specified the presumptive period as 45 days, although it is difficult to discern whether the 45-day period related to fingerprint checks, bio-checks, or both. While the subject of the memo is the bio-check process in naturalization cases, the language provides: "[t]he waiting period for FBI clearance will continue to be 45 days to allow for processing the fingerprint record check. [Emphasis added.]"

In our review, we found one other piece of information indicating that some offices adhered to a 45-day waiting period. When INS designed its NACS computer system (used to process naturalization applications), a case was considered "ready" to be naturalized provided 45 days had been allowed for the fingerprint check.

Consequently, as INS approached the start of CUSA, there was no clear articulation to the Field of what the presumptive policy was or how long it was supposed to last. INS instruction manuals and other memoranda mentioned timeframes of 40, 45, and 60 days. Given the lack of a comprehensive, written fingerprint policy for naturalization applicants and minimal (and conflicting) guidance with respect to the presumptive period, predictably there was no uniform understanding within INS of the length of the presumptive period, as described below.

      1. No uniform understanding of the presumptive period existed within INS

Most of the INS Headquarters officials interviewed by the OIG, including Commissioner Meissner, Deputy Commissioner Sale, David Rosenberg, Executive Associate Commissioner Aleinikoff, and Associate Commissioner Crocetti, indicated that the presumptive period had been 60 days prior to and during CUSA. However other INS officials, including Michael Aytes, who was the Assistant Commissioner for the Benefits Division and oversaw INS' service centers, and O'Reilly, who served as the Los Angeles CUSA site coordinator and later during CUSA was placed in charge of the naturalization program Service-wide, told the OIG that they believed the presumptive period before and during CUSA was 45 days. O'Reilly specifically recalled his work designing NACS to support his belief that the period was definitely only 45 days. In addition, notes taken by Elaine Kamarck of the Vice President's Office, who was briefed by Aleinikoff, Rosenberg, and other INS officials in February 1996 on the basic principles of naturalization processing, reflect that she was informed that the presumptive period was 45, not 60, days. The evidence thus indicates that it was not uniformly understood at INS Headquarters during CUSA that the presumptive period had been expanded to 60 days as the Commissioner described in her memorandum to the Attorney General.

More important than Headquarters officials' varied understanding of the length of the presumptive period, however, was the awareness at Headquarters that districts applied differing presumptive periods. As set out below, we found conflicting interpretations of the presumptive period in the Field, and no evidence of any attempt by INS Headquarters to clarify the policy before or throughout most of CUSA.

Witnesses in the Key City Districts offered the OIG a variety of descriptions of the presumptive period and how they understood it in the years preceding CUSA. In Chicago and New York, most District managers described the presumptive period as 60 days, but supervisors closer to the ranks of the DAOs in each District said the period had been 45 days until it was changed in the summer of 1996. In Los Angeles, the largest district in the country, District managers and line employees had different understandings of the presumptive period. While most managers told the OIG that they believed that the presumptive period was 60 days, others believed that it was 45 days. An informational packet sent to community groups participating in off-site processing also advised the organizations that 45 days were required after the receipt of the application for the FBI to complete the fingerprint check. Those who worked most closely with naturalization adjudications, however, understood the presumptive period to be 40 days, and a processing sheet for naturalization cases that was used during CUSA reflected a 40-day rule.277 In the Miami District, a 45-day presumptive period was used consistently before and during most of CUSA.278

      1. The risk created by shorter presumptive periods

Processing naturalization cases according to a policy that allowed 40-45 days for the fingerprint check increased the risk that the FBI would not be able to process the record in time. From January to September 1996, for example, approximately 76,986 cards took more than 45 days but less than 60 days for the FBI to process.

Much like Commissioner Meissner's testimony at the March 1997 Joint Hearing, many INS managers have stated that confusion about the exact length of the presumptive period was irrelevant because of the large application backlogs and the many months of delay between submission of an application to INS and the naturalization date. This view, however, is predicated on the assumption that an applicant's fingerprint card was sent to the FBI at the time the application was initially processed. In the absence of any directive from INS Headquarters specifying that fingerprint cards were to be sent to the FBI as soon as applications were received, districts instituted different practices whereby processing of fingerprint cards was delayed. Consequently, the supposition that large backlogs meant ample time for the fingerprint processing was flawed. Experiences in at least two cities contradict the inference that large backlogs obviated the need to clarify the presumptive period.

        (1) Los Angeles District

In the Los Angeles District, extremely large backlogs did not mean that more time was available in which an applicant's fingerprint card could be checked. Regardless of the length of the backlog, clerks did not strip the fingerprint cards from the applications until it was time to schedule applicants for interview. This practice persisted well into CUSA.

The Los Angeles District Director Richard Rogers noted in a memorandum to INS Headquarters in April 1994 that applicants were waiting approximately seven months to be interviewed,279 but the stripping and sending of cards was occurring almost three months after receipt of the application. While the memorandum stated that the four remaining months were still "an ample period" in which to obtain and interfile any related rap sheet, the memorandum nevertheless made clear that the cards were not being stripped and sent at the time applications were received by INS. Naturalization Section Chief (and later Deputy Assistant District Director for Adjudications) Donald Neufeld told the OIG that he recalled that before CUSA, data-entry and fingerprint card stripping was occurring only 60 days before interview.280 As Section Chief Neufeld pointed out, it was clear that during most of the time that applicants were waiting for their naturalization interviews, their fingerprint cards were still sitting at the INS office and were not being processed by the FBI.

        (2) Chicago District

The Chicago District followed a similar practice to the one followed in Los Angeles before CUSA. Naturalization applications were stored with fingerprint cards attached and not stripped or data-entered until two months before the interview date. In addition, clerks did not send fingerprint cards to the FBI as soon as they were stripped but collected them in a box in the middle of the room and sent them to the FBI every two to three weeks when the box was full. Although Chicago District officials made some changes to the process to address the delay in the submission of fingerprint cards in 1994, the staff could not keep pace with the increasing demands of production and the practice of sending the fingerprint cards to the FBI in a batch was continuing as late as December 1994. By not sending the fingerprint cards to the FBI until less than two months before the interview, the Chicago District increased the likelihood that an applicant would be interviewed before the fingerprint check had been completed.

    1. Failure to resubmit fingerprint cards rejected by the FBI

In addition to sufficient time to process an applicant's fingerprint card, a successful background check also required that the FBI receive a card that was capable of being compared to information in its databases.

The record shows that the FBI historically rejected a significant number of cards submitted by INS and that INS Headquarters was well aware of the high rejection rate. According to FBI officials interviewed by the OIG, INS' historic rejection rate of approximately 15 percent was higher than other agencies. In October 1991, INS Headquarters reported to the Field that the FBI was rejecting 25 to 30 percent of INS fingerprint cards.281 According to an internal INS Headquarters report, in fiscal year 1993 approximately four percent of the fingerprint cards submitted by INS were masthead rejects and approximately 11 percent of fingerprint cards were unclassifiable.

INS Headquarters officials asserted in interviews with the OIG that it had been INS policy to resubmit cards returned by the FBI. Associate Commissioner Crocetti even expressed his belief that guidance had been provided to the Field in this regard. However, the evidence shows that for many years before and throughout most of CUSA (until fingerprint cards were processed through the Fingerprint Clearance Coordination Center, an INS centralized processing facility that opened in June 1996) no such resubmission policy existed.

With respect to fingerprint cards rejected by the FBI because of masthead errors, INS failed to articulate any policy guidance to the Field before June 1996 on what they should do upon receiving the rejected card from the FBI, despite the fact that the failure to resubmit had been commented on in an OIG inspection report in early 1994. With respect to cards rejected by the FBI because of "unclassifiability," not only was there no policy requiring resubmission, we found evidence of a policy that affirmatively permitted the Field to adjudicate naturalization cases despite the rejection.

The contradiction between what INS Headquarters officials asserted was INS policy and what the evidence shows actually happened in the Field is similar to the dichotomy that existed about the presumptive policy discussed above. The lack of uniform understanding about INS' resubmission policy similarly resulted in disparate practices in the Field before and during CUSA. As discussed below, none of the five Key City Districts required resubmission of rejected cards in every case.

The Field's failure to resubmit rejected cards remained uncorrected through July 1996. Thus, as ultimately reported by KPMG, of the 1,049,867 persons originally identified as having naturalized during CUSA, 124,711 (or 12 percent) did not undergo a full fingerprint comparison because the FBI deemed their fingerprint cards "unclassifiable."282 In addition, the FBI had no record of ever receiving a fingerprint card for 61,366 persons naturalized during CUSA. Because the FBI does not keep track of cards rejected for masthead errors, it is impossible to tell with certainty why no records were found for these 61,366 naturalized persons. However, an applicant whose card was submitted to the FBI but rejected because of missing or incomplete masthead data would leave no trace in the FBI billing records.283

      1. INS' lack of a policy requiring resubmission of rejected and unclassifiable fingerprint cards

In the years before CUSA, INS Headquarters did not recognize or emphasize the importance of resubmitting rejected fingerprint cards. Neither the Operations Instructions nor other official sources described what the Field should do upon receiving a fingerprint card rejected by the FBI because of masthead errors. Prior to 1982, applicants whose cards had been deemed "unclassifiable" were required to submit new fingerprint cards that INS sent to the FBI. However, INS abolished this policy at the time it adopted the presumptive policy.

        (1) Instructions concerning unclassifiable cards before 1994

When INS adopted the presumptive policy in 1982 and directed the FBI to destroy the fingerprint cards it submitted for which there was no match after a full fingerprint comparison, INS also asked the FBI to destroy unclassifiable fingerprint cards. Because these procedures reduced the clerical burden on INS of having to process unclassifiable cards returned by the FBI (it would reduce the amount of incoming mail, clerks would not have to find the relating applicant file or prepare a letter to send to the applicant), this, like the presumptive period, was considered by INS to be a "system improvement." This request made of the FBI suggests that as early as 1982 INS had begun to underestimate the importance of a criminal history check based on a full fingerprint comparison for every naturalization applicant. Rather than expend the clerical resources required to resubmit a new card, INS chose to risk not having a full fingerprint check completed for applicants whose cards were deemed unclassifiable and returned by the FBI.

By 1990, however, when the FBI began to charge agencies like INS for fingerprint checks, including cards they found to be "unclassifiable," the FBI had resumed returning unclassifiable cards to INS. As a result, INS would know if it had been charged for the fingerprint check but had not received the benefit of a full fingerprint comparison. There is no indication, however, that INS then began requiring district offices to resubmit rejected fingerprint cards in every instance. INS Headquarters' only written policy guidance to the Field in this respect was Operations Instruction 105.10, discussed above, which instructed that unclassifiable cards be routed and attached to the applicant's file, but it did not explicitly require resubmission of a fingerprint card.284 Ironically, although INS had changed its fingerprint processing procedures in 1982 in order to preserve INS clerical resources, INS was wasting money by not resubmitting new cards for unclassifiable ones once the FBI began charging for fingerprint checks.

        (2) Instructions to the Field on how to save money in the submission of fingerprint cards

The FBI's decision to charge contributing agencies for their fingerprint card submissions prompted INS to address one aspect of the process. Beginning in 1990, the Field was exhorted to "minimize the number of errors on fingerprint cards"and "to upgrade fingerprint card submissions to the FBI" because the FBI was charging not only for each submission, but also for each resubmission unless proper procedures were followed. The memoranda reminded the Field that in order for INS to avoid being charged for the resubmission of an unclassifiable card, the rejected, unclassifiable card had to be stapled to the new card before being submitted to the FBI. The memoranda did not describe in what instances cards should be resubmitted, but simply instructed the Field to resubmit in accordance with "existing procedures." The earliest of these memoranda, one from September 1990 that is referenced in later directives, advised the Field to look to specific regulations and Operations Instructions for guidance with respect to resubmitting fingerprint cards. As discussed above, however, none of these regulations or Instructions required that a rejected card be resubmitted to the FBI.

The record thus shows that when INS Headquarters did provide guidance to the Field about the processing of rejected fingerprint cards, it was in the name of saving money. There was no reminder about the importance of ensuring that each applicant's fingerprints be checked or that the adjudication of cases be postponed in order to permit sufficient time to resubmit fingerprint cards. Instead, the emphasis was again on the "cost-effectiveness" of INS' fingerprint processing. As the Field's naturalization workload increased in 1993 and early 1994, there was no indication that Headquarters wanted the Field to pay particular attention to the processing of these record checks for naturalization applicants.

        (3) The discretionary resubmission of rejected cards

In March 1994, prompted by an OIG inspection of fingerprint processing procedures at INS that was published a month earlier, INS Headquarters addressed the processing of fingerprint cards in connection with benefit applications. In a memorandum to the Field dated March 17, 1994, James Puleo, then one of two Executive Associate Commissioners, noted that it was important to route information about the applicant's criminal history or about his or her "rejected unclassifiable" card to the naturalization adjudicator. His memorandum said, "[p]rocedures for attaching arrest reports and rejected unclassifiable fingerprints to related files and flagging applications to the adjudicator's attention must be ensured." However, although INS' Operations Instruction had left room for the adjudicator to infer that the appropriate response to seeing an "unclassifiable" card in an applicant file was to request a new card from the applicant, Puleo's March 1994 memorandum made explicit that resubmission was not required in every case. It noted that in "those cases where fingerprints are found unclassifiable," District Directors were "to exercise discretion in requiring the submission of new fingerprints and delaying adjudication of the application pending results." No criteria were offered to describe how this discretion was to be exercised.

Our investigation found that the March 1994 memorandum is the only policy statement, other than the Operations Instruction, that INS Headquarters offered the Field about the processing of rejected fingerprint cards before INS launched the CUSA program in FY 1996. As such, it was deficient in many respects. First and most obviously, the memorandum sanctioned a decision not to resubmit rejected unclassifiable cards. Second, although it noted that resubmission of unclassifiable cards was in the discretion of the District Director, by failing to address the second type of rejected card—the masthead reject—and also referring to "rejected [emphasis added] unclassifiable" cards, the memorandum was susceptible to interpretation that resubmission of all rejected cards fell within the District Director's discretion. Although many INS employees told OIG that they had understood the difference between the types of FBI rejections, as discussed below, a significant number reported that they did not, and these employees would have believed that resubmission of any card rejected by the FBI was discretionary.

The other significant flaw in the March 1994 memorandum is that it assigned to District Directors a matter about which there could be no meaningful exercise of even-handed discretion.285 To determine that an applicant need not undergo a fingerprint check was the equivalent of deciding that some applicants appeared not to warrant further investigation while others did. It is difficult to determine what factors would trigger a decision to investigate a particular applicant further except factors based on stereotypes about an applicant's age, socio-economic status, country of origin, and the like. Not surprisingly, we did not find evidence in the Key City Districts of any local guidance detailing how to determine whether resubmission was warranted. In two districts, we found that the exercise of this discretion was delegated to the individual adjudicator. Under such policies, an applicant who appeared to the adjudicator not to be the type of person who would have a criminal record would not have his or her record checked by the FBI. An applicant who fit the adjudicator's idea of a criminal, however, would likely suffer a delay not experienced by other applicants because his or her fingerprint card would be resubmitted to the FBI.286

The March memorandum had a final weakness worth noting here. It did not require, in those instances where the District Director resubmitted a fingerprint card to the FBI, that the naturalization adjudication be delayed until the FBI had a chance to check the second fingerprint card. Like the decision to resubmit, the INS policy left the decision concerning "delaying the adjudication pending the results" of the FBI review to the District Director's discretion. Of course, to resubmit a fingerprint card but not also wait to receive the results before adjudicating the case essentially rendered the resubmission meaningless. We found that most districts at least respected a second presumptive period that began when the adjudicator sent a new card to the FBI. However, the lack of any specific requirement to honor such a delay resulted in a haphazard application of it in the Field. Instead of requiring the applicant to return to complete the interview once the presumptive period had again run, some districts simply allowed the presumptive period to be tolled during the time between the interview and the ceremony date, hoping to receive any new information from the FBI before the applicant was sworn in. Furthermore, as production pressures increased during CUSA, districts had even less incentive to engage in a discretionary exercise that would slow down rather than speed up the completion of a case.

By 1995, then, the guidance that INS Headquarters supplied the Field concerning fingerprint processing was much like the guidance provided on permanent file policy as discussed in the previous chapter. On the surface, the policy statements issued by Headquarters gave the appearance that it wanted to ensure uniform standards in naturalization processing. In reality, however, INS Headquarters failed to provide any concrete guidance. In the absence of leadership about such details, especially as the new CUSA initiative and its production demands moved to center stage, the Field predictably inferred that taking the time to process each applicant's fingerprint card thoroughly, like taking the time to review the applicant's permanent file, was not a priority to INS Headquarters. As a result, to varying degrees it was not a priority in the Field.

      1. Key City District practices in regard to processing fingerprint cards rejected by the FBI

As noted above, the practices in the Key City Districts relating to rejected fingerprint cards before and during the CUSA program (until the changes brought by Direct Mail and the FCCC, discussed below) reflected the absence of any uniform policy requiring resubmission. The evidence from two Key City Districts, New York and San Francisco, showed comparatively more diligence in processing fingerprint cards than did evidence from the other three, and we discuss their practices below. No Key City District, however, had a consistent policy of resubmitting every type of fingerprint card rejected by the FBI.

In each Key City District, we found evidence of confusion between what managers believed was controlling policy and what employees described as actual practice regarding the processing of fingerprint cards. In this regard, the failure of communication about fingerprint processing within the Districts mirrored the failure of communication between Headquarters and the Field.

        (1) Districts that made some efforts to resubmit rejected fingerprint cards

        1. New York District

Witnesses in the New York District reported a more cautious policy than that articulated in Puleo's March 1994 memorandum. Most said that a second fingerprint card was submitted to the FBI when an applicant's first fingerprint card was rejected because of either a masthead error or its unclassifiability.287 Despite such a policy, the practice in New York resulted in rejected fingerprint cards not being resubmitted in every instance.

In the early 1990s, the New York District instituted a policy of requiring applicants to submit two fingerprint cards as a means to address the problems associated with rejected fingerprint cards. The rationale was that if a card was rejected, resubmission would be easier if another fingerprint card was readily available in the applicant's file.288 Although this policy did make resubmission less burdensome for clerical staff in the naturalization section who, we found, were generally aware of the District's policy to resubmit rejected and unclassifiable fingerprint cards, it was far from foolproof.

First, as reported by then-ADDE Mary Ann Gantner to INS Headquarters in response to the March 1994 memorandum,289 if the applicant originally submitted only one fingerprint card and that card was rejected by the FBI as "illegible," the New York District did not obtain a second set of fingerprints from the applicant. In addition, naturalization clerks were not the only clerical staff to handle rejected fingerprint cards, and other clerical staff were not necessarily familiar with the local practice concerning use of a second fingerprint card. Clerks in the District's main file room, where all District mail was received, processed rejected fingerprint cards when the computer systems indicated that the A-file was in the main file room or in another FCO.290 These clerks placed the fingerprint card in the file, and it was up to the adjudicating officer to decide whether to submit a new fingerprint card at the interview.291

With respect to rejected fingerprint cards associated with files in the naturalization section,naturalization managers and experienced clerks told the OIG that clerks obtained the related files and submitted the second fingerprint card to the FBI along with the rejected fingerprint card for both masthead rejects and unclassifiables.292 Clerks also indicated that they sometimes sent letters to applicants requesting additional information to complete the fingerprint card. Naturalization clerks did not delay processing of these cases, but date-stamped the N-400 a second time when the fingerprint card was resubmitted.

While the policy of requiring applicants to submit two fingerprint cards may have improved the likelihood that the second card would be submitted if the first card was rejected, it did not necessarily improve the chances that the second card could be reviewed successfully by the FBI. In fact, because fingerprint cards generally were obtained from the same source, and since clerks told the OIG that they submitted, in the first instance, the card that appeared to be of superior quality, submitting the second card likely meant that the FBI would encounter the same mistakes and reject the card again. Fingerprint cards rejected a second time were placed in the applicant's file and brought to the attention of an SDAO. The adjudicating officer then decided whether additional fingerprints would be requested and resubmitted.

While the New York District policy was to resubmit all rejected fingerprint cards, witnesses acknowledged that sufficient clerical resources were not always devoted to processing rejected cards. In the main file room, for example, processing responses from the FBI was secondary to handling applications and petitions that had fees attached and required immediate attention. According to Section Chief Rose Chapman, sometimes it was necessary to assign DAOs to assist the clerical staff with processing rejected fingerprint cards because of accumulating backlogs.

        1. San Francisco District

In the San Francisco District, we found no consistent understanding of any policy in regard to the resubmission of rejected fingerprint cards. However, the evidence indicates that offices within the District made efforts to varying degrees to resubmit cards rejected for "masthead" errors. This was a marked improvement over what an OIG inspection team had found in 1993 when the San Francisco District was discarding rejected cards.

We found that the San Francisco District did not resubmit every rejected card because they followed Headquarters' advice that permitted "discretion" in the resubmission of unclassifiable cards. Where there was confusion concerning specifically why the FBI rejected a particular fingerprint card, witnesses for the most part told the OIG that they adopted the more cautious rule and submitted the cards to the FBI.

Although witnesses consistently reported that they reviewed cards before sending them to the FBI in an attempt to prevent rejections for masthead errors, we found little consistency throughout the San Francisco District with respect to other procedures followed when handling fingerprint cards rejected by the FBI. Three naturalization supervisors and one clerical employee we interviewed (the latter individual later became her sub-office's "fingerprint liaison" to the FCCC) were unaware of the different types of rejects from the FBI and believed that the decision to resubmit a rejected card lay with the adjudications officer at the interview. Other witnesses, however, knew the difference between the two types of FBI rejections and reported that the District's clerical staff would ensure that "masthead rejects" were replaced or resubmitted with complete information, while "unclassifiable" cards would be placed in the applicant's file and new fingerprints obtained at the discretion of the officer. Because one of the witnesses who described this latter practice was the clerical supervisor responsible for administering the resubmission of masthead rejects, we assume that her report is more reliable in regard to such practices than reports from other staff who worked at greater distance from the fingerprint process. Finally, even though we found many other supervisors and adjudicators who did not know the difference between the two types of rejected cards, they all asserted that the District's policy was to resubmit a new card in every instance.

        (2) Districts that failed to resubmit rejected cards

As vulnerable a policy as it may have been to leave the issue of resubmission of a rejected card to an individual adjudicator's discretion, such a policy at least gave the experienced and diligent officer the opportunity to request a second fingerprint check. In three other districts, officers did not have that opportunity because rejected fingerprint cards were rarely routed to an applicant's file and instead were routinely destroyed.

        1. Chicago District

In response to the March 1994 memorandum, then-District Director A.D. Moyer reported to Headquarters that the Chicago District routed unclassifiable fingerprint cards to either the applicant's file or to a central location where they were filed by A-number along with rap sheets. According to Moyer, these documents were matched to the applicant's file before interview. He wrote that "[a]t interview, only case[s] with unclassifiable prints are given a new fingerprint card." Although we presume that the District Director meant that only applicants whose fingerprint cards had been rejected by the FBI were given new cards at interview (and not all applicants), the memorandum did not specifically address how the District handled cases in which the fingerprint card had been rejected by the FBI for masthead errors.

The process as described by Moyer, although deficient, at least suggested that adjudicators had the opportunity to request a second fingerprint check of some applicants whose cards had been rejected by the FBI. The evidence shows, however, that the policy described by Moyer was not the practice in the Chicago District. First, just a few months before the District Director made this representation to INS Headquarters about the handling of rejected fingerprint cards in Chicago, the OIG inspection had revealed that unclassifiable fingerprint cards were being routed to the basement of the Chicago District Office building and were burned. We found no indication that this practice changed before the District Director made his report or anytime before CUSA.

Second, the SDAO in charge of naturalization at the time recalled developing a new filing system to route rap sheets to a centralized file (as described by Moyer in connection with rejected fingerprint cards). However, this filing system did not include rejected fingerprint cards. In addition, the SDAO told the OIG that she was not aware of any practice of resubmitting rejected fingerprint cards. Interviews by the OIG of the Chicago District personnel revealed that the supervisory applications clerk and two long-term clerks also were not familiar with any procedure to resubmit rejected fingerprint cards.

        1. Miami District

The response from the Miami District to the March 1994 memorandum offered a realistic picture of the critical state of their fingerprint card processing efforts. In April 1994, the Assistant District Director for Examinations wrote to INS Headquarters that rejected fingerprints were "an impossible situation for an office the size of Miami," and that the Miami District did not resubmit fingerprint cards because it lacked sufficient clerical resources to do so.293 While the memorandum did not distinguish between masthead rejects and unclassifiable cards, interviews with Miami District officials confirmed that masthead rejects were rarely corrected and resubmitted. While several clerks told the OIG that they corrected and resubmitted masthead rejects, they conceded that this was not a routine part of their clerical workload. Fingerprint cards deemed unclassifiable were shredded without any further attempts to obtain a full FBI fingerprint check.

        1. Los Angeles District

The Los Angeles District Director Rogers informed INS Headquarters in April 1994 that the policy in his District was to route unclassifiable fingerprint cards to the applicant's file. The decision whether to resubmit a card was left to "the discretion of the adjudicator." Rogers did not specifically address in his response cards that were returned by the FBI as masthead rejects. Our investigation shows that contrary to Rogers' claims, unclassifiable fingerprint cards were destroyed and fingerprint cards rejected because of masthead errors were rarely resubmitted.

Rogers told the OIG that in April 1994, he did not know the difference between a card that had been "rejected" by the FBI and one deemed "unclassifiable," but said he believed that in both cases Los Angeles policy had been to route the card to the file for action by the officer at interview. Section Chief (and later Deputy Assistant Director) Donald Neufeld agreed that the policy articulated in Rogers' April 1994 memorandum regarding the handling of unclassifiable prints was the one with which he was familiar, though he also noted that from the time he became Section Chief later that year through the CUSA program, he had no personal knowledge of whether any fingerprint cards returned by the FBI were, in fact, routed to applicants' files. Assistant District Director for Adjudications Jane Arellano confirmed that, as stated in the April 1994 memo, District policy provided that unclassifiable records were to be routed to the file and new fingerprints requested in the discretion of the officer at the naturalization interview. When asked under what circumstances an officer was to request a new fingerprint card, ADDA Arellano simply stated that a request would be made when something in the file or remarks made during the interview suggested that there may be some criminal history that needed to be explored. However, despite these three managers' understandings of District policy, the practice in Los Angeles was clearly as deficient as that in the Miami District.

We found that employees who worked most closely with the fingerprint records understood the difference between cards "rejected" by the FBI and those determined to be unclassifiable. According to the employee who had primary responsibility for processing these cards both before and during CUSA, with masthead errors he would research the computer records to supply missing biographical data and resubmit the card to the FBI or, if the missing information was not in INS records, take steps to obtain a new card from the applicant. From at least 1994 until the summer of 1996, one officer with some clerical assistance essentially carried out this job as a collateral duty. The officer told the OIG that it was extremely hard for him to keep pace with the work given other matters—such as the routing of criminal history reports—that were of a higher priority.

However, in the Los Angeles as in Miami District, according to employees with firsthand knowledge of fingerprint card processing, unclassifiable cards returned by the FBI were destroyed, not routed to the file, if the FBI indicated that the applicant had no record based on a name-check. Similarly, rap sheets showing only unprosecuted immigration violations were discarded.

Los Angeles managers denied knowing that this material from the FBI was being discarded, either before or during CUSA, despite employees' reporting to the OIG that it was standard practice. However, even if managers did not specifically know of this practice, it was consistent with a general approach that prioritized the records that showed the most serious or disqualifying incidents over other records. Where the volume of FBI responses was great, ADDA Arellano told the OIG that Los Angeles employees "didn't get too worried" about matters such as criminal history reports that showed only immigration violations. She conceded that these issues might "pile up" because of the staff's concentration on records that reflected more serious crimes. In the high-volume, large backlog environment that existed in the Los Angeles District before CUSA, it was an unremarkable step for "discretionary" material like unclassifiable cards and records of administrative violations to go from being less important records to unimportant ones.

    1. Failure to ensure that rap sheets were available for review in conjunction with the naturalization interview

The naturalization adjudication system presumed that for those applicants whose FBI fingerprint checks revealed a criminal record, the criminal history report or rap sheet would be reviewed during the naturalization interview. The rap sheet was a tool to be used by the adjudicator at the naturalization interview to help determine whether evidence existed that would have an impact on the evaluation of the applicant's "good moral character."

The most common use of an applicant's rap sheet was when it indicated that the applicant had been convicted of a disqualifying offense. This information would lead the adjudicator to deny, rather than approve, the application. But even without disqualifying information, the rap sheet could prompt certain avenues of inquiry at interview designed to confirm the applicant's veracity in filling out the N-400, as well as to probe the credibility of answers provided during the interview.294

Even before CUSA, however, INS had begun to treat an applicant's rap sheet as the only source for determining whether the applicant was necessarily precluded from naturalizing because he or she had been convicted of a disqualifying offense. This narrow view of the rap sheet's role in naturalization adjudication was similar to the narrow view of the entire "good moral character" evaluation that had evolved in the Field by the early 1990s, as discussed in our chapter on interviews and adjudications. Apart from this changing mindset, the sheer volume of paperwork combined with a lack of adequate clerical resources often delayed the districts' processing of incoming rap sheets from the FBI. Taken together, these two situations led some INS districts to devise alternative methods of reviewing applicant rap sheets other than placing the record in the applicant's file before the interview date. This, in turn, often led to a failure to review the rap sheets before naturalization.

For example, as discussed below, in the Chicago District rap sheets were organized in boxes that adjudicators were supposed to check before conducting naturalization interviews. We found evidence that as production pressure mounted, adjudicators did not thoroughly review these applicants' criminal histories. Los Angeles developed a stop-gap policy of reviewing applicants' rap sheets for disqualifying information before the naturalization ceremony and then "pulling off" ineligible candidates on the date of ceremony. This "pull-off" system collapsed in Los Angeles under the volume of work during CUSA.

The failure to review rap sheets before interview made naturalization adjudications more vulnerable in several ways. First, it eliminated any potential value the rap sheet could bring to learning more about the applicant's background and in checking his or her veracity. Second, by creating makeshift alternatives to the interfiling of the rap sheet in the applicant's file, the districts avoided the underlying problems—the clerical inefficiencies and the disregard for the appropriate presumptive period in scheduling interviews—that were preventing the rap sheets from being available to adjudicators. Finally, by allowing the naturalization case to proceed independent of checking the fingerprint card and reviewing the rap sheet, INS laid the foundation for naturalizing applicants without reviewing their criminal history reports.

      1. Chicago District

Although rap sheets were supposed to be placed in applicant files for use during naturalization interviews, we found that, as discussed in previous chapters, insufficient clerical resources and other factors, like accelerated production at off-site processing sites, compromised the District's file-handling procedures. Rap sheets that were not inter-filed were placed in a bin in a central filing area (which eventually grew to multiple boxes), where they were organized according to the last three digits of the A-number, or organized in "terminal digit order." DAOs were expected to check this central filing area for rap sheets on the morning that files were received for interviews scheduled that day in the office ("office cases") or the night before, when files were received for interviews that were to be conducted at CBO facilities ("outreach cases"). This practice was followed as early as 1993 when the Chicago District practices were reviewed by the OIG's Inspections Division.

However, even at the time of the OIG inspection it was clear that this shortcut for making rap sheets available to interviewers was not working efficiently. Adjudicators reported that they rarely found arrest reports in the boxes, and OIG inspectors found that records in the central filing area were mostly fingerprint checks that had been completed more than a year before.

Although the Chicago District made some effort between the 1994 OIG report and CUSA to improve this "review box" system for rap sheets, by September 1996 more than 2,000 rap sheets were filed in the boxes, unreviewed. INS' review of these records, after congressional hearings about CUSA in September 1996, revealed that more than 1,000 of the rap sheets pertained to applicants who had already naturalized.295

      1. Los Angeles

From 1994 until August 1996, the Los Angeles District assigned one DAO to review FBI responses to fingerprint checks. Incoming rap sheets that were reviewed before the applicant's interview would be placed in the applicant's file, most often a temporary file housed at the District Office.296

However, because the Los Angeles District sent fingerprint cards to the FBI only two months before the date on which the interview was scheduled (as discussed above), according to DADDA Neufeld, rap sheets often failed to arrive prior to the interview and consequently were reviewed after the interview. The DAOs who had been assigned to this rap sheet review told the OIG, however, that not many records were reviewed after interview, but they did note that Los Angeles had developed a system for the post-interview review of the applicant's criminal history. If the applicant had already been interviewed by the time the rap sheet arrived, the DAO assigned to rap sheet review would review the information to determine if it contained any disqualifying or unrevealed criminal event. If it did, the DAO would add the applicant's name to a list of those who would be "pulled-off" at the naturalization ceremony.

Both the Section Chief and the DAO assigned to the rap sheet review project recalled that even before CUSA, the Los Angeles District would sometimes receive an applicant's rap sheet only after the applicant had naturalized. According to the DAO assigned to review rap sheets, if a rap sheet arrived after the ceremony, the file would be pulled to determine if the applicant should have been denied citizenship. If the applicant clearly should not have been naturalized, the file and the rap sheet would be forwarded to the litigation unit of the District Office for possible revocation procedures, then to Western Region, and then to the General Counsel's Office at INS Headquarters.

During CUSA, the Los Angeles District relied to a greater extent on this "pull-off" procedure. As the District opened new interview sites and files were dispersed throughout the District, the number of responses from the FBI also grew. Our investigation found that few applicant rap sheets were reviewed before the applicant's naturalization interview. Furthermore, the "pull-off" system was often unable to prevent the naturalization of applicants whose rap sheets showed disqualifying criminal histories.297

  1. INS' failure to respond to outside recommendations to improve the fingerprint process: the OIG and GAO reports of 1994

    1. Introduction

Unlike the vulnerabilities in other aspects of INS' naturalization processing procedures as described elsewhere in this report, the weaknesses of its fingerprint check procedures had been clearly identified by officials both inside and outside INS well before CUSA. By the end of 1994, both the OIG and the GAO had studied and reported on INS' fingerprint processing procedures—specifically in connection with applications for benefits—and had found significant problems.298 The problems included the weaknesses we have already addressed in this chapter—the reliance on a presumptive policy, the failure to resubmit cards rejected by the FBI, and the failure to ensure that rap sheets were made available to adjudicators in a timely manner. Both agencies made recommendations to INS to improve their fingerprint checking procedures; none was implemented before INS launched CUSA, however, with the exception of INS Headquarters' distribution of a directive to remind the Field to comply with fingerprint procedures.

As Chairman Lamar Smith of the Subcommittee on Immigration and Claims of the House Judiciary Committee pointed out in his statement to open the March 5, 1997, Joint Hearing, the GAO had "made specific recommendations and the INS agreed to implement them. Had they been implemented," he continued, "we would not be here today." Chairman Smith also specifically questioned Commissioner Meissner about why INS had failed to adequately respond to the 1994 OIG and GAO reports to improve fingerprint processing procedures before implementing CUSA. As he said to Commissioner Meissner:

You have said, Commissioner, that the chief mistake that you made was to, "apply outmoded practices to urgent and overwhelming demands." But those are the same outmoded practices that were criticized by the IG in 1994, and criticized by the GAO in 1994. It was a mistake to continue to apply what you had to be on notice were bad practices that were going to result in the mistakes that occurred . . .

Do you not think that you should have taken some steps after the GAO report of 1994 to fix the problem before you undertook a massive program to naturalize three and four times the number that were studied under the GAO report?

In her response, Commissioner Meissner said that INS had taken "a series of steps" that they "believed were responsive," but that "under the crush of the workload . . . they [were] not adequate to the task."

In the discussion that follows we describe the OIG and GAO reports and outline the steps INS failed to take in response to each. The record shows that, as alleged, INS did not fix the problems identified in either report before implementing CUSA.

The evidence confirms that even though INS took a series of steps in response to the OIG and GAO reports, they were inadequate to respond to the widespread problems identified by those reports. INS began an in-house discussion of potential improvements to the fingerprint process and set up a task force whose job was to provide short- and long-term recommendations. INS also held regular meetings to discuss its progress on responding to the GAO report. However, none of these plans or discussions resulted in any concrete change to fingerprint processing before INS implemented CUSA. Moreover, INS' principal solution to the deficiency created by the presumptive policy was an anticipated automated processing system that far from being either designed or operable.

Despite these somewhat urgent reports from the GAO and the OIG about the vulnerabilities in INS' fingerprint processing procedures, INS was slow to react. The record indicates, as discussed in the following pages, that INS officials' belief that fingerprint checks were not the sine qua non of naturalization adjudication contributed to their failure to take more substantive action in the wake of these reports. The Commissioner and other INS officials expressed the historical INS opinion that naturalization applicants were by and large a population unlikely to have a disqualifying criminal history. INS officials also believed that alternate methods were in place to learn about an applicant's criminal history, including the naturalization interview, and thus reliance on the costly FBI check was not the only option.

The small number of applicant criminal records that INS officials believed were detected by the process of submitting hundreds of thousands of naturalization applicant fingerprint cards to the FBI299 led INS to conclude that repairing fingerprint processing procedures in the short-term, before INS became the fully-automated INS of the future, was not worth the resources it would cost. By the summer of 1995, plans for CUSA took center stage and all efforts were directed toward increasing the number of interviews and increasing processing rates, production goals that were inimical to ensuring that each applicant's background was thoroughly reviewed. Increased production exacerbated some of the very conditions that had contributed to vulnerable fingerprint processing procedures in the first instance, and during CUSA INS' procedures broke under the strain.

It was not until the summer of 1996 that INS began to implement any changes in its fingerprint processing procedures as a result of the 1994 GAO and OIG reports. By then, as discussed in the next section of this chapter, the changes were too late and too poorly implemented to have their intended ameliorative effect. INS' superficial approach to fingerprint processing did eventually begin to change in the autumn of 1996 in reaction to intense scrutiny from Congress and the Attorney General. However, by that time, thousands more applicants had naturalized without having had fingerprint checks as part of the naturalization adjudication.

    1. The 1994 OIG inspection report and recommendations

      1. The inspection and findings

From July 1993 through October 1993, the OIG's Inspections Division examined INS' fingerprint procedures for naturalization, adjustment of status, and other benefit applications. Its February 1994 inspection report, "Alien Fingerprint Requirements in the Immigration and Naturalization Service," assessed "the necessity and effectiveness of conducting fingerprint checks" on applicants for INS benefits.

The OIG's first conclusion was that fingerprint checks were indeed necessary to the benefit application process. Citing their interviews of adjudicators and staff in INS' Office of Enforcement, the inspection team concluded that the fingerprint check was the only practical means available to positively identify an applicant's criminal history. The report noted that interviews with INS adjudicators and the team's own review of files showed that an objective criminal history check is necessary because applicants with criminal history often fail to disclose that history during their interview.

The OIG report then turned to the effectiveness of INS' fingerprint process and discovered significant weaknesses at every stage:

In response to the OIG's findings, INS noted its concerns about the time and effort involved in resubmitting applicant fingerprint cards after rejection by the FBI. INS emphasized that resubmission was labor-intensive, requiring a letter to the applicant and a new submission to the FBI that included both the new and rejected cards. INS also pointed out that interviews had to be postponed when applicants' cards needed more time to be processed.300 In response, the OIG reiterated that if INS exercised greater control over the entities taking fingerprints, this would increase the quality of the initial submissions and thus decrease the number of cards rejected by the FBI, thereby reducing INS' burden to resubmit the fingerprint cards.

      1. The recommendations

The OIG report recommended that INS gain control over the outside organizations that were providing fingerprinting services and develop procedures to verify that fingerprints submitted by applicants were their own. This recommendation addressed both decreasing the potential for fraud in the fingerprinting of the applicants and decreasing the number of unclassifiable fingerprints. In addition, the report recommended that INS instruct the district offices to ensure that fingerprint cards were mailed promptly to the FBI and that rap sheets were placed in the A-files before adjudication.

The OIG report discussed various options considered internally by INS to limit the number of fingerprint checks it was then requesting of the FBI. One such strategy discussed in a September 1993 INS report argued in favor of reducing fingerprint costs by limiting the number of applicants checked. The report presented various options for doing so, including forwarding to the FBI fingerprint cards only for applicants who fit a profile based on arrest statistics for age and gender. The submission of fingerprint cards to the FBI, the report suggested, could be discretionary, based on these criteria and generally informed by "the officer's impression of the alien during an interview."

James Puleo, then one of two Executive Associate Commissioners offered another cost-saving option in October 1993. Puleo recommended that INS eliminate the fingerprint check for most benefit applications except in limited circumstances in order to bridge a projected $20 million shortage in the Examinations Fee Account. The proposal, which was strongly resisted by INS' Office of Enforcement as "effectively subvert[ing] the integrity of the immigration benefit adjudication process," was rejected by then-Acting Commissioner Chris Sale. INS later resurrected this same idea in March 1994, again in the name of coping with a shortage of funds in the Examinations Fee Account (see discussion below).

While the OIG report concluded that a fingerprint check was still needed for benefit applications, the report noted that "changes may be in order to the current policy of conducting checks for all applicants."

      1. INS' response to the OIG report
        (1) INS concurs only in part that fingerprint checks by the FBI are necessary

INS' response to the OIG report clearly conveyed its reluctance to seriously overhaul its fingerprint processing procedures. As a preliminary matter, INS concurred only in part with the OIG determination that "fingerprint checks are necessary to detect criminal records which may make an applicant ineligible for benefits." According to INS, there were "many supplemental or alternative sources" of information of criminal activity available to INS in addition to the FBI-provided rap sheets.

Consistent with its position that fingerprint checks by the FBI were not necessary in all cases, INS also asserted that any changes made to address the fingerprint processing problem had to be balanced against the resources necessary to improve the process. According to INS' response, the weaknesses in the process were attributable to "an ongoing conflict between a changing, expanding workload and the inherent difficulties of resource allocation." In addition, INS indicated that it was reluctant to invest in revamping procedures that had been established years earlier when the process was entirely manual because INS anticipated moving toward a more automated process.

INS' claim that adjudicators had alternate sources available that would alert them to an applicant's potential criminal activity was technically accurate. INS investigations divisions had access to a number of automated criminal history databases, some of which had information about an applicant's local criminal activity that might be relevant and not reflected in FBI records. In addition, adjudicators also questioned applicants about their criminal histories at the interview. Finally, naturalization applicants presumably had undergone a criminal background check at the time they adjusted their status to permanent resident.

However, each of these alternative methods to check on an applicant's criminal history also posed serious problems. Adjudicators did not uniformly have access to the same databases as their criminal investigations counterparts, and the information was not authorized for use in conjunction with benefits applications. Also, the absence of an FBI record-check prevented the adjudicator from confirming an applicant's denial of a criminal history.301 Finally, even if a naturalization applicant had undergone the adjustment of status process, that record would necessarily be at least three or five years old and may have suffered from the same flaws inherent in INS' naturalization processing. The FBI database, on the other hand, is the most comprehensive nationwide source of criminal history information.302 In addition, FBI fingerprint checks were the only Service-wide method that INS had in place for checking applicant's criminal history records.

        (2) INS agreed that fingerprint processing required improvement but failed to take effective action

In response to the OIG report, INS conceded that its fingerprint processing procedures were in need of some attention. Apart from its partial concurrence with the finding that FBI fingerprint checks were necessary, INS officially agreed with the report's other findings and recommendations. In response to the recommendation that INS institute procedures to verify applicants' fingerprints, INS agreed to issue a policy on "fingerprint execution control" no later than March 15, 1994. In response to the recommendation that INS instruct district offices to mail fingerprint cards and file rap sheets promptly, INS agreed to direct district offices to review their fingerprint procedures and agreed to monitor the Field's compliance with existing procedures through site visits. However, we found that INS failed to follow through in any substantive way with either promised course of action.

By mid-April 1994, INS had done little to effect any real change in response to the fingerprint processing problems raised in the report. The policy on "fingerprint execution control" had not been drafted and INS Headquarters had failed to conduct any site visits to field offices. INS' only action in response to the OIG's recommendations was to issue a memorandum to the Field dated March 17, 1994, describing INS policy concerning "rejected unclassifiable prints" that was discussed in the previous section of this chapter. That memorandum also directed the Field to review their fingerprint procedures to ensure compliance with the policy and to submit a plan of action for addressing the findings in the OIG report. It also advised the Field of upcoming on-site inspections.

The March 17 memorandum was followed by two more from Headquarters, one in March and another in June, both soliciting information about fingerprint processing practices in the Field. No on-site inspections were conducted.

        1. Three fingerprint processing memoranda

As discussed above, the March 17 memorandum from Operations to the Field was ineffective. While it technically complied with INS' promise to the OIG to remind the Field to timely process fingerprint cards and criminal history reports, its reference to existing policy and procedures was of little value because those procedures were not clearly articulated in either the Operations Instructions or in other directives from Headquarters.

Information Headquarters received from the Field in response to the March 17 memorandum clearly showed that practices in the Field were far from ideal. In addition to the responses from the Key City Districts discussed previously (such as Miami's assertion that it simply could not process the rejected fingerprint cards), Headquarters received clear confirmation that what the OIG had found in the four districts it visited was happening throughout the country. New York, Dallas, and Detroit Districts all reported that rap sheets, or "kickbacks" as they were called in some offices, were sometimes received from the FBI after applicants had been naturalized.

While most of the responses from the Field indicated that the district offices and sub-offices were complying with INS fingerprint procedures, we found that some of these responses warranted a closer look. The Chicago and Boston offices, for example, both of which were cited for specific problems in the OIG inspection, reported that all procedures were being followed and that no changes were required. Headquarters did not follow up on the information provided by the Field. The only further action Headquarters took was to send two additional memoranda to the Field.

The first of the follow-up memoranda, dated March 29, asked the Field to review its timeliness in submitting fingerprint cards to the FBI. Accompanying the memorandum, Headquarters sent district offices FBI records reflecting the approximate date fingerprint cards for applicants, identified by A-number, were received at the FBI. The Field was asked to compare its records indicating the date on which they received a response from the FBI.303 The memorandum did not require the Field to report its findings, and we found no evidence that INS Headquarters made any further inquiries to the Field.

The second memorandum, dated June 8, requested information concerning each district's fingerprint processing practices, including whether districts were ensuring that FBI responses were being placed in files "before a decision" was made.304) When all but two offices responded that they were getting rap sheets into the applicants' files in time for the adjudication, INS Headquarters took no action to reconcile the positive results of this survey with the more sobering responses to the March 17 memorandum. Indeed, although the March 17 memorandum advised of on-site inspections to ensure compliance with fingerprint processing procedures, none was conducted.305

The only actual review of procedures that INS undertook in response to the OIG report was to review OIG procedures. Donald Crocetti, then Acting Associate Commissioner, tasked INS' Director of Statistics with determining whether the conclusions in the OIG report were supported by the data and whether the OIG's file review accurately portrayed the impact that missing rap sheets would have made on the adjudications. In both instances, the OIG findings were corroborated.

        1. Headquarters officials acknowledged to OIG that INS' response to the inspection report lacked substance

Two INS officials then responsible for the naturalization program—and, accordingly, for INS' response to the OIG report—acknowledged to the OIG that INS was aware the three memoranda to the Field in the spring of 1994 would not improve the quality of INS' fingerprint processing. Lawrence Weinig, who held the position of Acting Associate Commissioner for Adjudications (the office was later called the "Office of Programs") at the time of the OIG Inspection, and Crocetti, who came to INS Headquarters in early 1994 as an Assistant Commissioner under Weinig and who had replaced Weinig by April 1994, told the OIG in conjunction with the current investigation that both INS Headquarters and the Field believed no real change would occur without additional resources.

Crocetti described INS' memoranda to the Field as "lip service." When the districts reported that they were in compliance with proper fingerprint procedures, Crocetti said that even though Headquarters knew that it was not true, they still failed to take action. Weinig said that INS knew that fingerprint processing procedures were "hemorrhaging all over the place." He characterized the INS response to the OIG inspection as merely a way for INS to cover itself against criticism.

The only lasting effect of the OIG report (and subsequent congressional attention, discussed below) on INS' approach to fingerprint processing was the formation of the "Fingerprint Enhancement Working Group" (FEWG) in June 1994. However, even this group's recommendations, which we discuss later in the chapter, were not timely implemented.

    1. INS suspends fingerprint checks in March 1994

Just as INS responded to the OIG report by saying that it would take steps to improve the fingerprint process by ensuring compliance with procedures in the Field, it simultaneously took action to do away with the fingerprint check altogether. Even though the decision was quickly rescinded in response to an immediate outcry from Congress, we examine it here because it illuminates the prevailing attitude at INS about fingerprint checks at the time.

In March 1994, INS was anticipating a revenue shortfall in the Examinations Fee Account of approximately $30 million, and specifically a $4.6 million shortfall in funds to pay for FBI fingerprint services. Consequently, INS Headquarters resurrected the previously proposed and rejected plan to eliminate most fingerprint checks in order to save money. This time, INS translated the idea into action. Despite the OIG report and despite the memorandum INS had sent to the Field less than two weeks earlier ostensibly encouraging fingerprint checks, INS issued a directive to the Field on March 28, 1994, eliminating the fingerprint check requirement for almost all benefit applications, including naturalization. Associate Commissioner Weinig for EAC Puleo signed the memorandum.306

The memorandum instructed the Field that effective April 1, "all INS offices [were] to cease submission of all FD-258 fingerprint cards to the FBI," except in a few specific cases. An INS memorandum to the FBI explaining the decision indicated that this was a permanent change in INS procedures intended to reduce INS' costs by millions of dollars each year. INS Headquarters explained to the Field that it was taking this action "reluctantly" and that the "impact" of the decision, presumably on the integrity of the adjudications process, would be mitigated by the fact that INS was still going to conduct bio-checks307 and by the fact that "the percentage of hits on FBI fingerprint checks is very low."

The decision to eliminate fingerprint checks was consistent with the reticence to invest in fingerprint checks while awaiting the eventual improvements technology would bring, as had been articulated in Commissioner Meissner's response to the OIG report. INS was again citing the dual themes of the low-risk population and the high cost of checking their records. The decision to eliminate fingerprint checks was about "avoid[ing] paying the FBI 30 bucks a head," Deputy Commissioner Sale later told the OIG. However, it was also about what returns INS would get for its money.308 On balance, INS decided that it was not worth the money to check the criminal history of every naturalization applicant, since the majority were perceived to have "clean records." Commissioner Meissner, conceding INS' mistake in eliminating fingerprint checks, told the OIG the decision in March 1994 to eliminate fingerprint checks for naturalization applicants was "a very good example of the agency's view about the value of the fingerprint check."

Opposition from outside INS to the decision to suspend fingerprint checks was immediate and vocal. After media accounts appeared criticizing INS for its decision, Senators Robert Byrd and Ernest Hollings wrote the Attorney General on April 14, 1994, to express their "dismay and serious concern" over the policy shift. They emphasized that a naturalization program that did not include fingerprint checks was inadequate and asserted that even a small number of resulting disqualification represented a vital part of the system's integrity. Senator Dianne Feinstein registered similar concerns and expressed great dismay in a separate letter sent to the Attorney General the same day. She rebutted INS' complaints of inadequate resources by pointing out that INS had not sought congressional help with its budget shortfall before taking this "dangerous" step.

In light of this congressional concern and pressed by officials at the Department of Justice to justify the policy change, INS reversed its position and indicated that the decision to eliminate the fingerprint check requirement had not been given adequate consideration. On April 19, 1994, less than three weeks after it took effect, EAC Puleo directed the Field to hold the new policy in abeyance pending a "more exhaustive review." The Attorney General later responded to the congressional letters by indicating that INS' decision to eliminate the fingerprint check requirement had been reversed and that INS would review the data in the OIG inspection report in an effort to obtain more information about the efficacy of the fingerprint process before making any future changes. Despite INS' memorandum to the FBI that clearly indicated that the decision to eliminate the fingerprint check was a permanent change, INS later provided the Attorney General with information that its decision had been a temporary, 6-month pilot program prompted by a revenue shortfall.

The congressional scrutiny around the incident served to bring the topic of fingerprints and, specifically, the findings of the 1994 OIG inspection report to Commissioner Meissner's attention. It was her concern about the issue that resulted in a series of memoranda and meetings of INS Headquarters officials and the formation of the Fingerprint Enhancement Working Group (FEWG) whose work is described below. Congressional concern about how INS was handling applicants' fingerprints also prompted a request by 15 Members of Congress to the General Accounting Office (GAO) to review INS fingerprinting procedures. The GAO report, issued in December 1994, is addressed below after our discussion of the FEWG.

Shortly before Commissioner Meissner was scheduled to testify before the House Appropriations Subcommittee on April 20, 1994, about the decision to eliminate the fingerprint check requirement (among other topics), she directed the managers of the Adjudications Program at Headquarters to undertake a complete review of the fingerprint process. She received a response from her staff of the same quality as the guidance that staff had been providing the Field.

EAC Puleo assured Commissioner Meissner in a memorandum dated April 19, 1994, that INS Headquarters had already begun to address deficiencies in the fingerprint process. He said staff were responding to the OIG recommendations by requesting input from the Field on improving the fingerprint process309 and by reviewing the cases identified in the OIG inspection report to assess the effect that the FBI record would have had on the adjudication. Commissioner Meissner then reported to Congress at the April 20 hearing that INS was studying the entire fingerprint process and would provide Congress with a full report.

    1. The recommendations of the Fingerprint Enhancement Working Group

In June 1994, Headquarters convened the Fingerprint Enhancement Working Group as promised by Commissioner Meissner in her April testimony. The working group had an impossible mandate: although INS had told the OIG that "resource allocation" was the central obstacle to fixing fingerprint processing procedures, the task given the FEWG was to "improv[e] the present benefits-related fingerprint process . . . without requiring the expenditure of additional resources." Within a month, in July 1994 the FEWG issued a report detailing several steps INS could take to improve its fingerprint processing procedures.

While some of the recommendations suggested improvements that could be made over the long-term, others could be implemented immediately with very few additional resources. During the next few months, INS Headquarters reviewed the recommendations, seeking input from more than 50 senior managers and field managers. However, INS did not adopt any of the changes recommended by the FEWG before implementing CUSA in August 1995.

In the discussion that follows we describe the recommendations made by the FEWG. We postpone our discussion of INS' response to those recommendations until after we have also discussed the GAO report that was published in December 1994, soon after the FEWG completed its work. Even though the recommendations that grew out of the FEWG would have been responsive to many of the concerns raised in the GAO report, INS still failed to implement any of them.

      1. The FEWG's long-term recommendations

The FEWG, composed of INS Headquarters managers, staff members, and advisory participants from the FBI and OIG, considered issues affecting the quality of the fingerprint process, some of which had been pointed out in the OIG inspection report and some of which were identified by the working group.310 These issues included INS' lack of control over outside entities that took applicants' fingerprints, the lack of controls to ensure that fingerprint cards were complete and legible, the lack of controls to ensure that masthead rejects were resubmitted, and the lack of controls to ensure that unclassifiable cards and rap sheets were filed prior to adjudication. The group also considered whether INS should continue to require fingerprint checks of all applicants.311

The group's long-term recommendations to improve the fingerprint process focused on technology not yet available to INS. Such technology would automate both the transfer of biographic and fingerprint information to the FBI as well as receipt of the FBI response. This was the goal to which INS had previously pointed, in its response to the OIG inspection report, when it expressed reluctance to continue investing in paper-based procedures that were not consistent with this view of INS' automated future.

In discussing this automated future, the FEWG noted in its report that the four service centers would play an important role since INS was already moving toward Direct Mail for all benefit application processing. As a "mid-term" recommendation that could be implemented before Direct Mail began for naturalization applications, the FEWG recommended that INS explore moving the responsibility for fingerprint processing from all district offices to the four service centers. While this proposal was not adopted for all districts, the service centers processed fingerprint cards for four of the Key City Districts under CUSA's Direct Mail program. However, the FEWG's recommendation that INS concentrate the process at few locations with computer support would later influence INS' creation of the Fingerprint Clearance Coordination Center (FCCC), discussed below.

      1. FEWG's short-term recommendations

In addition to their long-term recommendations, the FEWG recommended several immediate changes in INS' fingerprint processing procedures. Two of the proposals were steps that INS had already committed itself to take in responding to the OIG report issued seven months earlier: the clarification of fingerprint procedures and the monitoring of compliance in the Field. In addition, the group recommended that INS use FBI's automated billing records in the fingerprint process and that INS establish a program for certifying providers of fingerprinting services.

Of these four proposals, INS acted on the recommendations concerning on-site inspections and certification of fingerprinting service providers during CUSA. However, neither recommendation was implemented in a timely fashion or improved fingerprint processing during CUSA. The failure to follow the group's recommendation to use FBI automated billing records in the fingerprint-checking process has been shown by subsequent events to have been particularly unfortunate. Early use of the billing records might have led INS to realize before CUSA the multiple ways in which these records could enhance the integrity of its fingerprint processing procedures. INS did not adopt such a strategy until November 1996, and only after it had been ordered by the Attorney General to abandon its presumptive policy. We examine each of the working group's four major recommendations in turn.

        (1) Reviewing and clarifying current fingerprinting procedures

The FEWG made the general recommendation that current processing procedures should be reviewed, clarified, and reinforced. The group found that additional training was warranted because its review of selected files showed that it was difficult to ascertain when or whether certain steps in the processing of fingerprint records had been accomplished. This was a non-controversial and low-cost recommendation.

        (2) Monitoring of INS field offices' compliance with procedures

To increase accountability, the FEWG's report recommended a "site-audit process" in which INS Headquarters would conduct regular audits of selected field offices and in which field offices would be required to verify and document that fingerprint processes were being followed. As discussed above, in February 1994 INS had committed to site visits in response to the OIG inspection report. In fact, INS Headquarters sent a memorandum to the Field in March 1994 indicating that on-site audits would take place.

Responsibility for these reviews was assigned to INS' Office of Internal Audit (OIA), which for some time had been considering implementing a field assessment program as part of its general oversight work. According to Kathleen Stanley, the director of the branch within OIA responsible for the field audits, INS obtained funding for the "Field Assessment Program" in September 1995. The program eventually became known as "Inspect," which stood for "INS Program for Excellence and Comprehensive Tracking."

The effectiveness of the on-site inspection program with respect to fingerprint issues, however, was diluted by having been expanded to include the review of field compliance with all INS procedures and not just those concerning fingerprint processing. INS conducted five reviews in FY 1996, but only two of those were reviews of district offices—Buffalo, New York and Harlingen, Texas—neither of which was a Key City District for CUSA. Moreover, neither district report was issued until fiscal year 1997, after the CUSA program had ended.

        (3) Certifying fingerprint services providers

INS' failure to ensure the authenticity of the fingerprints submitted with benefit applications was a central criticism of both the GAO and OIG reports. Consequently, INS paid relatively more attention to FEWG recommendations that were responsive to this concern. However, the record shows that INS moved too slowly to fix this aspect of fingerprint processing before it launched CUSA.

        1. FEWG's recommendation to certify outside entities providing fingerprinting services

The FEWG recommended that INS establish a certification process for the outside entities who conducted fingerprinting services in order to reduce the potential for fraud and increase the quality of applicants' fingerprints.312 The FEWG envisioned a certification process that would include training and monitoring of the outside entities by INS, the potential for de-certifying entities that performed poorly, and requiring outside entities to screen applicants for proof of identification. If INS developed the necessary regulations by September 1994, the FEWG projected that the certification program could begin by February 1, 1995.

        1. Designated Fingerprinting Services program

Implementation of the certification program, which was known as the "Designated Fingerprinting Services" or "DFS" program, was delayed several times. INS did not publish the proposed rule change until May 1995, and the final regulations were not published until a year later. Even though the final regulations indicated that the program would begin in November 1996, delays pushed the implementation date back to March 1, 1997, almost three years after the FEWG made the recommendation.

Shortly after the DFS program was implemented, it came under criticism from Congress as part of the April 1997 hearings on the overall failures of the CUSA program. Members of Congress alleged that INS had approved persons with disqualifying criminal records to provide fingerprinting services and that it had also certified entities that had no connection to fingerprinting services, such as liquor stores. In addition, according to INS records, once the DFS program was implemented, the percentage of fingerprint cards rejected by the FBI actually increased dramatically. These criticisms spawned legislation by Congress that effectively eliminated the DFS program as of December 3, 1997, by prohibiting INS from accepting fingerprint cards unless they were prepared by INS, a state or local law enforcement agency that was registered with INS, or U.S. consular or military offices abroad. INS quickly adopted regulations establishing new procedures in accordance with the statute that became effective in March 1998.313

        (4) Incorporating information from FBI's automated billing records

The most creative idea to come out of the FEWG's work was the recommendation that INS use the FBI's automated monthly billing records to check on the status of an applicant's fingerprint card. This strategy, as described below, would have dramatically decreased INS' vulnerability to processing errors created through its reliance on the presumptive period. However INS Headquarters, too, did not seize on this idea, before launching CUSA.

        1. The use of the billing records

The FEWG recommended that INS obtain the FBI's monthly billing records via computer-readable tape (the information had previously existed only in hard copy, but became available electronically as of August 1994). The billing records not only listed every case submitted to the FBI for a criminal history check, but also provided information on the status and results of each fingerprint check—i.e., whether it was unclassifiable, resulted in the discovery of a criminal history (an "ident"), or confirmed that no criminal history existed ("non-ident").314 Fingerprint checks that had not been completed were noted on the billing records as "pending." Failure of an applicant's A-number to appear on the list would be an indication to the adjudicator that the fingerprint card was never submitted or was being rejected by the FBI for reasons other than classifiability (e.g., a masthead reject).315

In its report, the FEWG had urged INS to incorporate use of the automated billing records in the "near term." Although this automated information from the FBI was not yet adequately "linked" to INS computer systems and could not be transferred electronically, the FEWG recognized that the automated billing data could still be consulted and used as both a tool for the officer to assess the status of the fingerprint check in relation to a benefit application and as a source of financial information concerning the fees the FBI was charging INS for applicant checks. The FEWG recommended the longer-term goal of incorporating the FBI's electronic response into the planned, reengineered CLAIMS system, and, ultimately, too electronically transmit the actual fingerprint impression.316

The automated billing records were not a perfect solution to the problem of tracking fingerprint cards because INS and the FBI sometimes had conflicting data concerning the same applicant and thus not all records could be instantly matched. However, the billing records clearly would augment and improve INS' existing processes. The records could confirm the absence of a criminal record so that the officer did not have to presume that an applicant's record was clear. They could indicate the existence of a rap sheet, thereby alerting the officer not to adjudicate a case prematurely. Of particular interest, the records could be used to examine how INS might be wasting its resources on fingerprint checks, the concern that had been foremost in INS' thinking about fingerprints since the FBI began to charge for the service in 1990.317

        1. INS postponed exploration of the automated billing records strategy

INS failed to explore use of the billing records either before or during the CUSA program. After CUSA, however, INS finally moved to take advantage of the automated billing data. When the Attorney General ordered INS in November 1996 to change its fingerprint policy to one requiring definitive record checks, EAC Aleinikoff wrote the FBI to request that the "monthly FD-258 fingerprint check billing data normally provided to the INS in hardcopy form be provided on a weekly basis in electronic form (magnetic tape cartridge)." INS' Office of Information and Resource Management (IRM) was then instructed to create a fingerprint tracking system using the automated billing records. Within a few weeks, IRM had created and INS district offices were using the "FBI Query" system in which the district office checked fingerprint status information using the applicant's A-number.

The relative ease with which INS, in December 1996, began to use the FBI billing records in lieu of reliance on a presumptive policy raised the question of why INS had failed to engage this valuable system before CUSA when it was first recommended. The record shows that this FEWG suggestion never made it beyond INS' Office of Programs. Ronald Collison, the Associate Commissioner for IRM, told the OIG that he was not aware that the data INS began to use in November 1996 had been available as early as August 1994. He told the OIG that it was "hard to believe that Programs knew about the automated billing records and didn't pick up the phone" to inform IRM.

The record shows that it was not just a "lack of communication," as Collison characterized it, that prevented the automated billing records idea from having an earlier incarnation. INS also failed to sufficiently prioritize the task of solving its fingerprint processing problems. To understand why the idea was postponed, however, it is first necessary to address the GAO report on INS fingerprint processing issued in December 1994. INS discussed the use of automated billing records during its consideration of how to respond to the GAO's findings. Ultimately, INS decided that instead of designing a method to use the billing records in the short-term, they would postpone their use until the records could be incorporated into INS' re-engineered computer system—technology that would not be installed in the Field until 1998 and 1999.

    1. The GAO report

      1. The review

As previously noted, the media accounts and congressional criticism of INS' March 1994 decision to eliminate the fingerprint check requirement resulted not only in INS reversing its decision but also in a request in May 1994 from 15 Members of Congress that the GAO initiate a review of INS' fingerprint process. Because many areas of congressional concern had been addressed by the OIG report, GAO focused its efforts on determining what actions INS had taken in response to the OIG findings. GAO concluded that INS' plans to implement a certification program for outside fingerprinting entities would address the OIG's concern about the lack of controls over those entities. However, the GAO found inadequate INS' efforts to correct the problems with the timely submission of fingerprint cards and the processing of rap sheets and rejected cards.

      1. GAO findings

GAO's first finding was that INS planned to implement a certification program that would increase INS' control over outside fingerprinting entities. Although INS advised GAO that it expected the program to be implemented in March 1995, the certification program was not implemented until March 1997.

GAO also found that untimely mailing of fingerprint cards, untimely filing of rap sheets, and failures to resubmit rejected fingerprint cards persisted in varying degrees in the district offices they reviewed, even after the OIG inspection report had identified these problems in some of the same offices. The GAO report noted that although INS had directed the district offices (via the March 1994 memorandum discussed above) to correct the problems, INS had not conducted any follow-up review to ensure that the problems had been corrected. Much like the FEWG, the GAO concluded that INS needed to monitor the district offices' compliance with fingerprinting procedures. To that end, GAO recommended that the Attorney General direct the INS Commissioner to monitor the districts' progress to ensure that they complied with proper procedures.

As the OIG and GAO reports made clear, adjudications officers could not reliably assume that the absence of criminal history information in the applicant's file meant that the applicant in fact did not have a criminal history. To address this problem, the GAO report recommended ending INS' presumptive policy by having the Attorney General direct the INS Commissioner to obtain results from the FBI of all of its fingerprint checks, including those for applicants without criminal history records, and make the results available to adjudicators before the applicant's interview. The GAO also noted that the FBI had reported that it could provide INS with the results of all record checks in electronic format.

      1. INS' response to the GAO report

The GAO report's recommendation that INS obtain the results of every fingerprint check before the applicant interview triggered alarm at INS Headquarters. Suddenly, the agency that nine months earlier had tried to do away with fingerprint checks altogether was being told that it should review the results of the checks in many more cases than it ever had before. INS Headquarters viewed such a policy that required definitive record checks for all applicants as so far beyond its reach that its response to the GAO report—which came five months after the report had been published—pointed to futuristic strategies that INS hoped would one day meet the recommended changes. INS either ignored the short-term solutions offered by the GAO or promised that they would be implemented and then failed to follow through.

        (1) INS repeats its commitment to monitoring compliance with procedures

INS agreed in its response to the GAO report—as it had more than a year before in response to the OIG report—that it needed to monitor the Field in order to ensure compliance with proper fingerprint procedures. Commissioner Meissner again indicated that INS would monitor these procedures through on-site reviews. This commitment was reiterated in another letter from INS to GAO five months later in which INS said that it was then "planning" a Field Assessment Program that would be implemented in early 1996.

As discussed above, the Field Assessment Program (later "INSpect") that was eventually implemented examined far too many issues within its purview and was not an effective policing mechanism for monitoring adherence to proper fingerprint processing procedures. Furthermore, no monitoring of a Key City District's practices was undertaken before or during CUSA.

        (2) INS resists ending the presumptive policy

Requiring INS district offices to obtain and file all responses received from the FBI—thereby effectively eliminating the presumptive policy—was viewed by INS Headquarters officials as an impossibility as long as fingerprint checking remained a manual process. The GAO report quoted an INS official as saying that INS had originally adopted the presumptive policy specifically because it did not have the resources to file all of the FBI responses. In e-mail messages to other INS Headquarters officials discussing INS' response to the recommendation, then Associate Commissioner Crocetti wrote in January 1995: "we, i.e., the field, couldn't possibly process all of these records. Many offices can't even handle the hits (as the IG appropriately pointed out)." In preparing the Office of Programs' response to the GAO report,318 EAC Puleo wrote in February 1995 that INS' manual systems were "slow and heavily labor intensive." Implementing the GAO recommendation to obtain a positive response to each fingerprint request "would demand a large increase in staffing, seriously slow the adjudication process and result in an unacceptable increase in backlogs."

        (3) The FEWG's short-term solution of using the automated billing records was not adopted

In January 1995, a staff officer who had been part of the FEWG advised Associate Commissioner Crocetti that the FBI's automated billing records could be used to comply with the GAO recommendation that INS obtain all positive and negative fingerprint responses. He recommended that INS develop software to sort the FBI data by district and then disseminate the information to the Field. The staff officer considered the idea "not perfect" but "workable."

However, this suggestion to use the billing records appears to have been quickly disregarded in the short-term. Crocetti remained concerned that GAO believed INS could actually manage all positive and negative responses from the FBI. He said in an e-mail message that it could be done "when we're automated," but not while the process was still manual. As INS' subsequent response to GAO made clear, "automation" meant when sufficient technology existed in the districts to access the FBI billing records data through INS' CLAIMS computer system. It did not mean the relatively simple and far less ambitious plan to sort the FBI data and distribute it to the Field as the staff officer had originally suggested. Unfortunately, INS' vision of a technologically advanced future disallowed the possibility that software could be designed in the short-term to create a "workable" solution.

When asked by the OIG why the staff member's recommendations were not adopted, Crocetti said that he "didn't know" what had happened and pointed back to the staff member. He generally chastised those below him in the chain of command for distancing themselves from "accountability." He told the OIG that he was concerned about how "individuals who [were] the owners, the primary staffer working those issues [didn't] take the necessary action and follow-up to make things happen."

The staff officer, for his part, recalled that after making his recommendations to Crocetti he spoke to anyone at INS who would listen about using the FBI's automated billing records. Implementing such a plan would have required communication from the staff officer's Office of Programs (through Associate Commissioner Crocetti), to EAC Aleinikoff, who in turn had the authority to direct INS' IRM Office. Although the staff officer recalled that Crocetti told him to explore the possibilities of how to use the billing records, he believed it was something he was supposed to do in his "spare time." He also recalled that he had "screamed" to Assistant Commissioner Michael Aytes about needing resources for his fingerprint projects, but always to no avail.

        (4) Exaggerating INS' progress in the response to the GAO report

In developing the agency's response to the GAO report, EAC Puleo recommended that INS concur "in principle" with the GAO recommendation and then explain what steps it was taking to become automated, including reengineering CLAIMS.319 Puleo recommended that INS advise GAO that it was "currently completing" the development of CLAIMS and was "beginning the local office installation process," that would give INS the capacity to transmit fingerprint check results directly to district offices. Commissioner Meissner's official response to the GAO report indicated that INS concurred with the recommendation that INS should obtain all fingerprint responses from the FBI. Adopting the exact language of the Puleo memorandum, she assured Congress that the impending installation of CLAIMS would resolve this recommendation.

However, INS' representation in this May 1995 letter that installation of CLAIMS in local offices was underway was a gross exaggeration. In-house discussions at the most abstract level about reengineering CLAIMS had begun, but the contractor that eventually developed the new CLAIMS system did not even begin work on the project until November 1995. Indeed, in response to a request from the Department's Justice Management Division for an update on INS' progress in addressing the GAO recommendations, then-EAC for Programs Aleinikoff reported in October 1995 through the Office of Internal Audit that INS was "beginning" installation of the new CLAIMS system that would be adapted to use the electronic FBI billing data. He also noted that the new system was not scheduled to be operational until 1997.320

    1. INS officials plan CUSA without fixing known deficiencies

By the summer of 1995, INS Headquarters had recognized that its presumptive policy was not an ideal approach to the processing of naturalization applicants' criminal histories. INS Headquarters had, however, rejected the GAO's call to immediately eliminate the policy on the grounds that the alternative was too costly. And, at the same time, it had failed to take even inexpensive steps—like clearly articulating model rules for processing fingerprint cards and rap sheets—that would have reduced the risks inherent in the presumptive policy. The flaws and vulnerabilities in INS' criminal history checking procedures had been repeatedly exposed, yet INS failed to take corrective actions.

So far in this chapter, we have examined various factors that contributed to INS' lethargic response to the obvious warning signs that its fingerprint checking system was breaking down. At the core of this institutional inertia was the sense that the FBI fingerprint check was not truly necessary for every naturalization applicant because historically this had been a low-risk population. In those relatively few instances when an applicant's criminal history might affect the decision to naturalize, INS believed that a variety of other "safety valves" in the system would expose any disqualifying characteristics. These attitudes encouraged INS officials in their decision that the cost was too great to administer the fingerprint check process properly. And so, INS launched CUSA without conducting a risk assessment or any other serious study of the impact to the integrity of adjudications of a drastically increased naturalization production.

Commissioner Meissner acknowledged to the OIG that the "fingerprint issue" was CUSA's "fatal flaw," but she believed that INS officials had taken "measures to deal with it" when they "began to see it happening." She said that while the problems had not been adequately foreseen, she pointed out that "you have to take . . . some risks if you're going to try to do things, [to] change the pattern of the way things have been done in the past." The historic pattern Commissioner Meissner dedicated herself to changing was complacency about long waiting times and huge naturalization backlogs. The risk, though, was that the system then in place was not prepared to handle current naturalization demands let alone a major initiative that would substantially increase these demands.

Two officials working with Commissioner Meissner as she launched CUSA had particular responsibility for overseeing the proper functioning of INS' naturalization program. First, Alexander Aleinikoff, who became Executive Associate Commissioner for Programs in June 1995, had been specifically tasked with this responsibility in late 1995. Second, Associate Commissioner Crocetti, who was the "designated process owner" for the FEWG's work, led the CUSA effort along with David Rosenberg.

Echoing the Commissioner's view that INS acted when it recognized problems, Aleinikoff told the OIG that he had not recognized any problems in the naturalization process before CUSA. Given his involvement with INS' response to the GAO findings, Aleinikoff, in fact, had considerable notice that INS' system was malfunctioning. However, he and other INS managers did not consider the fingerprint process as something that required attention before CUSA began. Instead, what required attention from their perspective was the production priority.

Crocetti, for his part, did not contend to the OIG that the fingerprint problems were either unknown or unforeseen. As noted throughout this chapter, he acknowledged to the OIG the many times prior to CUSA that INS failed—and thus, implicitly, that he failed—to take action to improve the process. According to Crocetti, as of fiscal year 1996, CUSA was the agency's priority and fingerprint check processing integrity was not one of that program's stated objectives.

In the following section, we describe what happened during CUSA as production was prioritized over the proper administration of applicant fingerprint checks.

  1. Criminal history checking procedures during CUSA

    1. Introduction

INS had not diligently undertaken the repair of its fingerprint processing system in the wake of the reports by the OIG and GAO. The mounting pressure, from the Commissioner and from outside INS, to aggressively tackle the naturalization backlogs also curtailed the opportunity that did exist in 1995 to seriously overhaul those procedures. The combined result of the attitude of INS officials toward the value of fingerprint checks, their failure to take timely ameliorative action in regard to known system weaknesses, and the production goals of CUSA further exacerbated the problems in INS' criminal history checking procedures.

In the discussion that follows, we examine the impact of the CUSA data-entry project on fingerprint processing procedures in early fiscal year 1996. We then describe the transition to Direct Mail, and outline the ways in which the poor training of the service center personnel resulted in serious flaws in the processing of fingerprint cards and criminal history checks. After our discussion of Direct Mail, we turn our attention to the final major processing change implemented during CUSA, the opening of the Fingerprint Clearance Coordination Center in Lincoln, Nebraska, in June 1996.

Before leaving our discussion of fingerprint checking procedures during CUSA, we examine the issue raised at the congressional hearings in early 1997: INS' failure to timely respond to increases in the time it was taking the FBI to conduct fingerprint checks. INS' failure to adjust the length of its presumptive period to this increase was one more way in which INS failed to properly administer its fingerprint processing procedures during CUSA.

Finally, we conclude our review of CUSA criminal history checking procedures by describing INS' "bio-check" process for screening naturalization applicants and the ways in which it, too, was improperly administered before and during CUSA.

    1. Fingerprint processing and the Naturalization Data Entry Center

      1. Introduction

In our overview chapter and again in our chapter concerning INS A-file policy, we described the Naturalization Data Entry Center (NDEC) in Laguna Niguel, California. A product of INS' "reengineering naturalization" discussions in the spring of 1995, the idea behind NDEC was that in order to institute serious backlog reduction, INS first had to address the voluminous "frontlogs" of naturalization applications that had been building over the course of several years. The largest "frontlog" of naturalization cases was in the Los Angeles District, where more than 220,000 cases were awaiting initial processing or "data-entry."

As previously discussed, NDEC was designed to quickly process naturalization applications so that, for example, the 220,000 Los Angeles District applicants could be scheduled for interviews as soon as possible. Detailees were expected to arrive within six weeks after clerks had begun to data-enter the applications in mid-August 1995 to staff a new naturalization office, and INS was counting on NDEC to have cases ready for them to adjudicate in October 1995. In those six weeks, the project succeeded in data-entering these applications, many of which had been sitting untouched in the Los Angeles District Office for more than one year with the applicants' fingerprint cards still attached. However, in CUSA project managers' rush to get these applications processed and to get the applicants scheduled for naturalization interviews, they paid little attention to the ramifications this data-entry project would have on two crucial aspects of thorough naturalization adjudication: the availability to the interviewer of an applicant's permanent file and results of a pre-interview fingerprint check by the FBI.

NDEC was a nationally designed project, overseen by Terrance O'Reilly from INS Headquarters who was sent to Los Angeles by Crocetti to serve as the CUSA site coordinator. Otherwise, managers administered NDEC from the Los Angeles District Office who supervised contract clerical staff. Consequently, NDEC's approach to processing fingerprint checks was influenced by both Headquarters' and the Los Angeles District's understanding of proper fingerprint processing procedures.

As the NDEC project got underway, INS officials did pay some attention to fingerprint checking procedures. Project organizers, cognizant of previous admonitions to the Field not to waste money by sending unsuitable cards to the FBI, put into place a procedure to review and replace unsuitable fingerprint cards before they were sent for processing. While this was going on, clerks at NDEC worked to data-enter all of the applications into INS' automated processing system. However, NDEC managers failed to ensure that the manual fingerprint processing steps, such as stripping and sending the cards to the FBI, were taken before the automated case-scheduling system presumed that the applicant's initial fingerprint card processing had been completed. As a result, the automated system scheduled naturalization interviews at a rapid pace independent of whether sufficient time had passed between the time the applicant's card had actually been sent to the FBI and the time of the scheduled interview or even naturalization. Accordingly, although NDEC procedures may have succeeded in reducing the number of unsuitable cards INS otherwise might have sent to the FBI, they did little to ensure that results from the FBI fingerprint check were available to the adjudicators before an applicant was interviewed or naturalized.

This error resulted from the series of problems that pervaded naturalization processing during CUSA. First, both O'Reilly, the NDEC site coordinator, and Los Angeles officials were unaware of proper fingerprint processing procedures. They also were unaware of how NDEC's automated scheduling system functioned, and therefore relied on this system to provide safeguards it could not and did not provide. These errors could easily have been avoided had INS Headquarters followed through on the commitment it made in response to the GAO report to clarify fingerprint processing procedures for the Field. In addition, INS officials both in Washington and at NDEC were almost exclusively concerned with "moving" naturalization cases. In their quest to meet CUSA's production goals, they failed to communicate about, or otherwise pay sufficient attention to, the integrity of the resulting adjudications.

In the discussion that follows, we examine NDEC's data-entry and fingerprint processing procedures and then address NDEC project managers' understandings about the presumptive period and INS automated systems. We conclude this section by describing various flaws in the NDEC project's approach to processing applicant fingerprint cards.

      1. Data-entry procedures and fingerprint processing at NDEC

The most important task for the NDEC project was to data-enter the Los Angeles District's "frontlogged" naturalization applications. Not only would this data-entry permit INS to accurately show the number of cases pending in the Los Angeles District, but it also was the action on which any further processing of these naturalization applications depended. The NDEC effort was CUSA's first large-scale experiment, and its success would be the first concrete indication that INS was serious about reducing application backlogs.

Before NDEC, naturalization applications awaiting data-entry were stored in the basement of the Los Angeles District Office according to the month in which they were filed. Some of the applications (approximately 60,000) already had their fingerprint cards stripped, a remnant from the era when the Los Angeles Records Division, and not Adjudications, stripped and sent the cards to the FBI before data-entry at the time they received the application fee. Fingerprint cards from approximately 168,000 of the "frontlogged" cases, however, had not been stripped.

Beginning in late August 1995, contractor employees at NDEC's Laguna Niguel facility began data-entering all the applications and then, when that task was completed, began stripping the fingerprint cards from the applications. NDEC clerks completed data-entry of all the frontlogged applications by September 29, 1995, and the stripping procedure by October 3.

Before the NDEC project, the Los Angeles District practice had been to mark each N-400 with a notation like "FD-258 sent" and the date when the card was stripped from the application. Fingerprint cards were usually sent to the FBI promptly after stripping, so the stripping date was considered the date on which the card was mailed to the FBI. The stripping date noted on the application was used in Los Angeles and other districts in order to assist the adjudicator at interview in determining whether the applicant had submitted a fingerprint card and whether the presumptive period had passed before the interview. During CUSA, however, adjudicators in Los Angeles presumed that the fingerprint cards had been sent to the FBI because the applications assigned to them for interview had been processed at NDEC.321 However, fingerprint "stripping" at NDEC was no longer necessarily contemporaneous with its sending the card to the FBI because NDEC had added a new step, described below, in between the date of stripping the card and the date on which the card was sent to the FBI. Thus, calculating the presumptive period from the date the fingerprint card had been stripped was no longer appropriate.

NDEC project managers, concerned about the quality of fingerprint cards that had been stored in boxes in the basement of the Los Angeles District Office, believed that many of the cards could have been damaged and would be unsuitable for processing at the FBI. They mistakenly believed that INS would be charged by the FBI for each rejected fingerprint card, so they instituted a procedure by which contractor employees reviewed the cards for suitability after they had been stripped but before they were sent to the FBI for processing.

After separating the suitable from the unsuitable cards, the contractor sent the suitable cards to the FBI. A letter notified applicants whose cards were deemed unsuitable that they had to submit a new set of fingerprints. To keep the newly submitted fingerprint cards separate from other incoming material, project managers set up a separate post office box on October 16, 1995, in Bell Gardens, California, to receive these new cards.

      1. CUSA officials allowed NDEC cases to be scheduled without regard for the presumptive period
        (1) No backlog of pending interviews at the Laguna Niguel office

The large backlog of cases in Los Angeles did not mean that cases data-entered at NDEC were only interviewed many months later. When NDEC opened in August 1995, it became the only data-entry facility in Los Angeles. When the new interviewing site at Laguna Niguel opened in October 1995, all of the cases scheduled for interview there had been data-entered at NDEC, while the downtown Los Angeles interviewing site worked on previously scheduled cases. Because a new office was opened for which the prior data-entry system could not have scheduled cases, some cases scheduled at NDEC were scheduled for interview at Laguna Niguel within one to two months, as further explained below.

        (2) Los Angeles officials took no steps to ensure that cases were scheduled for interview only after the presumptive period had passed

We found that NDEC project managers failed to take steps to ensure that cases processed at NDEC were scheduled for interview only after the presumptive period had passed. Despite having designed a system that resulted in stripping cards after data-entry and that separated the task of "stripping" fingerprint cards from "sending" fingerprint cards to the FBI, NDEC managers acted as if the only waiting period they needed to respect was the one between data-entry and interview. In addition, they mistakenly assumed that NDEC's automated processes would schedule cases for interview only after the appropriate presumptive period had passed. They relied on the clerical supervisor at NDEC to administer the scheduling. The clerical supervisor, in turn, told the OIG that he was unaware of any need to delay cases between data-entry and interview in order to ensure that the FBI had adequate time to check an applicant's fingerprint.

        1. Los Angeles officials mistakenly relied on automated systems to protect the presumptive period

As noted previously in our discussion of INS' implementation of the presumptive period, the Los Angeles District was one of the INS districts in which we found no uniform understanding of the length of the presumptive period. Some staff believed it was 40 days, some believed 45 days, and some believed the presumptive period was 60 days. Naturalization Section Chief Donald Neufeld (who, along with O'Reilly, served as coordinator of the CUSA project in Los Angeles) told the OIG that even though INS policy was a presumptive period of 45 days, he believed that INS' automated scheduling system in NACS had built in a 60-day period between initial processing of the application and interview, or at least between initial processing and the naturalization ceremony.322 He and other Los Angeles managers therefore believed there was a presumptive period "practice" of 60 days since before CUSA.

Site manager O'Reilly, who told the OIG that the presumptive period "wasn't a concern," also believed that the Los Angeles District could rely on NACS (in whose creation O'Reilly had participated) to ensure respect for the presumptive period.323 He believed the computer would keep track of the time since the fingerprint card had been stripped and sent to the FBI, and that no case would or could be scheduled for interview before a specified period had elapsed. Furthermore, O'Reilly told the OIG that when he arrived in Los Angeles, his focus was on getting the District's frontlogged cases ready for interview. He said that he did not pay attention "to the minutiae at the beginning of [the] program." Instead, he said he was concentrating on "the bigger need" of "getting facilities and getting people—hiring, getting computer systems in place, and that sort of thing."

To manage the details of case processing, both O'Reilly and Neufeld said they relied on Ygnacio Rosete, the supervisory naturalization clerk in charge of the contractor employees. When interviewed by the OIG, Rosete said that he was unaware of a requirement to allow the FBI a specified amount of time upon submission of the fingerprint card to check an applicant's criminal record. Rosete also said that no one in the Los Angeles District had emphasized the importance of this presumptive period.

          (a) NACS did not have "clocks" to prevent premature scheduling

Contrary to what these officials believed, NACS had no "flags" or "clocks" that prevented cases from being scheduled prematurely for interview. A clerk needed to "acknowledge" in NACS that certain steps had been taken, including that the fingerprint card had been stripped, in order for a case to become "available" for scheduling by the computer. However, once that acknowledgment was made the case became immediately available for scheduling based on when it had been data-entered.

The NACS system did have a safeguard relating to the time between an applicant's naturalization interview and the ceremony. When clerks updated cases after the interview, NACS would not schedule a case for ceremony unless 45 days had elapsed since an "acknowledgement" that the applicant's fingerprints were stripped during data-entry. Although this function created a protected window between the acknowledgment date and the ceremony date, it had no impact on interview scheduling since it was not triggered until the ceremony was scheduled, a step taken only after the naturalization interview.324

          (b) Fingerprint cards were not sent to the FBI when stripped at data-entry

Los Angeles officials had specifically designed a process in which fingerprint cards were not sent to the FBI at the time they were "stripped" and data-entered. Accordingly, any automated function that they believed prevented incursion into the presumptive period—an automated function that protected the 60-day period between data-entry and interview—could not possibly have served its purpose if the fingerprint card was not stripped and sent to the FBI until after data-entry. Thus, we found unreasonable Los Angeles officials' reliance on such a function as an adequate means of protecting the presumptive period.

As described above, clerks at the NDEC facility reviewed the 168,000 cards stripped at NDEC before sending those deemed "suitable" to the FBI for processing. It is unknown exactly how long this review process lasted, but the inferences from e-mail messages and from NDEC's task-by-task approach indicate that review of the stripped cards did not begin until October 3, 1995, the date by which data-entry had been completed,and that it was still underway as late as 17 days after the cases were data-entered. For those cards that were deemed "suitable" by NDEC contractor employees and sent to the FBI, no adjustment was made to the automated scheduling system to make up for the days lost during this review.325

For those cards that were deemed "unsuitable" by NDEC contractor employees and returned to the applicants, O'Reilly and other Los Angeles managers told the OIG they believed that Rosete had placed an automated "hold" on these applications and ceased any further processing until a new fingerprint card had been received. However, our investigation showed that this "hold" procedure was never implemented as envisioned.326 As a result, applications whose cards had been deemed "unsuitable" by contractor employees at NDEC were processed as if their cards had been timely sent to the FBI, even if no replacement card was ever received by INS.

The NDEC project did not keep track of the exact number of fingerprint cards that had been deemed unsuitable and thus returned to the applicants. Nor did it keep track of the number of new cards it received in response to the letters it sent to the applicants. Estimates of the number of rejected cards and the District's purported "hold" procedure were discussed in a memorandum prepared by Los Angeles District managers one year later in response to allegations made to Congress that Los Angeles was destroying fingerprint cards instead of sending them to the FBI. The memorandum reports that approximately 12,000 cards were rejected at Laguna Niguel. These allegations and the representations in the memorandum are discussed in our section of this chapter concerning fingerprint card and criminal history processing in the Los Angeles District during CUSA.

        (3) INS Headquarters implemented computer system changes founded on the mistaken belief that fingerprint checks had been initiated for NDEC cases

Even if INS' computer systems had honored the presumptive period between data-entry and interview, their own design for fingerprint card review after data-entry thus would have defeated this automated protection because cards were not sent immediately to the FBI when cases were data-entered. When a scheduling problem for cases data-entered at NDEC came to light, INS was provided a clear opportunity to correct the error and to take steps to properly postpone naturalization interviews. However, INS prioritized the rapid scheduling of naturalization interviews over more careful considerations of what these scheduling decisions would mean for processing applicant fingerprint cards.

        1. The systems change

In October 1995, once the frontlogged cases had been data-entered at NDEC, the Los Angeles CUSA project planned to open its first CUSA interview site in Laguna Niguel. Detailed adjudications officers were scheduled to begin work the first week of October. However, at that time INS discovered that the NDEC clerks had failed to take one procedural step during data-entry of all 228,625 cases. The omitted step was the acknowledgement of fingerprint stripping, that is, filling in the "fingerprints stripped?" field on the computer screen at data-entry. INS officials referred to this field as a "security flag" because it presumed that fingerprint cards were sent to the FBI after they were stripped from the applications, and thus the requirement of fingerprint stripping before an application could be scheduled for interview prevented the scheduling of interviews without the initiation of a fingerprint check. 327 The omission of this step during data-entry at NDEC prevented the 228,625 cases data-entered by October 2 from being scheduled for interview.328

CUSA site coordinator O'Reilly informed Thomas Cook, Acting Deputy Assistant Commissioner for Adjudications (later called Benefits) at INS Headquarters on October 2, 1995, that the data-entry and stripping of the NDEC applications had been completed. His message to Cook did not mention that the fingerprint cards, although stripped from the applications, had not been sent to the FBI but instead were undergoing the "suitability" review described above.

On the same day, Cook wrote to Assistant Commissioner for IRM Fernanda Young about the fact that the clerks' failure to make the "acknowledgement" during data-entry was preventing the NDEC cases from being scheduled for interview. He said that if the computerized information was not corrected by October 4, 1995, when detailed adjudications officers were scheduled to begin their work on these cases, the Los Angeles detail would have to be "terminated" and this would be a "disaster for the backlog reduction effort in LOS."329

Cook accurately noted in his message to Young that all NDEC cases had been "stripped" by that date and thus argued that the "acknowledgement" was no longer relevant as a security measure. He requested that the "security flag" in NACS be removed for all NDEC cases and that the cases already data-entered be made available for scheduling.330 Presumably because this change implicated security procedures, Cook sent a copy of his request to INS' Office of Internal Audit (OIA). The system change was implemented by IRM the next day, October 3, 1995, before OIA had responded.

The flaw in the reasoning undergirding the systems change, however, was that although all of the fingerprint cards had indeed been stripped from the 168,000 applications that arrived at NDEC with fingerprint cards still attached, these 168,000 cards had not been sent to the FBI for processing. NDEC contractor employees had not completed the "suitability" review process (described above) before the system change was implemented by IRM. This error affected cases whose fingerprint cards had not been sent to the FBI as of October 3, 1995—likely the majority of the 168,000 cases, since the "suitability" review appears not to have begun until October 2—and the representation that these cases were appropriately available for scheduling was a mistake. After the systems change, all cases data-entered at NDEC were "available" for scheduling regardless of whether a fingerprint check had been initiated. It affected all cases, including those in which the fingerprint cards were determined during subsequent review to be unsuitable for submission. The systems change therefore increased the likelihood that cases at NDEC would not have their fingerprint cards processed before the interview date.

        1. Knowledge of the risk

We found no evidence that either Cook or Young had inquired about or had been informed of NDEC's process at the time of the system change to strip the fingerprint cards but not send them immediately to the FBI. We found no evidence that INS Headquarters officials knew of the NDEC "suitability" review project. Even O'Reilly, when asked about this system change, did not recall the details.

As noted above, Kathleen Stanley, Assistant Director of OIA's Internal Review Branch, recommended caution in her October 13, 1995, response to Cook's request for a system change. As discussed previously, OIA was directed to conduct the "Field Assessment" of INS fingerprint processing procedures that had been touted that month in EAC Aleinikoff's memorandum prepared for the GAO describing INS' efforts to comply with the recommendations made in GAO's December 1994 report. Stanley proposed that Alice Smith, Special Assistant to Commissioner Meissner, who planned to travel to Laguna Niguel, review NDEC procedures to learn "what alternate controls [were] in place to ensure that all fingerprint cards are sent to the FBI." Stanley pointed out that both the GAO and OIG reports had raised issues concerning such controls.

Although Stanley's suggestion was dated only 11 days after Cook's proposal, it came ten days after the system change had been implemented. There is no evidence that any such review was undertaken, before or after the system change. O'Reilly told the OIG that he never received Stanley's response, and it appears that the document was never disseminated to officials at NDEC or at any other office in the Los Angeles District.331

There is no indication that this system change was undertaken in deliberate disregard for the risk it posed to the processing of NDEC cases. However, the evidence does suggest that INS officials were moving so quickly to schedule cases for interview that they failed to adequately check their assumptions before making sweeping changes. They had not even provided OIA more than a day to respond to their inquiry.

        (4) Computer data shows applicant interviews were not postponed until a presumptive period had passed

Additional evidence shows that CUSA site co-managers O'Reilly and Neufeld were mistaken in their assumption that INS' automated systems would prevent adjudicators from interviewing applicants until the presumptive period had lapsed. Data provided to the OIG by INS IRM shows that in October 1995 officers at the Laguna Niguel naturalization office were interviewing applicants whose applications had been data-entered as late as September 1995.

    1. Fingerprint and rap sheet processing under Direct Mail

      1. Introduction

INS' implementation of Direct Mail, like the NDEC project, was a CUSA strategy to increase naturalization production. As discussed extensively in the A-files chapter, INS believed that transferring the initial data-entry for naturalization applications to the service centers would increase its rate of data-entry and relieve the district offices of this clerical burden. However, INS had not clarified its fingerprint processing procedures by the time Direct Mail began even though it had promised to do so in response to the GAO report.

By this time, INS had incorporated in its automated systems a function designed to prevent cases from being available for scheduling for interview until 60 days had passed from data-entry. However, while INS was relying on this automated function to protect against processing naturalization applications prematurely, it was manipulating this function to ensure the rapid scheduling of interviews, thus undermining the automated protection. Direct Mail was designed to promote the rapid scheduling of naturalization interviews, not to ensure that the requisite tools for a quality adjudication—the permanent file and the applicant's criminal history check—would be available to the officer at interview.

During our investigation, we found no evidence that CUSA planners paid any attention to using Direct Mail to improve the quality of INS fingerprint processing. While service center directors, district managers, and Headquarters officials had at least briefly discussed the impact of Direct Mail on the availability of files, the record is silent concerning any discussions about the impact of Direct Mail on fingerprint processing. The transition to Direct Mail as implemented—that is, with the service centers responsible for "owning" the case—resulted in an unexpected increase in workload for the service centers with respect to all phases of fingerprint and criminal history processing, tasks with which they were largely unfamiliar. And yet, employees at the service centers were provided little guidance or training on how to perform these tasks. The result was that under Direct Mail fingerprint cards were not always submitted to the FBI in a timely manner, rejected fingerprint cards were not reliably reprocessed, and rap sheets were not timely matched with applicant files. Although the poor condition of INS' fingerprint card and rap sheet processing before the transition to Direct Mail does not provide a good benchmark, the evidence shows that INS' performance under Direct Mail during CUSA was similarly poor and cannot be considered to have been an improvement.

INS provided a very different "assessment" of the impact of Direct Mail at the Joint Hearing before the House Subcommittees on March 5, 1997. During this hearing, Commissioner Meissner pointed to implementation of Direct Mail during CUSA as a "critical improvement" to fingerprint processing. Commissioner Meissner, in her prepared statement for the hearing, said that the service centers "[screened] fingerprint cards and [mailed] them to the FBI promptly, overcoming the problem of tardy submissions of fingerprint cards. Direct Mail [allowed] data entry and other paper handling to be performed much more efficiently . . ." Although Direct Mail processing may have improved since the end of fiscal year 1996, no such success had been achieved with respect to criminal history checks during the CUSA program.

In the following sections, we describe how INS failed to adequately plan for the processing of fingerprint cards and resulting criminal histories under Direct Mail. We then show, by examining a prominent example brought to Congress' attention in October 1996,332 how that failure to plan resulted in data-entry errors that ultimately had a significant adverse impact on the processing of criminal history checks in at least two Key City Districts—New York and Miami. We show how INS' response to the problem, much like its response to the data-entry error at the NDEC project, failed to take into account the ramifications of its solution on the integrity of criminal history checks for the affected applicants. Finally, we address other fingerprint processing planning failures in these two Key City Districts during Direct Mail.

      1. Service center staff received insufficient training and guidance

The service centers were unprepared to take on their new role in fingerprint processing for many of the same reasons that they were unprepared to obtain A-files. First, INS was rushing to get Direct Mail started after INS decided in late November 1995 to implement it. In addition, much like the service centers' lack of experience obtaining files from other offices before Direct Mail, the service centers had previously processed only a few applications for INS services or benefits that required fingerprint checks. In addition to the service centers' lack of experience, Labat-Anderson, Inc. (LAI), was a new contractor that had just replaced the former contractor in three of the four service centers in the month before Direct Mail began. Although most employees of the former contractor stayed on under LAI, LAI hired a large number of new employees in 1996 and replaced on-site managers with its own managerial employees. We found that contract and INS employees alike were given no training and belated written guidance on fingerprint processing that was found to be incorrect or incomplete. They were expected to use a new computer system whose only "trial run" had been at the NDEC project, although INS had failed to conduct any review of this prototype system's successes or failures. By all accounts, both INS and contractor employees at the service centers were in a constant state of confusion with respect to fingerprint processing throughout CUSA.333

        (1) Naturalization-specific training of service center staff was not prioritized

Previously in this report we described the failure of INS to train its service center staff about naturalization processing and handling of A-files before the implementation of Direct Mail. Several INS managers we interviewed said that training for service center staff was not a priority because INS viewed the N-400 application as "just another application" and because the service centers already had expertise in data-entering INS applications.

The service centers, however, did not have expertise in processing fingerprint cards and rap sheets on a large scale. Furthermore, this lack of experience was exacerbated by the fact that INS had recently redesigned the service centers' role under Direct Mail, giving them responsibility for more of the process than just the initial tasks relating to application data-entry and fingerprint stripping. Without guidance about how to manage the subsequent stages of fingerprint processing, such as how to handle rejected cards and criminal history reports sent from the FBI, the service centers' efforts to complete the initial processing correctly would be pointless. INS Headquarters eventually issued written guidance for the processing of N-400s, but not until after Direct Mail had begun. The draft "Standard Operating Procedures" (SOPs) were not issued until March 5, 1996, six weeks into the Direct Mail season, and guidance on some crucial topics was not issued until September.

        (2) Deficiencies in the written guidance concerning fingerprint processing under Direct Mail

The SOPs issued to the service centers in draft on March 5, 1996, were generally described to the OIG as ambiguous and incomplete. This was particularly true for the SOPs related to fingerprint processing.

        1. Service centers were given the wrong "originating agency identifiers"

To ensure that responses from the FBI to fingerprint queries were returned to the service center instead of the district office, the service centers used their own originating agency identifier or "ORI" on the fingerprint cards. We found that the SOPs distributed to the Field listed the wrong ORI codes for the four service centers. INS realized the error almost four months after Direct Mail had begun when a Texas Service Center (TSC) employee sent an e-mail message to INS Headquarters in June 1996. INS Headquarters addressed the mistake later that day by sending an e-mail message with the correct ORI codes to all four service centers. A few days later, INS Headquarters informed the service centers via another e-mail message that each office would have to "work out a reasonable corrective action with the Bureau" for any mis-addressed fingerprint responses from the FBI.

        1. SOPs did not explain how to review fingerprint cards for completeness or how to process rejected cards

Important aspects of the fingerprint process were not adequately addressed in the written guidance. The SOPs contained no explanation of INS' presumptive policy. Although the SOPs offered a general instruction that the "FD-258 when received must be reviewed," they failed to explain what such a review should include other than to say that the "required" information had to be completed on the card. It was not until August 1996 that INS Headquarters specifically directed the service centers to review the fingerprint cards for completeness before submitting them to the FBI, and it was not until a month later that INS Headquarters provided the service centers with specific information about which fields in the masthead of the FD-258 were "required."

The SOPs also failed to specifically address how the service centers should handle fingerprint cards rejected by the FBI. It mentioned "FBI responses," but this phrase also referred to rap sheets sent by the FBI. The SOPs contained no explanation about the differences between masthead rejects and unclassifiable cards. For example, the SOPs failed to advise service center staff that an unclassifiable card needed to be resubmitted with a new card in order for INS to avoid being charged twice by the FBI for the same applicant.

The SOPs contained a suggestion that the service center should contact an applicant who had submitted an "unusable" card. That suggestion was in a sample form letter provided as an attachment to the SOPs. However, the instructions on the processing sheet only referred to sending letters to applicants who had failed to submit any fingerprint card with their applications.

In the absence of written SOPs, service center employees did not have any other written guidance from INS to turn to if they wanted to research the matter on their own. As previously discussed, the Operations Instructions were also silent about these matters.334

      1. The ramifications of inadequate planning and guidance

To illustrate INS' lack of preparation for the transition to Direct Mail during CUSA with respect to fingerprint processing, we begin by describing a data-entry error—the omission of a required step indicating that fingerprint cards had been sent to the FBI—that occurred in all four service centers and affected approximately 85,000 cases. The data-entry error caused by INS' poor planning and failure to train the contractor staff at the service centers, resulted in an inability to schedule these cases for interviews.

Unlike what occurred at NDEC, however, INS' resolution of this scheduling problem evidenced a conscious disregard for the proper processing of criminal history checks of some applicants in New York and Miami. In these Districts, INS moved forward to adjudicate these cases without respect for the presumptive period.

This data-entry error led to other problems in New York and Miami Districts that we discuss below. Consequently, many thousands of applicants were naturalized before their criminal history checks could be completed.

        (1) Background on software design innovations to promote the proper processing of fingerprint cards

With the implementation of Direct Mail, INS revised its CLAIMS software, introduced for the NDEC project, to add automated functions to promote the proper processing of fingerprint cards and the transfer of files.335 Before the implementation of Direct Mail, the CLAIMS software developed for NDEC, like NACS, required only one data-entry step for fingerprint cards. Data-entry clerks had to acknowledge that the fingerprint card had been stripped (and presumably sent to the FBI). If this acknowledgement was not made, the computer would prevent the case from being scheduled for interview.336 For Direct Mail, INS revised the CLAIMS software, adding an additional data-entry step to ensure that applicants scheduled for interview had in fact submitted a fingerprint card and changing the data-entry field that was the acknowledgement that the fingerprint card had been sent to the FBI. The revised CLAIMS software also included an internal "clock" to prevent the scheduling of a naturalization interview before 60 days had elapsed, during which time the service center would process the fingerprint card and obtain the applicant's A-file.

In the revised software, the first and new data-entry step was an acknowledgement that the applicant had submitted a fingerprint card with the application.337 This step, contained in a "pop-up" screen that appeared when the clerk clicked on a button labeled "Evidence," asked whether the naturalization applicant had submitted a fingerprint card and two photographs.338

After the data-entry clerk made that acknowledgement, the second data-entry step was the one that had previously been required in the version of CLAIMS used at NDEC and in NACS, an acknowledgement that the fingerprint card had been stripped and sent to the FBI. However, the data-entry field was different than what was in NACS or the previous version of CLAIMS. At the bottom of the N-400 screen and below the "Evidence" button, the system contained a data-entry field next to the words "Fingerprints Processed." If the applicant had submitted a fingerprint card and it had been sent to the FBI, the data-entry clerk was supposed to mark the field.

In addition to these changes, the revised CLAIMS software included a "clock" that was triggered by this second data-entry step. This clock would not allow a case to be available for scheduling for an interview until 60 days after the "Fingerprints Processed" field had been marked. Once the 60 days had passed, the case would change "ownership" in NACS from the service center to the appropriate district office and would thus be available for scheduling by the district office.339 If this data-entry step was not performed, the computer system would not trigger the 60-day clock, and the case would never become automatically available for scheduling.

        (2) INS failed to provide formal training on new software

Despite use of the CLAIMS software for N-400s for the first time in the service centers with Direct Mail, data-entry clerks received no training from INS Headquarters, IRM, or EDS, the contractor who had developed the new software. According to an INS Headquarters official who helped oversee the service centers, the INS service center employees were expected to provide any necessary training to the data-entry clerks at the service centers.340

        (3) The data-entry error and the resulting "sweep"

        1. The scheduling problem that led to the discovery of the data-entry error

In our chapter on A-file policy and practice during CUSA, we described the crisis at INS Headquarters in May 1996 when officials there learned that the New York District was unable to schedule cases that had already been data-entered at the Vermont Service Center. The Miami District had also reported to INS Headquarters that it was experiencing the same difficulty scheduling cases that had already been data-entered by the Texas Service Center. The problem in the automated scheduling system "jeopardize[d] the agency's priority and all of Doris' commitments" according to Michael Aytes, Assistant Commissioner for Benefits. He had requested an immediate solution from Headquarters IRM staff.

In researching the scheduling problem, INS discovered in June 1996 that the second data-entry step was not being consistently executed at the service centers because of confusion about the meaning of the "Fingerprints Processed" field. As a result, many cases that had been data-entered were not becoming available for scheduling. A senior-level employee from INS Headquarters who worked with the service centers told the OIG that data-entry clerks did not understand that "Fingerprints Processed" meant "fingerprint card sent to the FBI." Some clerks apparently thought that the step was to be performed only to update the computer system after a response had been received from the FBI.

Despite becoming aware of this mistake in May 1996, INS did not issue guidance immediately to service center staff.341 Documents show that by mid-June 1996, the articulation of proper procedures in regard to the "Fingerprints Processed" field was still in draft form and under consideration. Even though information about the data-entry error was spreading informally throughout the data-entry corps, the misunderstanding lingered in some quarters well into July. INS determined that as of July 10, 1996, approximately 85,000 cases had been affected by omission of this data-entry step since the service centers began data-entering naturalization applications in February 1996.342 Approximately 60,000 cases, or 70 percent of these cases, were from the Vermont Service Center (for the New York District) or the Texas Service Center (for the Miami District), the two districts that had reported to INS Headquarters in May 1996 that they lacked a sufficient number of cases to schedule for interviews.

        1. INS' Headquarters response to the scheduling problem
          (a) An automated "sweep" to change N-400 data was approved with few precautionary measures

INS responded to the scheduling problem with an automated solution. IRM reported that it could "sweep" all 85,000 cases, which meant that a computer program could be written that would automatically fill in the field that had been skipped in these cases. However, such a sweep would mean that INS was assuming that fingerprint cards had been sent to the FBI in all of the affected cases.

Before approving the sweep, Deputy Commissioner Sale requested that the Benefits Division confirm that despite the lack of indication on the N-400 screens, service center employees were, in fact, sending fingerprint cards to the FBI for processing. INS had no records to turn to for confirmation because it did not keep a log of what material it sent to the FBI for processing or of when it was sent.343 Service center personnel reported to Aytes and his staff, however, that it had been their practice to regularly submit cards to the FBI for processing, and the fact that the "Fingerprints Processed" field had not been checked in so many thousands of instances was attributable only to the data-entry misunderstanding described above. Asking service center staff whether fingerprint cards were submitted was the only step INS took before implementing the automated sweep.

Aytes requested the sweep in a July 26, 1996, memorandum to IRM. He noted that the sweep was a "priority" because "four subject sites [were] nearing a crisis situation regarding eligible cases to be scheduled for interview." Within a week, EDS wrote a computer program to effect the sweep. Once the sweep was conducted, these 85,000 cases and all additional cases affected by the error that were data-entered between July 10 and the date of the sweep became available to the district offices for scheduling.

At the same time he requested the sweep, Aytes asked IRM to select a random sample of 1,000 cases—that is, select a random sample of A-numbers—from among the affected cases in each of the Key City Districts and compare those numbers against the FBI's automated billing records to see if the fingerprint cards had been processed. This sampling was not conducted before Aytes ordered the sweep but, instead, was intended to confirm what Aytes said he had learned from Service Center Directors about their processing of fingerprint cards.344

          (b) Results of the random sampling conducted after the automated sweep

To allow enough time for the FBI to have completed its fingerprint card processing, the IRM sample included cases that had been data-entered at least 60 days before the sweep. The A-numbers of these cases were checked against the FBI's August 8, 1996, billing records to ensure that the presumptive period of 60 days had elapsed for every case in the sample.

IRM found that the FBI had received and processed the fingerprint card in less than half of the cases data-entered for each district—31.4 percent for the California Service Center, 28.9 percent for the Nebraska Service Center, 37.0 percent for the Vermont Service Center and 44.4 percent for the Texas Service Center. Lack of a match could have meant that the fingerprint card had never been received, that it had been received but rejected, that the FBI processing had not yet been completed, or that the FBI and INS had different identifying information for the same applicant.

This finding should have caused concern at INS Headquarters. Although the sample established that fingerprint cards for some of the 85,000 cases had been sent to the FBI for processing, the finding was that the FBI had not processed fingerprint cards for a majority of the applicants in the sample. Because of the low matching rate, IRM was directed to conduct a second random sample approximately two weeks later. The results of this second sample were similar, if not worse.345

The results from these samples should have triggered an alarm at INS Headquarters regardless of whether Service Center Directors believed they had sent all the fingerprint cards to the FBI. If INS' presumptive policy was being administered effectively, the FBI should have processed all of these cases. The fact that the majority had not been checked was a clear indication that something was amiss: either INS was not sending cards to the FBI promptly, the cards were being rejected at an extremely high rate, the FBI was not processing the cards in a timely fashion, or the two agencies had drastically different identifying information for the same applicant.

Aytes and his staff, however, did not interpret the results in this fashion. To the contrary, they found the results encouraging. Because the sampling established that some fingerprint cards had been processed, it was proof that the service centers were, in fact, sending the cards to the FBI despite the absence of the "Fingerprints Processed" indication in the automated system.

          (c) Despite indications of fingerprint processing shortcomings, INS did not delay the adjudication of the 85,000 cases

In his interview with the OIG, Aytes acknowledged that, in retrospect, the results of the sampling should have prompted further investigation. He said that INS should have investigated the matter of FBI processing times, but neither he nor anyone on his staff raised the issue at the time. As Aytes told the OIG, INS Headquarters' and the Field's primary concern with regard to the sweep cases was getting them scheduled for interviews, not waiting for confirmation on a fingerprint check. Because of this focus on production, Aytes never reconsidered his decision to move forward with the adjudication of the 85,000 cases despite the findings that all of the fingerprint checks had not been completed.

          (d) The deliberate removal of the automated protection of the presumptive period

The sweep put thousands of the 85,000 cases at significant risk in another way. Although INS implemented an automated solution that respected the 60-day presumptive period for cases data-entered fewer than 60 days before the sweep in Los Angeles, San Francisco, and Chicago, this was not the case in New York and Miami. Because these districts were the ones most in need of additional cases to schedule, at the direction of the Benefits Division,346 the automated solution implemented by IRM included an exception that negated the "60-day clock" for all cases data-entered before July 1, 1996, at the TSC and the VSC, regardless of whether 60 days had passed since data-entry. In other words, once the sweep had been conducted, INS had no automated process to protect the 60-day presumptive period in these two cities, and all cases became available for immediate scheduling.347

As discussed in detail below, however, INS had already risked eliminating automated protections for the presumptive period for New York and Miami cases at least once before. In June 1996, in response to the scheduling difficulties mentioned above, INS "released" cases for scheduling no matter when they had been data-entered—even if the 60-day clock had not lapsed. Accordingly, the sweep extended what had become a familiar risk with an additional 60,000 cases in New York and Miami. With respect to New York and Miami cases, INS had come to depend only on manual efforts by district or service center personnel to ensure that no case was adjudicated earlier than 60 days after the date on which the fingerprint card was sent to the FBI.

As described below, thousands of naturalization interviews took place before the presumptive period had elapsed because of INS' failure to provide sufficient guidance to service center and district staff about fingerprint processing procedures. In addition, INS' failure to appropriately monitor the presumptive period contributed to the naturalization of applicants before their criminal history reports were received from the FBI.

      1. Fingerprint processing under Direct Mail: the Texas Service Center and the Miami District

Although at the beginning of CUSA Miami had enough pending cases to constitute a 2-year naturalization processing backlog, by May 1996 few cases were "available" in NACS for the District to schedule for interview. Miami's lack of cases in NACS was caused by a variety of processing errors at the Texas Service Center and failures within NACS' automated scheduling system. INS Headquarters decided, at the Miami District's urging, to make thousands of Miami District cases available for automated scheduling immediately after data-entry and not 60 days later, as was customary.

Once INS made this exception for the Miami District cases, the vulnerability it created was exacerbated by poor communication between INS Headquarters and the Miami District and within the Miami District itself. The Miami District scheduled thousands of interviews over the summer of the CUSA project, but this scheduling success was at the expense of respecting even the minimum requirements for fingerprint processing.

        (1) Reasons why the Miami District did not have cases ready for naturalization interview

In addition to the 22,000 Miami cases subject to the service center data-entry error (described above) that could not be scheduled, Miami had developed a second "frontlog" of 25,000 cases that required initial data-entry in between the time its first "frontlog" had been sent to NDEC for processing and the date on which the TSC took over the initial processing responsibilities under Direct Mail.348 Other Miami District cases were unavailable for scheduling through the automated systems despite having been data-entered more than 60 days before due to the "leading zero" data-entry error at TSC discussed in our chapter on A-file practices. Finally, because of the mercurial interface between the CLAIMS system and NACS (see chapter on A-Files), cases that had been data-entered became "stuck" in the automated system for reasons IRM officials could not explain. During CUSA, the Miami District reported that at least 50,000 Miami cases data-entered as far back as the previous year were stuck and could not be scheduled.

        (2) INS permits Miami cases to be scheduled regardless of the date on which their fingerprint cards were sent to the FBI

Miami's CUSA site coordinator, John Bulger, urged Headquarters to "release" cases for scheduling. Aytes then approved of a strategy to schedule cases faster than the computer would ordinarily allow. At Aytes' instruction, IRM made all of the Miami cases available for scheduling as of June 7, 1996, no matter when they had been data-entered. In addition, the automated clock for Miami cases remained off for some period of time after the June case release in order to ensure that cases being data-entered could also be scheduled quickly.349 According to documents obtained in this investigation and a review of files adjudicated in the summer of 1996, releasing cases, removing the protection of the 60-day clock, and conducting the automated sweep (discussed above) permitted cases throughout June, July, August and September 1996 to be scheduled for interviews fewer than 60 days after data-entry.

        (3) Awareness at Headquarters that Miami needed to protect the 60-day period was not adequately communicated to the Field

Recognizing that the case-release process and subsequent removal of the 60-day flag would allow some cases to be scheduled for interviews and perhaps even sworn in before the presumptive period had lapsed, INS Headquarters personnel discussed the need to have the Miami District pay attention to processing dates. However, this concern from Headquarters, even if it had been adequately conveyed, was limited: officials hoped that Miami would monitor its cases to prevent anyone from being naturalized in fewer than 60 days. There was no evidence of any concern from INS Headquarters about the shrinking window between when the fingerprint card was sent to the FBI and the date of interview.

However, even this limited concern from Headquarters was barely communicated to Miami managers. One IRM official recorded in an e-mail message that he had spoken with Bulger, the CUSA site coordinator, about the need to monitor cases to ensure that they were not adjudicated or naturalized prematurely. We found no other record in the many thousands of documents provided by INS to the OIG that documents any kind of cautionary advice.

Bulger told the OIG that he was either never informed of the case release or subsequent flag change, or that he had misunderstood the information provided to him. He said that he mistakenly believed that the only cases for which the 60-day automated protection had been removed were the frontlogged cases whose fingerprint cards had been sent to the FBI a few weeks before the applications were sent to the TSC for data-entry.350 He did recall general discussions with Headquarters about the need not to naturalize applicants in less than 60 days, but said that this was not in the context of a specific scheduling discussion. Elaine Watson, Miami's Deputy Assistant District Director for Naturalization during CUSA, did not recall any discussions about delaying interviews for cases that had been only recently data-entered or any cautionary instruction about the removal of the automated 60-day protection.

Despite not recalling any particular instruction about the need to protect the presumptive period for Miami cases, DADDN Watson said she nevertheless made efforts to teach DAOs that they should be mindful of the date an applicant's fingerprint card had been stripped. She told the OIG that she did this consistent with her pre-CUSA understanding of the presumptive period.

However, as discussed previously in this chapter, Watson, like other employees in Miami, believed that the presumptive period was 45 days, not 60 days. In addition, although DAOs were instructed that the presumptive period should have elapsed before they approved an applicant at interview, it also was acceptable to simply ensure that an applicant did not naturalize within 45 days of the fingerprint card having been sent to the FBI.351

By September 1996, despite Miami's processing rules, Watson and her staff became aware that some Miami applicants were being naturalized less than 45 days after their fingerprint cards had been stripped and sent to the FBI.

        (4) IRM and FBI records show that the presumptive period was not respected

As part of this investigation, the OIG requested that INS conduct a search to determine the number of persons interviewed and the number of persons naturalized in the Miami District during the summer of 1996 (June through September) within 60 days of the date of data-entry of their application. IRM advised the OIG that 10,905 people were interviewed between June and September 1996 within 60 days of their applications being data-entered.352 IRM also advised the OIG that 3,206 people were naturalized during these same four months in less than 60 days. For these 10,905 cases, INS Headquarters had sanctioned the disregard of the 60-day presumptive period in the name of production.353

        (5) Rap sheets received only after the applicant naturalized in cases where the 60-day period was not observed

The Miami District was processing naturalization applications at unprecedented rates in the summer of 1996, and during this time the naturalization office began to receive rap sheets after applicants had naturalized or so close to the naturalization date that the staff did not have time to interfile the rap sheets in the A-files. By the end of 1996, the Miami District had accumulated 908 rap sheets that had not been matched with files.354

The OIG obtained copies of the 908 records reviewed by the Miami District at the end of 1996 and chose 100 at random for review. These 100 rap sheets were matched with a case history printout from NACS and examined for the date of data-entry ("clock-in date") by INS, interview date, ceremony date, and date the fingerprint card was processed at the FBI. The 100 rap sheets corresponded to 97 unique records. Of these 97 cases, 90 were naturalized during CUSA. Thirty of these 90 cases—all data-entered in the summer of 1996—were interviewed within 60 days after the date of data-entry. Sixteen of these 30 cases were also naturalized within 60 days of data-entry.

      1. Fingerprint processing under Direct Mail: the Vermont Service Center and the New York District

The New York District suffered problems similar to those in the Miami District in that it did not have enough cases available for scheduling after the transition to Direct Mail. More than 37,000 cases were delayed by the data-entry error related to the "Fingerprints Processed" field. In addition, approximately 17,000 cases were unavailable because of an unexplained automated systems error. INS Headquarters' solution to New York's scheduling problem was the same "fix" it had offered to Miami: disregard the 60-day clock and make cases available for immediate scheduling as of June 1996, regardless of when the case had been data-entered. As a result, the New York District began interviewing cases in less than 60 days after they had been data-entered.

The frenetic pace of scheduling cases during CUSA and the subsequent disregard for the automated clock, however, were not the only factors that contributed to profoundly impaired fingerprint processing procedures in the New York District after the transition to Direct Mail. The Vermont Service Center staff had not been trained adequately about fingerprint processing and made numerous mistakes. Within two months of the implementation of Direct Mail, the VSC discovered that contractor staff had been storing, rather than processing, rejected fingerprint cards and applicant rap sheets. Despite purported attempts by VSC staff to clarify procedures, thousands of unprocessed, rejected cards and rap sheets were found in June 1996. Numerous mistakes at the Service Center, coupled with a lack of communication between the VSC and the New York District, contributed to the naturalization of applicants whose fingerprint checks were either not conducted or were not reviewed before naturalization.

        (1) Errors at VSC

The VSC, like other service centers making the transition to Direct Mail for N-400s, did not receive guidelines on fingerprint processing from INS Headquarters, so it often implemented procedures on a trial-and-error basis, making adjustments as problems came to light. Unfortunately, numerous problems did arise and we cite two as examples.

As of April 1996, the second month of Direct Mail processing, the contractor staff at the VSC did not know how to process rejected fingerprint cards. During that month, a crate of fingerprint cards that had been rejected by the FBI for a variety of reasons was discovered during a routine audit of the file room. A message taped to the crate simply noted that the cards were to be placed in the corresponding A-files. The INS employees who conducted the file room audit sought clarification from their supervisor about whether the fingerprint cards should be filed or whether the contractor employees should be directed to request new fingerprint cards from the applicants.355

Contractor staff at the VSC also did not properly process rap sheets returned by the FBI. In late May 1996, an INS employee at the Service Center reported to her supervisor that "stacks" of rap sheets had been found in a crate under a table in the file room. The INS employee also indicated that some of the rap sheets had been received as early as April 11. Because many of the cases were already scheduled for interviews, she sent the rap sheets via overnight mail to the New York District. The INS employee's explanation of the problem in a contemporaneous e-mail was, "[i]t seems that Labat [the contractor] was not sure what to do with these so they just put them in a stack." The INS employee said in the e-mail that the Assistant Center Manager for Labat assured her that she would address the problem to ensure that the contractor employees followed proper procedures in the future.

        (2) New York District learns of VSC processing errors and discovers additional mistakes in the processing of rap sheets and rejected fingerprint cards

On June 19, 1996, the New York District Naturalization Section Chief Rose Chapman learned that her District had received a shipment of fingerprint cards from the VSC.356 New York employees could not determine whether the cards needed to be filed in applicant files or whether they were supposed to have been sent to the FBI for processing.357 New York applicants were required to submit two fingerprint cards in case the first one was not suitable for submission, so having two cards for one applicant would not have been unusual. Chapman called the VSC to ask about the shipment of fingerprint cards.

Chapman's inquiry inadvertently led to the discovery at the VSC of 12 crates of rap sheets and rejected fingerprint cards in the office of a supervisor of the contractor employees. VSC staff reviewed the material and found approximately 2,000 rap sheets. Chapman advised INS' Eastern Regional Office of the discovery of the rap sheets at the VSC and of her belief that the shipment of fingerprint cards sent to New York instead should have been sent to the FBI for processing.358 Chapman said she also alerted INS Headquarters to the problem.

A review of the 2,000 rap sheets revealed that all of them related to cases scheduled for interviews at the New York District's Garden City site, the naturalization office staffed almost exclusively with new, temporary DAOs (see chapter on interviews and adjudications). VSC staff interfiled rap sheets into applicants' case files that were still on hand at the Service Center. The remaining rap sheets were sent to the New York District.

Chapman sent an experienced DAO to the Garden City site on June 21 to review the cases for which the VSC had provided a rap sheet. Rap sheets were placed in the filesof cases that not yet been interviewed. For cases that already had been interviewed and scheduled for ceremonies, the experienced DAO and several TDAOs prepared notices to "motion" or pull the applicants from the ceremony and schedule the cases for another interview. The DAO reviewed cases that had already been naturalized to determine if de-naturalization proceedings were warranted.

It is unclear what action, if any, was taken concerning rejected fingerprint cards that were also in the crates. Although VSC Director William Yates assured INS Headquarters via e-mail on June 21 that his staff had met with LAI managers and that LAI managers were "fully aware of their responsibilities in handling these FD-258s," the evidence shows that this clearly was not the case. The next month, for example, the Labat Center Manager was asking the VSC INS managers what information the FBI required on the FD-258.

The New York District's concerns about VSC's fingerprint processing procedures continued in July. When New York first began to experience the scheduling difficulty that was later solved by the NACS sweep described above, Chapman requested naturalization files from VSC so that the New York District staff could schedule cases manually. When the files arrived, she saw that many contained fingerprint cards and no indication whether any card had been sent to the FBI for processing. Furthermore, she told the OIG that the VSC sent New York fingerprint cards for cases that had already naturalized. Chapman recalled that these cards had been sent to the VSC by applicants in response to a request for a new fingerprint card after the first card had been rejected but had never been forwarded to the FBI for processing.359 As a result of these events, Chapman requested an emergency teleconference with INS Headquartersand on July 17 sent an e-mail message directly to Associate Commissioner Crocetti detailing her concerns. This was the situation that Crocetti, in a subsequent e-mail message discussed at the March 1997 Congressional hearings, referred to as a "time bomb." 360 He instructed Assistant Commissioner Aytes and others to "get right on top of this."

By then, however, it was too little and too late. With only two months remaining in the fiscal year, most of the cases that would be processed in the New York District during the CUSA year had already gone through initial processing at the VSC.

        1. New York District learns that cases scheduled at the Garden City site either had rap sheets or rejected cards from the FBI of which the adjudicators had been unaware

Unconvinced that the VSC had corrected its data-entry and fingerprint processing procedures, Chapman visited the VSC at the end of July 1996.

Chapman told the OIG that she discovered that VSC contractor employees were unaware that they could place a case on automated "hold" that would prevent the case from being automatically scheduled upon receiving a rap sheet or rejected fingerprint card from the FBI. She also discovered that in cases in which contractor employees requested a new fingerprint card from an applicant because the first card had been rejected by the FBI (and a second card could not be retrieved from the file), they marked the naturalization processing sheet "FD 258 sent," a notation adjudicators interpreted as the date the card had been sent to the FBI, not to the applicant. Thus, the adjudicator's calculation of the presumptive period based on such notes would necessarily be inaccurate.

The most troubling processing error, however, concerned the "Fingerprints Processed" field discussed above. Like many contractor employees at other service centers, the VSC staff did not understand that the "Fingerprints Processed" field was used to indicate that the fingerprint card had been sent to the FBI. Worse, many VSC contractor employees believed the field was to be used to indicate that some return—a rejected card or a rap sheet—had been received from the FBI. Thus, cases that were made available for automated scheduling—that is, cases for which the "Fingerprints Processed" box had originally been filled in—were only those cases whose fingerprint cards were rejected by the FBI or cases in which the background check had uncovered a criminal history.

        1. VSC's Director recognizes that fingerprint processing under Direct Mail lacks integrity

Several months into Direct Mail employees at both the VSC and the New York District expressed frustration about fingerprint processing problems. Employees from both offices attributed the problems to a lack of information and guidance about the process and a lack of communication between the offices. The New York District managers voiced concerns that contractor employees at the Service Center were not being monitored closely enough. VSC employees told the OIG that they were unprepared for the number of N-400 applications that had to be processed and the amount of work associated with handling fingerprint cards.361 VSC employees consistently reported to the OIG that they were not given adequate guidance on what procedures were to be followed, and did not receive adequate answers to their questions. The resulting confusion led VSC Director Yates to tell Associate Commissioner Crocetti and others at INS Headquarters in an e-mail message dated July 18, 1996, "we need to start over in designing a processing system with integrity."

    1. The Fingerprint Clearance Coordination Center

      1. Introduction

In November 1995, EAC Aleinikoff became concerned when he learned that INS had yet to begin making the improvements in fingerprint processing it had assured Congress it would make in response to the GAO report. At approximately the same time in November, managers in the Programs office received a serious jolt: it realized that INS Headquarters' Records Division employees had been destroying, rather than processing, "hits"362 provided to INS by the FBI in response to "bio-check" requests, INS' second method of checking applicants' backgrounds for possible criminal activity (the bio-check process is examined later in this chapter). This discovery generated serious concern at INS Headquarters that INS could be naturalizing known criminals.

These concerns prompted a concentrated effort beginning in mid-December 1995 by a member of the Benefits Division staff (who had also been a member of the FEWG) to develop ways to improve INS' performance in processing both fingerprint and bio-check responses from the FBI. One recommendation that grew out of this effort was a suggestion that INS centralize the receipt of all fingerprint check and bio-check responses to quickly and efficiently sort the FBI responses and immediately relay relevant information to the Field.

This recommendation was approved in late January 1996, and the centralized office to process these responses became known as the Fingerprint Clearance Coordination Center (FCCC). The FCCC opened in June 1996 at the Nebraska Service Center and began to receive responses from the FBI for dissemination to the Field on June 24.

Much like INS' transition to Direct Mail, the FCCC was undermined by its flawed implementation at the height of CUSA. It was opened without sufficient staff to meet the demands placed on it. In addition, its procedures had not been fully developed at the time it opened for business, and the procedures that did exist were not adequately communicated to the Field. Instead of improving INS' fingerprint processing, the FCCC actually served to further weaken practices that were already deficient.

      1. Purpose of the FCCC

INS established the FCCC in an effort to provide greater integrity and accountability to its criminal history checking procedures. By dedicating staff in a centralized location to the processing of fingerprint-related information from the FBI, INS hoped to more promptly alert adjudicators to applicants' criminal histories and to any fingerprint card rejection by the FBI. In addition, INS hoped to increase accountability by designating particular employees to handle the fingerprint responses distributed by the FCCC to the field offices and to increase the likelihood that rap sheets and other responses from the FBI would be addressed sooner because they were being sent to a specified individual in each district rather than to a general mailroom address.

During implementation of the FCCC, INS emphasized another aspect of its purpose, one more in keeping with the tenor of the CUSA program. According to a March 18, 1996, memorandum from Associate Commissioner Crocetti to the Field, the FCCC also was intended to "speed the adjudication process." The memorandum assumed that the resubmission of rejected fingerprint cards was delaying cases, and thus the FCCC's more efficient processing of these records would accelerate adjudications. In addition, the Naturalization Process Changes memorandum of May 1, 1996 (also discussed in our chapter on interviews and adjudications), asserted that having to "wait 60 days for FBI to return fingerprint checks" was one of the "procedural barriers" inhibiting naturalization efficiency. It also noted that the FCCC was envisioned by its designers as a reform that would "allow [INS] to shift from a standard 60 day wait period to wait periods based on actual processing times."

      1. Design of the FCCC process

The concept of the FCCC was simple: the center would receive all fingerprint and bio-check responses from the FBI, sort and distribute them by "requesting office" or "ORI" designation, obtain and resubmit new fingerprint cards when required, and notify the field offices or service centers when the FBI needed more information from applicants or when new fingerprint cards had been resubmitted.

The FCCC was established as a separate unit within the Records Branch of the Nebraska Service Center and staffed with NSC contractor employees.363

FCCC clerks date-stamped each incoming piece of mail and separated it into rap sheets or "hits," unclassifiables, rejects (cards rejected because of masthead errors), responses from applicants to the FCCC's requests for more information, and undeliverable mail. Data-entry clerks then keyed identifying information from the fingerprint cards and rap sheets into a local database created for this purpose, the Agency Check Tracking System (ACTS). At the end of each shift, ACTS sorted the masthead rejects, unclassifiables, and rap sheets by ORI designation and applicant A-number. This information was sent automatically by facsimile to the appropriate INS office. File clerks followed up this facsimile transmission by sending the responses to the field offices by regular mail. According to the FCCC draft Standard Operating Procedures and a "gentleman's agreement" with the contractor, rap sheets/hits were supposed to be mailed to the Field within four or five days of receipt by the FCCC.364

To process masthead rejects, FCCC staff attempted to locate the missing information by accessing INS' databases. If they could locate the necessary information, they would fill out the fingerprint card and resubmit it to the FBI. If FCCC staff could not find the missing masthead data, the fingerprint card was returned to the applicant with a "request for information" (RFI), copies of which were forwarded to the appropriate district office. For unclassifiables, clerks sent an RFI to the applicant asking for new fingerprints. Copies of these RFIs were also forwarded to the field offices.

Rejected and unclassifiable fingerprint cards were supposed to be processed within 10 days of receipt at the FCCC. When applicants sent additional information pursuant to an RFI or a new fingerprint card, FCCC staff had to locate the applicant's first card and update the database before sending the new card or additional information to the FBI. FCCC staff sent a notice to the Field once a new card was submitted to the FBI.

FCCC planners contemplated that the field offices and service centers would use the faxed notifications to delay interviews and ceremonies until rap sheets were received, or until 60 days after they received notice that an applicant's new fingerprint card had been submitted to the FBI.

      1. The implementation of the FCCC
        (1) The March 18 memorandum

As pointed out in earlier chapters, March 1996 was a month in which significant changes were occurring within the CUSA program. New interview sites were opening, NPR officials were visiting the Field, and, at the end of the month, the Key City Districts learned that they were going to receive additional resources provided they committed themselves to adjudicating every case received through March 1996 by the end of the fiscal year. It was into this hectic mix that INS Headquarters' Benefits Division issued its March 18 memorandum notifying the Field of the creation of the FCCC. The memorandum directed each office to designate a "fingerprint specialist" who would receive the faxed notifications as well as the rap sheets and other responses from the FCCC.

Associate Commissioner for Benefits Crocetti issued the March 18 memorandum. Although the memorandum suggested that the FCCC would enhance fingerprint processing integrity, it did not explain how the new program would achieve that end. On the one hand, the memorandum specifically required that the designated fingerprint specialist from each adjudications division should maintain a log reflecting all information received from the FCCC by applicant name and A-number, as well as a "tracking system to ensure timeliness in the receipt of information from the FBI and in the delivery of this information to the appropriate adjudicator."365 On the other hand, the memorandum did not identify any other procedures these specialists should adopt to ensure that the FBI responses were being timely used in adjudications. Instead, offices were simply instructed to develop "local procedures" to ensure that adjudicators were aware of the information provided by the FCCC. Our investigation found that the thrust of the memorandum was perceived as keeping track of the FBI's fingerprint processing efforts, not on shoring up potentially weak practices within INS offices. Furthermore, the fingerprint specialist was not expected to work on these matters full-time, but rather on a "collateral duty" basis.

Crocetti told the OIG that the district fingerprint specialist appointed pursuant to the March 18 memorandum was envisioned as someone who would understand the entire fingerprint process and take responsibility for ensuring that everything that needed to be done was accomplished. He acknowledged that in a large district like Los Angeles such a position would likely be a full-time job. When the OIG asked Crocetti why the memorandum referred to the assignment as one that could be assumed as a "collateral duty," he modified his earlier description of the job to say that the intention "was to use the person perhaps just as a point of contact. Not necessarily the overall coordinator and manager."

The OIG also asked Crocetti about the timing of the transition to the FCCC. He acknowledged that the March 18 memorandum would have been likely to generate a negative reaction in the Field because districts had been "overwhelmed for years" by the naturalization workload. Crocetti described for the OIG how he imagined the Field responded to his memorandum: "We're telling them, oh, on top of everything you're doing, add this collateral duty. So the very nature of requiring an extra responsibility on top of a program that's overwhelmed is going to be received in a negative way."

Thus, the author of the memorandum recognized that a negative reception by the Field managers to creation of the FCCC would likely undermine the program's intent. Much like Crocetti's characterization of the memoranda INS Headquarters sent to the Field in the spring of 1994 in response to the OIG inspection report, the FCCC memorandum ran the risk of paying only "lip service" to the notion of fingerprint processing integrity. Given the timing and the lack of specificity of the March 18 memorandum, it was not surprising that the role of the fingerprint liaison position was widely misunderstood in the Field, as explained below. Before we address the Field's reaction to the implementation of this CUSA innovation, however, we discuss it from the perspective of the FCCC.

        (2) The FCCC was immediately overwhelmed

In late May and early June 1996, INS notified the FBI that beginning on June 1 it wanted responses to all Adjudications Branch-related fingerprint checks sent to a post office box at the Nebraska Service Center (NSC). The FBI agreed on June 17, 1996, to send all responses to the NSC, although the FCCC did not begin receiving FBI responses until June 24.

Although the FCCC was established, in part, to address the lack of resources in the Field to handle processing of rap sheets and rejects, the FCCC itself was handicapped from the outset due to insufficient staffing. In February 1996, the contractor manager of the FCCC estimated that 35 employees would be required to staff the FCCC based on the number of rap sheets, rejects, and unclassifiables reportedly processed by the FBI in 1995 plus a projected 10 percent increase to capture the increase in filings in 1996.366 By mid-March, FCCC managers had prepared a draft of the FCCC SOP along with a cost and staffing estimate of 29 employees.367 The Benefits Division staff official in charge of the implementation of the FCCC agreed with these staffing estimates.

In mid-April, however, Assistant Commissioner Aytes authorized only 14 contractor positions for the FCCC. In an e-mail to other Headquarters officials, explaining how the FCCC staffing decision was made, Naturalization Branch Chief Pearl Chang stated that Aytes had originally planned for the FCCC to be staffed by "8 or more but certainly not [29] contract personnel."368 We found that Aytes' calculations had been based on two erroneous premises, each of which served to underestimate the FCCC's workload. First, he based his calculations on the number of fingerprint cards processed by the FBI for INS in 1995, just as the contractor had, but he did not adjust that number to reflect the anticipated increase in filings that would occur in fiscal year 1996. Second, Aytes underestimated the percentage of fingerprint cards submitted by INS that would require additional processing at the FCCC. For example, Aytes estimated that 5 percent of the fingerprint cards submitted to the FBI would result in "hits," but the accurate figure was 10 percent. We were unable to determine the source of the numbers used by Aytes in his calculations.

The assistant manager at the FCCC told the OIG that his contractor staff was immediately backlogged after receiving one of its first shipments from the FBI. With respect to this first shipment, NSC Director Natalie Vedder reported to INS Headquarters that the FCCC had received approximately 3,200 responses and their projection was to process approximately 2,300 per day. She also reported that the second shipment of responses was almost twice as large. The INS Headquarters Benefits Division staff officer to whom she reported this information (the same officer who had first suggested the FCCC concept to Headquarters officials), responded that he had checked the number of INS fingerprint requests since January 1996 and that if the pace continued INS would submit over two million fingerprints in 1996. This number was almost double the estimate used by Aytes to calculate the staffing level needed at the FCCC. By way of an apology to Vedder, the INS Headquarters official explained in an e-mail, "the estimates we gave you earlier were based on [the] past. We got caught in CUSA." This same officer did not immediately relay the NSC director's concerns about inadequate staffing; he simply reported to Aytes that the FCCC was receiving responses "at a major and accelerated level."

FCCC staff continued to express concerns about the workload. The staff worked several weeks just to process all of the documentation received in the early shipments. By mid-July, the manager of the FCCC's contractor staff requested an increase in personnel from 14 to 36, indicating that the work was "much more time consuming than anticipated" and that the volume of work was "exceeding anticipated workloads." Two weeks later, the same manager reiterated her concern to the INS FCCC project manager, stating that she was "very uncomfortable" with the current staffing level and that she was required to pull other NSC staff from their duties to undertake the fingerprint work.

Staffing shortages at the FCCC predictably led to delays in sending fax notifications to the Field and in sending districts the actual rap sheets. When the manager of the contractor staff looked back, in early 1997, at the FCCC's workload during the summer of 1996, she noted in an e-mail that timely processing had been "near to impossible." Field personnel responsible for receiving the fax notifications and FBI responses from the FCCC also complained to the OIG about the FCCC's processing delays. Although the March 18 memorandum had emphasized that the Field should watch how long it was taking the FBI to process records, the FCCC process itself was slowing the process down.

In early August 1996, the Benefits Division officer who had made the early recommendations for centralization of FBI responses sent an e-mail message to Aytes more pointedly advising him that the FCCC needed more staff. He warned that if the staff was not augmented, INS would "delay agency check responses and either ignore the consequences (approval w/o checks) or slow adjudication." Aytes immediately forwarded the message to Associate Commissioner Crocetti. Thus, the highest level of the Benefits Division had information that the FCCC was understaffed and was struggling to meet workload demands. The FCCC eventually did receive approval to hire additional staff, but not until November 1996.369

        (3) Other results of inadequate planning at the FCCC

In addition to inadequate staffing, the FCCC was also beset by other problems that could have been obviated by improved planning and coordination. For example, the FCCC staff did not have a correct list of the district fingerprint specialists' names and fax numbers. Once the Field began to report problems with fax notifications not arriving and FBI responses being sent to wrong offices, an FCCC staff member requested that field offices confirm the name and fax number for the fingerprint specialist in each office. A month after the FCCC began receiving FBI responses, this list still had not been completed, and the FCCC sought help from INS Headquarters to compile this information.

Fingerprint processing at both the California and Vermont Service Centers was delayed by this rudimentary mistake. In the case of California, the FCCC was sending the FBI responses to the attention of the wrong employee, who was away on a detail for several months, and thus packages sat unopened for what one employee described as a "long time." Although the CSC notified the FCCC of the problem in mid-July, as of October 1996 packages sometimes were still being directed to this employee. In Vermont's case, the FCCC was sending information to the wrong building. The VSC reported in mid-July that the FCCC was sending fax notifications, rap sheets, and fingerprint cards intended for the VSC to the St. Albans Sub-office of the Portland District Office, which was located in the same city as the VSC but in a different location.

Another problem that surfaced soon after implementation of the FCCC created additional work for the fingerprint specialists in the field offices and contributed to processing delays. The faxed lists of applicants for whom rap sheets or rejected/unclassifiable fingerprint cards had been received were supposed to be sent in terminal digit order according to the last three numbers of the A-number, but the lists were often not in any order due to a mistake in the FCCC's computer program. In addition, the lists were not segregated according to the benefit for which the applicant had applied.370 This made it more time-consuming for the fingerprint specialists or other staff to find the related applicant files.

The New York District's practice after implementation of the FCCC illustrates how time-consuming the processing of disorganized lists could be. At the Brooklyn and Garden City CUSA sites, supervisors received the lists for the entire New York District via fax from the fingerprint specialist (who worked in the New York District Office in Manhattan) and posted them on the wall. Adjudicators were expected to consult the lists to determine whether an applicant they were scheduled to interview had a rap sheet or a rejected fingerprint card. Because adjudicators would have had to consult a list each day the FCCC was in operation, and because each list could be multiple pages, the fact that the lists were often not in any particular order only increased the amount of time that would be required for an adjudicator to ensure that the FCCC did not have information relevant to an upcoming adjudication. Given the disincentive in the New York District CUSA offices for spending any extra time on a naturalization application (see our chapter on interviews and adjudications), the fact that the FCCC lists did not lend themselves to quick analysis inherently discouraged adjudicators from conducting thorough searches for information about applicants' fingerprint checks.371

        (4) Confused implementation in the Field

The success of the FCCC was dependent upon not only an adequate, well-equipped staff at the FCCC, but also on the work of the fingerprint specialist in each office who received the fax notifications and fingerprint responses from the FCCC. The evidence shows that this second part of the equation was similarly unprepared.

        1. FCCC processing not immediately understood

Crocetti told the OIG that it was exactly because of inefficiencies in district offices like Los Angeles that led to the creation of the FCCC. However, for the FCCC to improve such offices, its procedures had to be clearly communicated and understood. The March 18 memorandum from INS Headquarters announcing the FCCC process, as discussed above, provided little guidance. Without such guidance the Field failed to make the transition smoothly.

In Los Angeles, the Deputy Assistant District Director for Naturalization Donald Neufeld specifically recalled that the District never received this March 18 memorandum from INS Headquarters explaining the FCCC and requesting that they designate a fingerprint specialist. He said the District first learned of it when they received a call from Headquarters asking for the name of the person Headquarters assumed they had already assigned. According to Neufeld, the Los Angeles District managers did not fully understand what the fingerprint specialist job entailed and thus assigned the task to an adjudicator who had no previous experience processing rap sheets and rejected fingerprint cards.372 Similarly, the employee selected in New York was an Examinations Assistant who had no prior experience with or knowledge of fingerprint processing before becoming the fingerprint specialist.373

District managers exacerbated the situation by failing to provide adequate guidance to these fingerprint specialists. The designated specialist for Los Angeles, for example, indicated that not only did she not have any experience with fingerprint processing procedures, her supervisors also never informed her that she had been designated. In fact, she told the OIG that she had never heard of the FCCC until she started to receive faxes from the facility. Fingerprint specialists in New York, San Jose, and at the Texas and California Service Centers all told the OIG that they had received no advance notice of their new duties.

In Chicago, the transition to processing fingerprint responses through the FCCC did not immediately trigger new procedures in the District. The District's designee was the Deputy Assistant District Director for Examinations who, when he learned of this designation upon receiving a faxed list from Nebraska, simply passed on all the FCCC material to the supervisory DAO in charge of naturalization. The supervisory DAO told the OIG that she had assigned a team of officers to sort material from the FCCC, but all three of the officers on this team recalled that they were not assigned this task until the later stages of the CUSA program. One of the DAOs on this team recalled that the assignment was made only after four Chicago employees testified before Congress at the September 10, 1996, hearing.

        1. The "hold" procedure not understood or implemented throughout the Field

One of the fingerprint processing weaknesses the FCCC was intended to specifically address was INS' reprocessing of cards rejected by the FBI. To that end, the FCCC developed procedures to notify applicants when they needed to submit new cards. The FCCC's efforts to obtain new cards from applicants were wasted, however, if the application was adjudicated in the absence of a fingerprint check. And yet, because FCCC procedures were not adequately explained to Field staff—indeed, because fingerprint procedures had never been clarified since the OIG called for such clarification in 1994—the Key City Districts failed to consistently take steps to prevent the scheduling of interviews and ceremonies even when an applicant's fingerprint card was still being processed by the FCCC.

Based on INS regulations, the FCCC allowed applicants 87 days within which to submit a new fingerprint card after an FCCC request. If the FCCC received the new fingerprint card before the 87-day period had lapsed, the new card was submitted to the FBI and a notice was sent to the appropriate field office. If 87 days passed and the applicant had not submitted a new fingerprint card, the FCCC sent a notice to the field office indicating that the applicant had failed to timely respond. The field office was then supposed to call the applicant in for an interview and deny the case based on the applicant's failure to produce the required evidence. According to FCCC guidelines, applicants whose fingerprint cards were still in some stage of processing could be prevented from being scheduled for interview by the field offices placing such applications on a computerized "hold."

Among the service centers, however, only the California Service Center trained its staff on how to use the automated "hold" function in NACS to prevent a case from proceeding to interview or ceremony when the FCCC had indicated that a response from the FBI was pending. At the Texas and Vermont Service Centers, staff had no such training. Supervisors at the Texas Service Center mistakenly believed that the FCCC would put these cases on "hold" and that this would prevent those cases from appearing on computer-generated lists of cases that needed to be sent to Miami for naturalization interviews.374 Similarly, employees at the Vermont Service Center were unaware of the procedures for preventing cases that appeared on an FCCC list from being scheduled for interview until mid-August 1996 when employees from the CSC demonstrated how to perform this function in NACS.

We found that even districts that were familiar with the NACS "hold" function failed to consistently use it in conjunction with information received from the FCCC. The Miami District's fingerprint specialist told the OIG that he understood his responsibility as limited to ensuring that the faxed lists from the FCCC reached the applicant's file before adjudication. The New York fingerprint specialist also told the OIG that she was unaware of any requirement to delay these cases.

        1. Designated "specialists" were overwhelmed

We found that the districts' fingerprint specialists were quickly overwhelmed by their new responsibilities. The combination of the large number of benefit applications and the labor-intensive procedures inherent in the FCCC's design resulted in much more work than could reasonably have been considered a "collateral duty."

As described above, the field offices received fax notifications that rap sheets and rejected fingerprint cards for certain applicants had been received from the FBI. Several days later, if everything worked according to plan, the fingerprint specialists received the actual FBI rap sheets. To ensure that adjudicators were made aware at the earliest possible moment of the existence of a rap sheet or that an applicant's fingerprint card had been rejected, several of the CUSA field offices (service centers and district offices) placed a copy of the faxed notification from the FCCC in the applicant's file. This task required staff to search for the applicant's A-number in the database, locate his or her file in the database, then retrieve the file or send the FCCC notification for interfiling. The entire process would have to be repeated a few days later upon arrival of hard copies of the rap sheets and copies of notices requesting new cards. The process was labor intensive and fingerprint specialists consistently told the OIG that they were unable to perform all of the duties associated with their position.

      1. The Field's assessment of the FCCC process

Both INS staff and contractor employees in the Field criticized the processing of FBI responses through the FCCC.

Within a few weeks of the FCCC's implementation, contractor staff at the service centers began to express concerns about the duplication of effort required by FCCC processing. For example, on July 9, 1996, an employee at the Texas Service Center wrote to his manager, "I keep looking at the FD-258 interfiling process and my head hurts . . . . Should we not wait for the hardcopy and interfile when we receive it? This would ensure that we are only handling the file once for FD-258 interfiling." Similarly, on August 1, the manager of the contractor staff at the Vermont Service Center wrote to the FCCC manager in Nebraska:

After working with this program for a month now, I'm convinced this fax business does not work! We need to persuade INS that the only positive way to stay on top of these is for you to [overnight] [the FBI responses] directly to the ORI and forget about the fax. This not only will ensure the quickest update to the applicant's file, but will also cut in half the number of lookups and interfiles we currently perform.

The Los Angeles District Director Richard Rogers told the OIG that he had complained about the FCCC proposal when he first heard about it because of its anticipated requirement that someone would have to look up a case and interfile information twice in a short period of time. Once Los Angeles had experience with the process, Rogers described the FCCC and its faxed "hit lists" as "useless" and the transition to the FCCC a "nightmare." Rosemary Melville, Deputy District Director, called processing fingerprints through the FCCC "ridiculous" because it added one more meaningless step in the process—the "stop" an applicant's record would make at the FCCC before it was sent on to the district office.

Reviews of the FCCC were critical even in a district that had better than average fingerprint processing procedures. San Francisco Assistant District Director for Adjudications David Still told the OIG that he did not understand how the FCCC could have been considered a "streamlining" idea. As Still put it, the "logic of [the FCCC] was not immediately apparent."

      1. The end of the FCCC

The goal of improving INS fingerprint processing through centralization failed because INS did not adequately staff the FCCC or adequately explain its procedures to the Field. As INS approached the end of fiscal year 1996, its only tangible attempt at improving fingerprint processing procedures in response to critical findings in OIG and GAO reviews was unsuccessful. Even with the FCCC, INS still had not developed reliable procedures to prevent cases from being adjudicated prematurely. In fact, fingerprint processing under the FCCC had become more cumbersome. By the time INS adopted the definitive record check in November 1996, INS used the FCCC to process record checks for only limited types of benefit applications. Responses from the FBI to requests for criminal history checks on naturalization applicants were once again sent directly to the Field.

Improving fingerprint processing was not the only benefit INS failed to obtain from its FCCC strategy. INS officials had hoped that implementation of the FCCC would clarify the issue of exactly how long the FBI was taking to process fingerprint checks. INS harbored the hope that if the FBI was taking fewer than 60 days, then INS, too, could reduce its adjudication time accordingly. Ironically, the strategy INS adopted in part to examine FBI processing times showed that INS could not keep pace with its burgeoning workload. However, at the same time that the FCCC was injecting delay into the fingerprint check process on INS' end, the FBI was experiencing significant backlogs, as discussed below.

    1. INS was slow to adjust to longer FBI processing times

      1. Introduction

As noted at the beginning of this chapter, INS asserted that its failure to adequately review applicants' criminal histories during CUSA was a result of its presumptive policy and changing processing times at both INS and the FBI. INS explained that as INS processing times decreased, FBI processing times increased, and strict adherence to the presumptive policy essentially blinded INS to the resulting problem of not being able to check applicants' criminal histories before interview or even naturalization.

Throughout this chapter, however, we have identified crucial aspects of the fingerprint check process wholly independent of the presumptive policy that were compromised during CUSA—the submission of suitable fingerprint cards, the replacement of rejected cards, and the scheduling of interviews only after the 60-day presumptive period had elapsed. INS ignored proper administration of these procedures, despite detailed and repeated warnings from the GAO and OIG, because it was determined to meet the demanding production goal it had set. Consequently, extreme production pressures were superimposed on a system that already devalued the importance of the fingerprint check. Thus, the record shows that INS' explanation of why it failed to adequately check applicants' criminal histories during CUSA is incomplete.

This is not to say that FBI processing times had no impact on INS' fingerprint processing procedures during CUSA. FBI processing times did, in fact, increase during fiscal year 1996 as compared to previous years. As the FBI's own backlogs grew, it began to require more than the allotted 60 days to process an applicant's fingerprint card. Increased processing times at the FBI thus made INS' reliance on its presumptive policy even more of a risk to adjudication integrity than it had been before. However, just as INS failed to respond to other indications that its fingerprint checking procedures were in need of repair, it also failed to examine the implications of information it received about what was happening at the FBI.

      1. Attempts to reduce FBI processing times to permit a presumptive policy of fewer than 60 days

Since its adoption of the presumptive policy, INS had paid little attention to actual processing times at the FBI. This lack of interest changed when INS considered a backlog reduction program and naturalization streamlining initiative. INS Headquarters officials identified FBI processing times as a potential obstacle to reaching ideal processing times during reengineering discussions in the spring of 1995. INS' plans to reduce naturalization application processing times included consideration of whether FBI processing time—then presumed to be 60 days—could also be reduced.

By March 1996, INS Headquarters officials anticipated that INS would soon be in a position to reduce processing times from the date of an application's data-entry all the way to the naturalization ceremony to less than 60 days. However, until August 1996, INS never sought to compare actual FBI processing times with INS processing times nor did INS consider the notion that it might have to slow down its ambitious timetable if FBI processing times no longer fell within the parameters set by the presumptive period.

In March 1996, as evidenced by the FCCC memorandum discussed above, INS rekindled its interest in FBI processing times. Headquarters was considering naturalization streamlining, and it wanted to explore what time might be gained by relying on actual FBI processing times instead of what they believed was the longer 60-day presumptive period. Accordingly, to determine whether INS could recommend decreasing processing times to fewer than 60 days, Branch Chief Chang and another INS Headquarters Benefits Division staff officer contacted the FBI in late March 1996 to determine FBI processing times.375

According to notes from those contacts, the FBI indicated that it was receiving about 45,000 fingerprint cards per day (from all contributing agencies) and could only process approximately 36,000 cards each day. The notes also indicated that FBI officials informed Chang and the other Headquarters official that FBI processing times were increasing not only due to increased volume but also because they were relocating half their staff to West Virginia, because of government furloughs, and because of closures of the FBI fingerprint facility due to weather. The FBI's growing backlog meant that "non-hits" were taking six to eight weeks to process, while "hits" or "idents" required eight to ten weeks. In short, it was taking the FBI most of the 60-day presumptive period to complete the fingerprint checking process for applicants who did not have criminal records. Worse, on average it took the FBI longer than the presumptive period to locate the record of applicants who had a criminal history.

Chang and the Benefits Division staff officer advised EAC Aleinikoff and Benefits Division staff of what they had learned. A few days later, this Benefits Division staff officer drafted the section of the Naturalization Process Changes memorandum (the final version of which was issued in May) that addressed fingerprint checks. In his draft written on April 4, he noted that the FBI's "current processing time" was "in excess of 8-10 weeks."

During the same period as these meetings with the FBI, EAC Aleinikoff obtained separate information that the FBI completed fingerprint checks in an average of 20 calendar days and that 10 days should be added for mailing—a total of 30 calendar days or a little more than four weeks. Because this conflicted with information he had received at meetings with the FBI, the Benefits Division staff officer working with Chang reviewed other recent information he had received from the FBI about its processing times.376 He sent an e-mail message to EAC Aleinikoff, Associate Commissioner Crocetti, Assistant Commissioner Aytes, Deputy Assistant Commissioner Cook, and Chang on April 18 indicating that the FBI had a backlog of 1.6 million fingerprint cards and, based on a reported daily output of 47,000 cards, had a backlog of 37 working days.377This backlog meant an additional five weeks' processing time for the fingerprint card in addition to the actual processing time required by the FBI to perform the check. Chang also indicated that the FBI's processing time was "a lot longer than before" in an e-mail message sent on April 18 to Cook, Aytes, Crocetti, and Aleinikoff.378

After learning this information from the FBI, INS did not revoke its presumptive policy or otherwise inform the Field that the FBI could not complete the fingerprint check in fewer than 60 days. In addition, INS did not share with the Field how long the FBI was taking to process cards, in particular those cards that resulted in "hits." The only action taken by Headquarters staff was EAC Aleinikoff's decision to change the original draft of the Naturalization Process Changes memorandum from warning that FBI fingerprint processing took "in excess of 8-10 weeks" to "in excess of 30 days." The memorandum that finally issued on May 1, 1996, did not include any indication that criminal history "hits," in particular, might take the FBI even more time to process.

      1. Concerns from Headquarters staff and from the Field that naturalization processing times did not allow sufficient time for an applicant's fingerprint check

As previously discussed, many Headquarters and district officials believed that interview backlogs provided some security against interviewing applicants before fingerprint checks were completed. However, this belief rested on the assumption that INS always stripped and sent fingerprint cards to the FBI soon after it received the application. During CUSA—notably at the NDEC project—we have seen that this was not an accurate assumption. Even if such an assumption could have been made in 1995, however, by the spring of 1996 INS was rapidly decreasing its interview backlogs and by then the idea of completing the entire process in fewer than 60 days—was more credible.

By June 1996, INS was striving to decrease to 30 days the amount of time from receipt of application to interview date. INS was no longer requesting information about FBI processing times379 or monitoring the receipt of FBI responses in the Field. At the same time, INS Headquarters staff, aware of recent steps INS had taken to rapidly schedule cases, expressed concern that INS could be adjudicating applications before it received completed fingerprint checks. For example, a Benefits Division staff officer who had been working on implementation of the FCCC reported in a June 10 e-mail to Pearl Chang, Thomas Cook, and others that INS' recent decision to permit New York and Miami Districts to schedule cases without regard to the 60-day clock (as discussed above under "Direct Mail") threatened INS' ability to timely check applicants' fingerprints. He wrote:

According to the FBI, their current processing time is 20-25 workdays or 6 to 8 weeks, on roughly 90% of the prints we send them. However, even at that rate the chances of them making [sic] to the files are slim. Now with MIA and NYC dropping their cases almost immediately for scheduling, the time frames become even more critical.

Staff in INS' Office of Information and Resource Management also recognized the problem. One staff member advised Assistant Commissioner for IRM Fernanda Young in a memorandum that she should "speak with Don Crocetti on FBI taking 3-4 months to return a fingerprint hit match back. This would jeopardize Adjudications, since Adjudications makes an assumption that a case is OK after 45 days [sic] we have not received a response from the FBI."

By June 1996, this vulnerability had also occurred to INS managers in the Field. William Yates, Director of the Vermont Service Center, expressed concern in a June 27 e-mail to Cook, Aytes, and other INS Headquarters officials that:

[E]ven if [the contractor] employees do everything correctly and timely we may not get the record checks back from the FBI in time to include them in the files. NYC has set up such an aggressive schedule for interviews that they want them as soon as the 60 day period is up… . It is my understanding that we wait the 60 days then go with what we have. Of course this may lead to naturalizing a small number of people who may be ineligible. Is this correct?

Cook responded to Yates in an e-mail message that an INS Headquarters employee had met with the FBI and "reported that he was encouraged by their processing times." Cook indicated his concurrence with this assessment of FBI processing times when the statistics reported should have generated concern. Cook offered statistics to reassure Yates, noting that the FBI processed 76 percent of the fingerprint cards within two to three weeks and the remaining 24 percent within four to six weeks. However, Cook also indicated that some searches could take "up to 2-3 months" and that "[t]he longer the case takes to clear, the more likely it is to be a positive hit." Cook's lack of concern about the statistics he had passed on to Yates reflects the troubling disregard INS Headquarters officials then had for FBI processing times.380

      1. The slow change to a 120-day presumptive policy

The evidence indicates that INS Headquarters did not begin to discuss increasing the length of the presumptive period until an August 5, 1996, teleconference involving INS Headquarters and representatives from CUSA offices and service centers. Notes from that teleconference indicate that participants discussed the fact that the FBI had recently reported that it took 60 days to process "clean" cases, that the response time was "running into the 90 day period" and that for a "diligent search, they [were taking] up to 120 days to respond." The notes also indicate that an unidentified official from INS Headquarters recommended that INS change its presumptive period to 120 days.

However, it was not until Commissioner Meissner began to ask pointed questions of her staff that INS began to move in the direction of changing its policy of a 60-day presumptive period. The topic of processing times was raised at INS' Third Quarter Priorities Review meeting that took place on August 7 and 8, 1996, at INS Headquarters. At the meeting, David Rosenberg and Associate Commissioner Crocetti reported on the status of the CUSA program. After Crocetti generally mentioned that there were some problems with rap sheets, Commissioner Meissner reportedly asked "bluntly" whether INS was naturalizing applicants with criminal records. After a Field Operations employee pointed out that INS had information that the FBI was taking 120 days to process cases with rap sheets, the Commissioner, according to a summary of the meeting, suggested that INS consider adjusting its processing times to allow criminal records checks to be completed. According to Crocetti, the Commissioner expressed "grave concern" at the meeting about INS' fingerprint process and requested a "position paper" on the entire fingerprint process. However, Headquarters' staff was slow to respond to the Commissioner's concerns even though INS continued to receive information in August that applicants were being naturalized before rap sheets and notifications of rejected cards were received from the FBI.

The New York District acted on its own and told INS Headquarters on August 22 that it was delaying scheduling cases for interview until 120 days had passed from the filing date. According to Deputy District Director Mary Ann Gantner, New York was concerned about the number of applicants that were being naturalized without a complete fingerprint check in light of information that the FBI was taking 90-120 days to process fingerprints.

On September 27, 1996, INS finally issued a memorandum to the Field directing an increase in the presumptive policy to 120 days.381 Between the time Commissioner Meissner recommended slowing down processing times at the Priorities Review Meeting and the issuance of this memorandum, INS had adjudicated approximately 300,000 cases.382 Two months later, the presumptive policy was eliminated altogether upon order of the Attorney General, as noted at the outset of this chapter.

      1. Conclusion

We asked senior INS Headquarters' officials why they failed to respond to evidence that INS' naturalization processing times were not allowing the timely processing of applicants' fingerprint checks. Crocetti, Aytes, and others, much like Aleinikoff had done, pointed to the lulling effect that INS' large naturalization backlog had on their sense of urgency about such matters. However, given the indications that the backlog was considerably reduced by late summer of 1996, we find that it was not the backlog but the pursuit of its elimination that dulled their senses. By focusing so intently on the goals of CUSA, they failed to respond to the sacrifices in processing integrity required to reach these goals.

INS' response to congressional criticism in March 1997 of this diminution in processing integrity would have been more accurate if it had not placed FBI processing times in such a central role. Instead, INS should have noted that it failed to respond to increasing FBI processing times because of its single-minded focus on rapid scheduling and on reaching the production goals of CUSA. INS' presumptive policy, in combination with FBI and INS processing times, did not cause the many errors of CUSA. The presumptive policy simply allowed INS to ignore clear indications that its fingerprint procedures were in dire need of immediate repair.

    1. Bio-check processing for naturalization applicants

As noted at the beginning of this chapter, fingerprint checks were not the only way to conduct a criminal history background check of a naturalization applicant. During a "bio-check," the applicant's name and other biographical information were compared to data maintained by the FBI and the CIA to determine if the applicant was implicated in any federal investigation.383 These databases contained information concerning individuals associated with intelligence, counter-intelligence, organized crime, or terrorism investigations. In addition to investigative data, these databases could reveal an applicant's criminal history from a foreign country or reveal outstanding domestic or international arrest warrants.

INS' bio-check process was not specifically targeted in the congressional allegations.384 However, because we found that these procedures, too, were impaired during CUSA, and because this had a bearing on the quality of the criminal history checks conducted for naturalization applicants, we include a description of our findings concerning INS' bio-check process. Because the OIG's Audit Division recently conducted a review of INS' bio-check procedures, we refer readers to the March 1999 report, "Fingerprint and Biographical Check Services Provided by the Federal Bureau of Investigation to the Immigration and Naturalization Service," for more detail.

A review of the bio-check process as it existed during CUSA shows the same kinds of deficiencies that marred INS' processing of naturalization applicants' fingerprint cards and resulting criminal histories. INS' presumptive policy governed bio-checks as it did fingerprint checks and created the same risks of assuming that no news was good news. Procedures for handling responses from the FBI as a result of a submitted bio-check form were unknown or misunderstood, much as we saw in the processing of rejects and unclassifiable fingerprint cards. As sometimes happened in INS' processing of fingerprint cards, staff ignorance of appropriate bio-check procedures also resulted in the destruction of relevant material returned from the FBI.

Headquarters officials who were working on INS' fingerprint processing procedures learned of bio-check processing problems in late 1995, after Records Division staff inadvertently discovered that certain bio-check responses were being destroyed. Shortly thereafter, INS implemented new procedures for processing bio-check responses and eventually created the Fingerprint Clearance Coordination Center, as described earlier in this chapter, to process both bio-check and fingerprint responses from the FBI.

However, INS stumbled when it turned its attention to restoring bio-check processing integrity: when it was working on improving bio-check procedures in the spring of 1996, INS delayed the processing of more than 500,000 bio-check requests. Some of the delay was a function of the high volume of requests that were being generated at NDEC early in the CUSA program. Some delay was caused by INS' efforts to identify the reasons districts were submitting duplicate automated requests. Finally, some of the delay was purposeful, as INS requested that the FBI halt processing of bio-checks while it continued to work to solve the duplication problem. The net result was that the processing of bio-check requests generated electronically from October 1995 through April 1996, affecting more than 500,000 applicants, was not resumed until July 1996. As a result, many CUSA applicants did not have bio-checks completed before they naturalized.385

      1. Background information on bio-checks
        (1) INS procedures for submitting bio-check requests

INS has inter-agency agreements with the FBI, the CIA, and the Department of State to conduct bio-checks in connection with various benefit applications, including naturalization. Because the problems in the bio-check process we address below arose out of INS' interactions with the FBI, we limit the background information provided to bio-checks conducted by the FBI.

Historically, INS used Form G-325A to request that the FBI conduct bio-checks for naturalization applicants, and the district offices submitted these forms directly to the FBI. In the early 1980s, INS began working with the FBI to automate bio-check request submissions for adjustment of status applications. In the mid-1980s, INS began to reprogram NACS to automate bio-check submissions for naturalization applications. By 1989, district offices were using NACS to generate tapes of bio-check requests for the FBI. INS Headquarters issued a memorandum in February 1989 directing that INS offices equipped with NACS discontinue use of the paper Form G-325A to request bio-checks for naturalization applicants. All five of the CUSA Key City Districts were equipped with NACS.386

NACS tapes requesting bio-checks were created after district office clerks or service center personnel (after the transition to Direct Mail) data-entered applications into NACS or CLAIMS. Once this information was uploaded to the NACS mainframe in Texas, EDS, the INS contractor, initiated the processing of the data tapes, which were automatically transferred to a computer center at INS Headquarters. At INS Headquarters, the bio-check data became the responsibility of another INS contractor (Maxima Corporation) that was supposed to send the tapes to the FBI on a weekly basis.387

        (2) The cost of bio-checks

As noted earlier in this chapter, Congress authorized the FBI to charge fees for both bio-checks and fingerprint checks beginning in 1990. For bio-checks, the FBI fees were determined, in part, by whether the requests had been submitted in hard copy or on the automated tape. If the FBI received complete information on a tape, it could conduct an automated search of its databases; hard-copy submissions required more costly manual searches by FBI staff. The cost of the check was also dependent on the results of the search, that is, whether the FBI had to send INS written results of any information it found.

During FY 1996, the FBI charged INS $12 each for requests submitted in hard copy, regardless of the result. If the request was submitted by automated tape and no record was found, the FBI charged only $1.40. If the automated search resulted in a "hit" or a "possible hit," the FBI charged $10.65.388

FBI billing procedures during FY 1996 recorded all G-325A requests (or manual submissions) under INS' "adjustment of status" program regardless of the applicable benefit. For its part, INS did not track its manual bio-check submissions to the FBI. Therefore, the total number of bio-check submissions for naturalization applicants during FY 1996 is unknown. However, statistics do show how many automated requests were submitted through NACS during FY 1996, and all of those requests related to naturalization. FBI records reflect that between October 1, 1995, and September 30, 1996, INS submitted 969,575 bio-check requests on tape. Based on those figures, INS spent well over one million dollars in FY 1996 for automated bio-checks for naturalization cases alone.389

      1. FBI procedures for processing bio-checks
        (1) The search for information

Once the FBI's Information Resources Division (IRD) received tapes requesting bio-checks from INS, they ran an automated search comparing that data against information in their Central Records System. This process could result in either a definitive "no record" response or an indication that further matching efforts were required to determine if a record existed. Bio-checks that did not conclusively result in "no record" were known as "possible hits" and generate an Indices Popular (IP) response form. The "possible hits" undergo one or more additional search procedures at the FBI, after which the IP response form is used to report to INS the result of those additional searches.

The "possible hits" are first sent to the FBI's Name Search Unit where FBI personnel manually check the names against a database to eliminate "possible" records (records of individuals with the same name and other similar biographical information, who are not the subject of the bio-check). For this reason, the more information that INS can provide about an individual, the more quickly and thoroughly FBI staff can complete the bio-check. This Name Search Unit is also the office that initially receives bio-check requests made manually on a Form G-325A. If the Name Search Unit can eliminate all "possible hits" and determine a "hit" or a "no record," the processing is complete. In this scenario, the IP response form or the Form G-325A, as appropriate, is stamped with the final result and sent to FBI Headquarters for transmission to INS.

If the Name Search Unit cannot eliminate the "possible hits," the request is forwarded to the FBI File Review Unit. This unit manually pulls its files for the potential candidates and compares any biographical and geographical data provided to determine whether the applicant is, in fact, someone with "no record" or someone who matches information on file at the FBI (a "hit"). Again, the IP response form or the Form G-325A, as appropriate, is stamped with the final result and sent back to the FBI Headquarters for transmission to INS.

INS receives the IP response form or the Form G-325A along with any information about the applicant from the FBI. If the FBI information cannot be disseminated because it is "classified," for example, FBI file analysts draft a memorandum providing a synopsis of the releasable information. Because it will not provide information to INS related to another agency's investigations, the FBI may refer INS to another investigating agency to obtain the necessary information. This action is called a "third agency referral." The burden then falls on INS to obtain the necessary information from the agency indicated. If INS was the source of the information found in the FBI database, the FBI stamps the response: "no information in addition to that already known to your agency."

Not all bio-checks result in definitive findings. In such instances, the FBI would advise INS that it was unable to determine if the person about whom it has information is identical to the INS subject or applicant. Again, the ability of the FBI to determine with specificity whether the person in its records is the same person applying for naturalization was influenced, in large part, by the adequacy of the bio-check information provided by INS.

        (2) Advising INS of the results of the search

For bio-check requests submitted on tape, the FBI would provide the result of its search on the same tape that was submitted by INS. In addition, the FBI provided INS a summary that included a list of the names on each tape that it had accepted for checking and the names that could not be considered further because of errors in the information submitted. The FBI also notified INS of any duplicate requests.390 Attached to the summaries were the results, listed in alphabetical order by name, and of the bio-checks conducted. The FBI sent these summary documents along with the returned tapes to the Information and Resource Management (IRM) office at INS Headquarters.

For bio-checks submitted on tape by INS that resulted in "no record" after the FBI's initial automated search, no paper response was generated by the FBI. For bio-checks that were submitted in hard copy and for automated bio-checks that resulted in an initial determination of a "possible hit," the FBI would also return a hard-copy response to INS that identified the result of its searches. The response could take any one of a number of forms. For example, the FBI might return the applicant's Form G-325A stamped "NR" for "no record," or the Form G-325A with "unclassified" information attached. Or, the response might be an IP response form, the processing form used by the FBI to report the results of additional searches staff conducted on "possible hits." The IP response form could indicate that the person had no record, or it might refer INS to another agency or to an office within INS for information about the person who was the subject of the check. The IP response sheet could also cover "unclassified" or "classified" information that the FBI had found as a result of the check.

The FBI sent these hard-copy responses to different offices within INS. The FBI employee who was responsible for disseminating these responses until March 1996 told the OIG that she sent this material to either the Records Division at INS Headquarters, the Naturalization Office at INS Headquarters, or the Liaison and Records Searcher at INS' Washington District Office, depending on the type of response and the reason the bio-check had been requested by INS. In general, the FBI sent all Form G-325A responses, except those with "classified" information, to the Records Division. It sent IP response forms, whether the result was "no record" or whether "classifiable" or "unclassifiable" information was attached, to the Naturalization Office, except for responses concerning asylum and adjustment of status applications. Finally, the FBI sent G-325A forms with "classified" information and some IP responses for adjustments of status and asylum to the Liaison and Records Searcher in the Washington District Office.

        (3) Processing time

As previously discussed, the length of time required by the FBI to complete the bio-check was dependent on whether information existed in FBI databases about the applicant and on the completeness of the submissions. It also depended on the size of any backlogs for bio-check searches.

FBI officials interviewed by the OIG reported that they experienced notable backlogs during 1996. For example, in July 1996, according to FBI statistics, 24 percent of INS' requests for bio-checks were generally cleared within four to six weeks from the date of submission, although some required up to three months to complete.391 As was true in the processing of fingerprint checks, those checks that required longest to process were the ones most likely to result in "hits," or the ones most likely to reveal information that would have an impact on an adjudicator's evaluation of an applicant's eligibility for naturalization.

      1. INS procedures for processing results of bio-checks
        (1) Receipt by INS

Since 1990, the job of processing bio-check responses from the FBI was assigned to a staff member in the Naturalization Branch at INS Headquarters as one of several duties. All "unclassified" responses were forwarded to the district office that had requested the check. All "classified" responses were sent to the Alien Files & Naturalization Branch of the Records Division in Washington, and that office disseminated the responses to the Field, in accordance with INS regulations.

The Liaison and Records Searcher at the Washington District Office said that when he received bio-check "hits" from the FBI, he logged the information and then forwarded the responses to the appropriate office in the Field.

Although the FBI indicated that only "hits" ("classified" G-325A information, all asylum IP forms with "hits," and "classified" IP forms for adjustment of status) were sent to the Liaison and Records Searcher, the employee who has held this job since late 1994 told the OIG he received "four or five boxes monthly" from the FBI filled with other kinds of bio-check responses. He said that he did not know what this material was, but presumed it was some kind of receipt for the bio-check process and he forwarded it to INS Headquarters. It is unknown what processing, if any, occurred in relation to this material at INS Headquarters.

The FBI also sent information directly to INS Headquarters' Records Division and the Information and Resource Management office. However, neither office processed the bio-check responses it received. Because these gaps in processing were not discovered until the autumn of 1995, we address them below in our discussion of bio-checks during CUSA.

        (2) The presumptive period

In our earlier discussion about the presumptive period for fingerprint processing, we noted that some INS officials believed that the waiting period was considerably shorter than the 60-day presumptive period INS Headquarters officials claimed governed its practices during CUSA. This belief was based, in some instances, upon INS employees' understanding of the presumptive period for bio-check processing which, according to INS' Operations Instructions, clearly was 40 days.

INS' practice regarding the presumptive period for bio-checks, the OIG discovered, generally mirrored whatever presumptive period that district (or employee) followed for fingerprint checks. Consequently, a naturalization case was considered "ready" in all regards after either 40, 45, or 60 days had passed since an application was data-entered.

The reliability of the presumptive period for bio-checks, like that for fingerprint checks, depended in large part on how quickly INS sent the requested information to the FBI and how quickly it disseminated any information it received in response. It also depended on how much time the FBI required to process any particular record.

Although INS' automated submission of bio-check requests obviated many of the clerical delays associated with the processing of fingerprint cards or hard-copy G-325A forms, the responses from the FBI took a more circuitous route back to the requesting office. Because the responses were sent to INS Headquarters for processing and not directly to the field office that requested the information, an additional processing step was built-in to the procedure for receiving any bio-check "hits." In addition, both FBI and CIA officials told the OIG that they advised INS not to rely on a presumptive policy in regard to the bio-check process.

However, the question of the reliability of the presumptive period with regard to bio-checks was rendered moot during CUSA. We found that bio-checks were improperly processed during CUSA, at one point deliberately suspended, and suffered weaknesses wholly unrelated to the presumptive period policy. These problems and their consequences are discussed below.

      1. Bio-check processing during CUSA

The evidence shows that except for the dissemination of classified responses by INS to the Field, INS' bio-checking procedures were deficient through the end of 1995. Other "hits"—searches that had resulted in an "unclassified" response to a manual G-325A request and had been returned to the Records Division—and other types of bio-check responses were routinely destroyed. Also, despite the enormous costs associated with the bio-check process if the FBI had to resort to a manual search, INS did not consistently ensure that its automated systems provided the FBI with enough information to actually conduct an automated records check on all applicants included on the NACS-generated tape. We found no evidence that INS ever reviewed either the summary information provided by the FBI indicating that the FBI had not been able to conduct a bio-check because of insufficient information, or the lists concerning the duplicate requests made (and thus duplicate costs incurred) by INS. As a result, INS was spending large sums of money each year for a process that failed to produce bio-checks for all applicants.392

Indeed, we found that no one at INS Headquarters was paying attention to the bio-check process until an inadvertent discovery in November 1995 that INS was destroying certain responses. As a result, INS recognized processing errors it had been making with respect to bio-checks for many years. However, just as had occurred in INS' evaluation of its fingerprint card checking procedures in March 1994, once INS recognized its failures with bio-checks, it considered whether to retain the bio-check process at all.

While INS considered the value of the bio-check and worked to identify why NACS was sending the FBI so many duplicate automated bio-check requests, it inadvertently delayed the processing of approximately 500,000 bio-check requests that were already on tape. While INS was working to fix this problem, INS requested that the FBI suspend for approximately one month (April-May 1996) its processing of bio-check requests already submitted on tape.

In May 1996, INS Headquarters officials who recognized the importance of the information that could be revealed through a bio-check won the day, and INS continued to make bio-check requests for naturalization applicants. INS used the FCCC in Nebraska as a centralized location to process bio-check responses. Even those requests that had been delayed were eventually submitted to the FBI. However, this resumption of bio-check processing did not occur until July 1996, by which time the majority of the affected applicants, if otherwise found eligible, would have naturalized.

In the discussion that follows we offer a brief review of the bio-check processing errors that were discovered in November 1995 and the efforts INS made to correct them. We examine INS' delay of the bio-check requests for more than 500,000 CUSA cases. The record shows that INS' weak bio-check procedures were another significant failure by INS to ensure that adequate background checks had been conducted for applicants naturalized during the CUSA initiative.

        (1) Information provided by NACS insufficient to prevent the need for manual searches at FBI

At a minimum, the FBI required the individual's name, race, sex, place of birth, date of birth, and originating agency identification number (ORI code) in order to conduct a bio-check. In addition to the minimum data, optional information could be supplied, including social security number, employer information, former addresses, aliases, and parents' names. The FBI preferred to receive as much of the optional data as possible to help match records in the database.

The hard-copy Form G-325A contained as much of this more expansive identifying data as had been submitted by the applicant. When INS submitted the information via NACS, however, the data tape transmitted only the applicant's primary name and the date and place of birth. Bio-check searches on applicants who submitted only this minimal amount of information could take more time for the FBI to complete.

As noted above, the five Key City Districts during CUSA used NACS and thus submitted their bio-check requests by electronic tape. After the transition to Direct Mail, the service centers continued to submit bio-check requests to the FBI using NACS. Accordingly, all of the bio-check requests made by INS during CUSA provided only this minimum amount of identifying information to the FBI. Consequently, many of these bio-check requests would require manual searches because of the lack of identifying data about an applicant. The FBI informed INS of this processing weakness by the FBI in April 1994, and by the middle of the CUSA project INS estimated that 30 percent of the bio-checks submitted by tape required manual searches. Ironically, the transition to the NACS automated bio-check request process had not guaranteed that the check could be more quickly or less expensively conducted.

        (2) Neither IRM nor the Records Division processed information returned as a result of the bio-check process

        1. IRM

The FBI regularly returned to INS a tape that recorded the results of all of the bio-check requests submitted electronically. With the tape, the FBI provided a summary of the status of the checks. The tape and summary also indicated which names had not been checked because of errors in the data, and which submissions were duplicates.

INS had no procedures outlining how its Headquarters or Field staff should use this information. As a result, the tapes returned from the FBI were regularly reused without review of the data, and the written summaries forwarded to INS Headquarters were unused and eventually discarded.

INS IRM staff interviewed by the OIG were not aware of any written INS policies and procedures related to electronic bio-check requests other than the Naturalization System-NACS Production Control/Operations Manual, dated February 1995. This manual, developed by EDS, describes the process for creating the weekly data tapes of bio-check requests. However, the manual does not address how to monitor or track the tapes submitted, how to reconcile the results returned from the FBI for billing purposes, or how to process the results.

INS personnel provided the EDS manual to another contractor, Maxima, whose job included receiving the tapes containing the electronic responses and the associated summaries. However, because the Maxima employees who received the bio-check tapes and written summaries from the FBI were unaware of any related procedures, they returned the tapes to a central pool for reuse without review. These contractor employees forwarded the associated summaries to INS IRM personnel, who in turn forwarded them to the Benefits Division at INS Headquarters. At INS Headquarters, we found that no one claimed responsibility for reviewing this material, and the evidence shows that the summaries were usually discarded.

The tapes returned from the FBI to INS with the results of the completed bio-checks contained either information that INS would have destroyed anyway—the "no record" results393—or hits, information that should have existed in hard-copy format at one of the three INS offices to which the FBI was sending paper responses. In this regard, the recycling of the tapes and destruction of the summaries made little substantive difference. However, information on the tapes and summaries about INS' submission errors (including duplication) were not available from any other source. By overlooking this information until February 1996, INS missed the opportunity to correct the mistakes it was making—in particular, those related to the submission to the FBI of incomplete or duplicate data, both of which resulted in INS wasting money on fees charged by the FBI.

        1. The Records Division

As discussed above, the FBI sent all hard-copy G-325A responses, including "hits," rejects, some IP responses, and "no record" responses, to the Records Division of INS. Records Division managers and other records employees told the OIG that they had been unaware of the procedures for processing bio-check responses from the FBI until November 1995. In fact, a data input employee within the division had been routinely burning boxes of FBI bio-check responses since 1987 or 1988. Neither the data input employee nor his supervisor knew what the responses were or why they had been directed to burn them. There is no evidence to suggest that the data input employee sorted through the responses to cull out "hits" or rejects prior to destruction. The former Acting Chief of Records Operations who the employees said gave this direction did not recall ever instructing anyone to burn the FBI responses.

        (3) INS' response to its discovery of improper bio-check processing procedures

        1. The discovery of the burning of bio-check responses and the transition to the FCCC

The fact that the Records Division was not processing FBI bio-check responses came to light in November 1995 when Records Division managers discovered that employees within their Division had been burning FBI responses to bio-check requests for years. They contacted the Benefits Division staff officer who had been working on fingerprint card processing issues at INS Headquarters for advice. Although the staff officer was not familiar with bio-check processing procedures, he studied the matter and quickly discovered a variety of irregularities in how INS processed these FBI responses.394

The concern engendered by the discovery that INS had been burning bio-check responses was part of what prompted INS to open the FCCC in the summer of 1996. The Records Division stopped its practice of burning bio-check responses in late 1995, shortly after it had been discovered. Until the FCCC took over the processing of bio-check responses in June 1996, INS arranged with the FBI to have all bio-check "hits" sent to the Liaison and Records Searcher in the Washington office, and to have all other responses—those with "no record" or rejected G-325A forms—sent to a specific person in the Headquarters Adjudications Office for further processing.

        1. Recognition that INS had been submitting duplicate requests for bio-checks

INS managers also realized in March of 1996 that INS was wasting huge sums of money because of its duplicative submission of bio-check requests to the FBI.395 Once IRM officials began to review the summary documents the FBI continued to send to INS, they discovered that the FBI had been billing INS for duplicate submissions—and notifying INS of the duplication—since 1992. Had INS reviewed the FBI documents or tapes, this problem could have been revealed years earlier. As one IRM official pointed out, until early1996, the Office of Programs had no single point of contact responsible for reviewing, analyzing, or verifying the FBI invoices.

        1. The origin of the duplicate requests

INS also determined in March 1996 that the source of the problem was in the NACS software: NACS created duplicate bio-check requests when information was inputted into certain data fields, like name or date of birth, to correct or update previous entries. According to IRM, the software had functioned in this fashion for about ten years. This weakness in NACS had been compounded by the volume of cases inputted at the beginning of CUSA (see NDEC, above).

The number of names on the tapes generated by NACS during CUSA also caused problems. The amount of data on the tapes affected the length of time required to transfer the information from the NACS mainframe in Texas to INS Headquarters, where the bio-check request tapes were created and sent to the FBI. According to INS documents, this transfer process could take as long as 10-14 hours. Sometimes, as one IRM official reported in a contemporaneous e-mail, the "…sheer volume on the system [NACS] would not allow the job to complete…" and restarting the tape-generation process contributed to the high number of duplicates. INS installed a temporary processing change in April 1996 that sorted and removed duplicate records from the tapes prior to their weekly submission to the FBI. However, duplicate submissions on NACS tapes remained a problem throughout CUSA and continued to be a problem that was referenced in the OIG's 1997 audit of INS fingerprint and bio-check procedures.

The volume of bio-check requests on the NACS tapes also caused problems for the FBI. Its tape-processing limitations necessitated that these large tapes from INS be segmented into smaller tapes containing only 10,000 names. For example, INS submitted one tape from the NDEC project in Laguna Niguel that contained approximately 300,000 names. Further, the FBI could process no more than 30,000 names per day, a capacity limitation that helped create backlogs of tapes to be processed.

        1. Bio-check processing was delayed as INS worked to solve duplicate problems

Once the Benefits Division staff officer was aware of the problem of duplicate data and began collecting information from the FBI about the bio-check process, INS took steps to eliminate duplicate entries on previously submitted tapes. In March 1996, INS IRM began to retrieve the large tapes from the FBI that had not yet been processed (mostly those tapes created between October 1995 and March 1996), organized them into manageable sizes, and eliminated duplicates so that the tapes could be resubmitted to the FBI.

In April 1996, while INS was still trying to create smaller tapes and eliminate duplicate entries from tapes retrieved from the FBI, INS continued to submit newly generated tapes. However, in mid-April, INS determined that the FBI was still processing older tapes that INS had never retrieved and corrected. In response to this discovery, an IRM official requested via fax on April 24 that the FBI halt its processing of any INS bio-check requests. The request stated in pertinent part "…please cease processing INS Naturalization Name Check tapes until we (INS) make the correction that will eliminate duplicates, and resolve the attendant burden on both the FBI and INS in processing these unnecessary records." The fax indicated that INS would retrieve any tapes it had submitted to the FBI and that the tapes would be resubmitted once the problems were fixed.

Throughout the month of April, INS Headquarters officials were engaged in an internal debate about whether it would still be worth the cost of checking the older tapes—those created as much as six months earlier—that INS staff was then working to break down into smaller tapes without duplicate entries.396 INS did not decide until early May 1996 that it would resubmit these older tapes for processing, and the FBI did not begin processing this information until approximately July.397

In May 1996, the IRM Division reported that between October 25, 1995, and May 3, 1996, NACS had generated 901,420 bio-check requests. Of these, 237,667 (or 26 percent) were found to be duplicate submissions, leaving 663,753 unique requests. Of these unique requests, IRM reported that approximately 100,000 had been processed by the FBI as of May 3, 1996. Accordingly, because INS delayed the FBI's processing of tapes by submitting tapes that were too large, by retrieving these tapes to eliminate duplicate entries, and by requesting the FBI to halt its processing, approximately 500,000—almost half the total number of applicants naturalized during CUSA—did not have a bio-check conducted until after the presumptive period had lapsed. In the majority of those cases, no bio-check was completed until several months after the case had been data-entered by INS.

        1. INS considers the value of the bio-check process

The Benefits Division staff officer who analyzed the bio-check problems in late 1995 wrote a report describing bio-check procedures. He advised his superiors in December 1995 that INS did not "…reliably get hits to the adjudicator in time" (before adjudication). He recommended, among other things, that Headquarters instruct district offices to establish accountability for bio-checks and monitor compliance through the Field Assessment Program, the program INS was to establish to monitor field office compliance with INS policy as recommended by the GAO. As discussed above, he also recommended the centralization of responses through the Nebraska Service Center.

INS did not act immediately on the officer's suggestions.398 Instead, review of the bio-check procedures continued, and INS considered abolishing the check for naturalization applications. The same staff officer was asked to prepare a paper discussing INS' options for the bio-check process, including whether INS should continue to conduct bio-checks at all. In March 1996, the staff officer submitted the "options" paper to Assistant Commissioner Aytes. In his paper, the staff officer characterized bio-checks as "…unique and potentially of exceptional value" because they provided "the only source of information which might associate the applicant with criminal or subversive activity for which he or she has never been arrested or convicted." At the same time, he pointed out that such checks were of value only if the resulting information was made available to the adjudicator prior to naturalization.399

During the same time period that INS delayed the entire bio-check process while it attempted to fix the problem of duplicate requests, INS officials also considered the elimination of the bio-check process. Associate Commissioner Crocetti wrote to the FBI on April 30 that INS "question[ed] the value of such checks [bio-checks] " and asked whether the FBI believed such checks were necessary. The FBI's May 14 response referred Crocetti back to its own regulations, saying that if INS was required to conduct "national agency checks," the bio-check should be retained. The record does not indicate whether INS officials continued to discuss the complete elimination of the bio-check process.

      1. Conclusion

INS' administration of applicant bio-checks during CUSA thus suffered from the same inattention that characterized its fingerprint processing procedures. Governing policies either did not exist or were not understood. Fortunately, when the many flaws came to light in late 1995, INS took some action to improve the system. However, those improvements were too late to bolster the integrity of CUSA adjudications.

At the same time that INS was delaying the bio-check process in the spring of 1996, INS was also dramatically decreasing the time it took to process a naturalization case. By August 1996, even Los Angeles, the largest Key City District, had begun to process cases that had been data-entered only six months earlier. Such quick processing times indicated that, in general, cases that had been data-entered before March 1996 (when INS began retrieving tapes from the FBI) had already been processed and many of those applicants had been naturalized. The delay in processing the backlog of more than 500,000 bio-check requests, combined with application of INS' presumptive period, resulted in several hundred thousand persons being naturalized during CUSA without the benefit of a completed bio-check.

  1. Widespread errors in the processing of applicant criminal histories and INS' unreliable reports to Congress

    1. Introduction

As described throughout this chapter, INS' fingerprint processing procedures were profoundly impaired during CUSA. Despite this evidence, senior INS officials emphasized to the OIG that these errors were simply not of serious consequence. They pointed to the low number of persons identified in the post-CUSA reviews supervised by KPMG as having been naturalized despite a disqualifying criminal history (369 of the 1,049,867 persons naturalized). Referring to those numbers, these officials emphasized that given the number of applications processed, INS' performance, though not ideal, did not result in serious harm.

For example, Terrance O'Reilly, who until April 1996 was the Los Angeles site manager for CUSA and, later, the head of the CUSA program for Field Operations, told the OIG:

You look at the numbers that they're focusing in on of cases that we may have naturalized improperly, and you take the total number of cases we adjudicated and the little bit of number they've got left over here, what is it? One-tenth of one percent? That's more pure than Ivory Snow, which is 99.9 percent pure. Okay? When you're dealing with two or three million cases . . .

He went on to say that "300 and something" cases that might lead to revocation were "small stuff." He asserted that while this was still not an "acceptable error rate," he was a "realist" who understood that "nobody, absolutely nobody, is perfect." Others, including EAC Aleinikoff, suggested that the low number of problem cases identified actually fell within INS' "tolerable error" range:

You know, at some point you are going to tolerate a small degree of error to accomplish an overall goal. I'm not sure any government process is 100 percent accurate as it goes through. What we wanted was a, you know, a tolerable degree of error.

Other officials did not describe the KPMG review findings as exonerating, but they nevertheless pointed to them to show that INS' procedures simply did not cause the degree of harm originally alleged. Project manager David Rosenberg told the OIG:

The review shows that a very small proportion of people who were ineligible on the underlying criminal offense, very small, it will probably be, when the final numbers are in it will certainly be less than 500. . . .Now those are not the numbers to be proud of, but the fact is as you look as an error rate percentage, it's very low for any benefits.

He went on to say that he did not "expect a pat on the back," but he believed that the OIG needed to show that "all of these potential and real vulnerabilities did not lead to a wide open hole." Similarly, Associate Commissioner Crocetti told the OIG, "one could argue that it may not have been as bad as it could have been when you look at the actual percentage of cases [requiring revocation]."

In making such assertions, these officials used narrowly defined criteria to assess the degree of harm wrought by INS' flawed processing procedures. For them, it was ultimately the number of ineligible persons who in fact became citizens that determined whether INS had failed in its mission. Because the percentage of such cases was quite low, according to what they understood of the KPMG-supervised review, they expressed some vindication. We found that their assessment of the KPMG findings was consistent with their basic attitude in approaching CUSA—that naturalization applicants were not a population likely to have extensive criminal histories.

However, solely the number of presumptively ineligible persons who became citizens should not measure the integrity of CUSA's criminal history processing procedures. Instead, these procedures should be evaluated according to the gamble INS took by naturalizing people without much concern for what a fingerprint or bio-check might reveal. This degree of risk is illustrated by the number of persons INS naturalized without ensuring that a fingerprint card had been submitted, and by the number of persons with criminal histories whose rap sheets were not reviewed before their applications were granted.

The KPMG-supervised reviews begun in 1997 were designed to identify the number of persons naturalized despite having a disqualifying criminal record. Indeed, the INS officials quoted above seem to have interpreted the KPMG reports only insofar as they identified this "bottom line." However, examination of the results of the KPMG-supervised reviews offers little comfort to those who want to believe that CUSA did not create a "wide open hole" in INS' naturalization practices because only "one-tenth of one percent" of the naturalizations were compromised. On the contrary, the results of the reviews show an extremely high error rate resulted from INS' flawed procedures.

First, 71,413 persons naturalized by INS during CUSA were not even included in the universe of cases subject to the KPMG-supervised review because INS was unable to timely identify them as completed 1996 naturalization cases.400 Second, of the 1,049,867 persons originally believed to have naturalized during CUSA, 186,077—or 18 percent—never had a fingerprint comparison conducted by the FBI. INS officials who focus only on the number of persons whose records were later checked and found to be presumptively ineligible ignore the significant risk inherent in INS' failure to ensure that fingerprint checks were conducted for almost one-fifth of the applicants naturalized during CUSA.

What the KPMG reviews cannot tell us, however, is the number of times INS approved a naturalization application without first reviewing the applicant's criminal history report, even if in the end INS was fortunate that the criminal history by itself did not preclude naturalization. The reviews cannot determine in how many cases a criminal history report that by itself was not disqualifying might have led an adjudicator to ask questions that would have resulted in denial instead of approval of the application. To appreciate the degree of risk inherent in INS' faulty procedures, we examined how often applicants' criminal history records sat in district offices unreviewed and how often INS managers failed to get these records to the adjudicators before the naturalization interview. The record shows that INS ran a significant risk in every Key City District during CUSA.

INS officials' focus on what they interpreted as the low number of ineligible persons naturalized during CUSA was not a new position they took during the OIG investigation. When Congress and the media began to scrutinize the CUSA program in August, September, and October 1996, senior INS officials offered the same defensive response to congressional inquiries. Instead of acknowledging INS' flawed fingerprint processing procedures—as shown in the OIG and GAO reports of 1994—INS officials, on these occasions including Commissioner Meissner, consistently turned the debate toward what little proof existed of how many ineligible persons had, in fact, been naturalized. In so doing, they understated the nature and degree of INS' many errors in criminal history processing of which they were aware.

Our review of the events of August through October 1996 illustrates the most serious flaws of INS' criminal history checking procedures and the unreliable nature of INS' reports to Congress about the CUSA program. During a period when Congress was pressuring INS to answer specific questions about naturalization applicants' criminal histories, INS learned about the extent to which applicant rap sheets had been ignored in the Key City Districts. As these errors came to light, however, INS Headquarters officials failed to provide Members of Congress with information that would have answered some of their questions.

In the discussion that follows, we first summarize the findings from the KPMG-supervised reviews. Next, we describe the evidence that corroborates the media reports and employee allegations toward the end of CUSA that INS was not reviewing applicants' rap sheets before naturalization. We review the testimony offered by senior INS officials to the Subcommittee on National Security, International Affairs, and Criminal Justice of the House Committee on Government Reform and Oversight in September 1996 and point out what was left unsaid about the errors of which INS Headquarters was already aware. We then discuss INS' flawed "Fingerprint Clearance Process Survey," the grossly inaccurate results of which INS provided to Congress in the hope of obviating the need for a more systematic review of FY 1996 naturalization cases. We found that INS' lack of care in reporting issues concerning the criminal histories of naturalization applicants reflected the same inattention to important detail that characterized the flawed criminal history checking procedures during CUSA.

    1. The results of the KPMG-supervised reviews

The KPMG-supervised reviews of criminal histories relating to citizens naturalized during CUSA resulted in more than 6,000 cases being referred to INS' Office of General Counsel for possible denaturalization proceedings, including the 369 cases that had been originally deemed presumptively ineligible.401 The review team based its analysis (and thus the resulting identification of 6,000 cases) on a review of those cases originally identified by INS as in the "universe" of CUSA cases, of which only 82 percent had actually undergone a full fingerprint comparison. Of the 1,049,867 persons originally identified by INS as naturalized during FY 1996, the FBI had deemed fingerprint cards from 124,711 "unclassifiable" and had no record of conducting a fingerprint check on 61,366 others.

      1. Case stratification

INS first had to identify the cases naturalized during CUSA before comparing these cases to FBI billing records and criminal history databases. This process was known as the "case stratification." According to KPMG's first CUSA-related report, the "Naturalization Review Case Stratification Report" dated July 28, 1997, the goal of the case stratification was to "identify all cases with potentially disqualifying criteria."

Case stratification required a comparison of data maintained by INS and the FBI. For INS, the information was extracted from both the Central Index System (CIS) and NACS. For the FBI, the information was extracted from its automated billing records and from its Criminal Justice Information System (CJIS), the FBI's repository of information on criminal and civil enforcement history. KPMG stressed in its Stratification Report that using these various databases to identify cases with criminal histories was difficult because of the "incompatibility between INS and FBI processes, data and automated systems." In addition, the report specifically noted that although the case stratification process relied exclusively on electronic matching of data, "INS and FBI record keeping systems were not designed to perform electronic matching" and that "controls were not designed into the systems to ensure data integrity across the systems."

INS' first effort to identify the universe of individuals naturalized between August 31, 1995, and September 30, 1996, resulted in 1,049,867 unique matches.402

Through matching efforts using FBI billing and criminal history databases, the KPMG-led team found that:

      1. Criminal History Case File Review

Once the case stratification process was completed, INS collected rap sheets and A-files "to determine if any naturalized citizens failed to establish good moral character at the date of naturalization" based on the applicant's criminal history or failure to reveal a criminal history. The results of this review were documented in a KPMG report entitled "Criminal History Case File Review" dated December 19, 1997.

The FBI provided the review team rap sheets for those naturalized persons who had undergone a full fingerprint comparison and for whom the FBI had some criminal history record. The review team sorted the rap sheets into one of three arrest classifications: rap sheets reflecting arrests for felonies and/or crimes involving moral turpitude (CIMTs), rap sheets reflecting misdemeanors, and rap sheets indicating administrative or immigration violations.

For the December 1997 Criminal History Case File Review report, the review team examined the files of the 17,257 cases in which the rap sheet indicated a felony and/or CIMT arrest.405 INS was unable to provide the files for 399 cases, so in the end 16,858 cases were examined in the criminal history case file review. Rap sheets that indicated arrests that were either non-CIMT misdemeanors or administrative violations were not included in the case file review.

If the review team determined that the FBI's criminal history report on the applicant did not refute INS' original decision to naturalize, the case was determined to be "proper." If the rap sheet or case file did not contain dispositions of arrests or other information necessary to make such a determination, the case was determined to "require[s] further action," as were cases in which the applicant failed to reveal his/her criminal history in the adjudication process.406 The third category, "presumptively ineligible," identified cases in which the information in the file or rap sheet demonstrated that the applicant should have been precluded from naturalizing because of his or her criminal history.

These KPMG-led reviews of the 16,858 cases resulted in the following findings:

Of the 5,954 cases that were deemed to require further action, 3,580 had rap sheets indicating crimes that could have rendered the applicant ineligible for naturalization. However, these cases needed further review because information about dispositions of arrests and other matters was not in the file. The remaining 2,374 cases concerned applicants' possible "failures to reveal" relevant criminal history. These cases, together with the 369 "presumptively ineligible" cases, were submitted to INS General Counsel's Office in order to obtain the additional information needed to make a more definitive judgment on whether denaturalization proceedings were warranted.

      1. Supplemental review

After initiation of the initial KPMG-supervised reviews, the Department of Justice obtained almost 2,600 additional rap sheets and other miscellaneous documents from INS offices that had been found in the Los Angeles, San Francisco, Miami, and Chicago Districts that had not been interfiled in naturalization files.407 In addition, 300 rap sheets had not been considered in KPMG's initial review because they had been "in transit" at the time the case history review began. The KPMG-supervised review team subsequently analyzed all of these additional records in 1998.

Of these approximately 3,000 rap sheets and other documents, the review team found an additional 192 cases had been naturalized during CUSA even though the applicant had a record of a felony/CIMT. The team requested the A-files for these 192 individuals, but INS was able to provide only 167 of the files. Of these supplemental cases, 1 case was determined to be "presumptively ineligible," 126 "required further action," and 34 were determined to be "proper."408 The "presumptively ineligible" case and the 126 "requires further action" cases were referred to INS' General Counsel's Office for review.

    1. Evidence of unreviewed rap sheets in the Key City Districts in contrast to INS' reports to Congress

      1. Headquarters' August 1996 survey in response to media reports concerning the mishandling of rap sheets

In mid-August 1996, INS field offices began to receive inquiries from The Washington Post concerning alleged problems in INS' fingerprint processing, including a rumor that a field office had received rap sheets after applicants were naturalized. In order to be prepared to respond to any such allegations, Associate Commissioner Crocetti directed Benefits Division staff to work with their colleagues in Field Operations to obtain information from the Field about whether applicants were being naturalized without fingerprint checks being completed and, if so, whether denaturalization proceedings were warranted. On August 15, 1996, Mary Ellen Elwood, an INS Headquarters Field Operations staff officer, sent a message via fax to the three Regional Directors and the same message by e-mail to the regional offices and other INS officials requesting information by the end of that day concerning The Washington Post's inquiries. Regional offices were directed to respond to Stella Jarina, who was then the Acting Services Branch Chief in Field Operations.

By August 19, 1996, all five Key City Districts had responded in some fashion to Elwood's inquiry.

Thus, within a week, INS Headquarters had learned from the Field that more than 100 cases were under review for possible denaturalization proceedings because of an applicant's failure to reveal criminal history or because post-naturalization review of a rap sheet uncovered a potentially disqualifying criminal history report (10 cases in Miami, 40 in New York, 61 in Los Angeles). In addition, INS Headquarters learned through these communications about more than 750 other cases, not including the August shipment from the FBI to the VSC, whose corresponding rap sheets still needed review (686 in Los Angeles, 71 in the VSC, 10-20 in Chicago).

      1. The House subpoena

By September 1996, the media interest in possible problems with the CUSA program continued to grow. Prompted by these media reports and other accounts of fingerprint processing problems, William Zeliff, Chairman of the Subcommittee on National Security, International Affairs, and Criminal Justice of the House Committee on Government Reform and Oversight (hereinafter the Subcommittee), wrote to Commissioner Meissner on September 17 requesting information about the CUSA program. He wrote:

Specifically, we have reason to believe that in the cases of numerous applicants for naturalization, the INS has not waited a sufficient time for applicants' fingerprint reports and law enforcement records to be delivered from other agencies or offices and become part of the applicants' naturalization files. As a result, many applicants have been naturalized who have subsequently turned out to have criminal records, which should have precluded naturalization and which, in many cases, should have led to deportation instead.

Chairman Zeliff directed INS to provide to the Subcommittee by September 27 the name and other identifying information of each person naturalized since August 31, 1995, who had any felony arrests or convictions. On September 27, after an intervening hearing at which the Subcommittee learned of additional allegations of INS' mishandling of applicant rap sheets, Chairman Zeliff sent a subpoena on behalf of the Subcommittee requesting the information specified in his September 17 letter and requiring the information to be produced by October 1.

The Chairman's requests prompted significant discussion between INS staff and Subcommittee staff. Eventually, this subpoena and other congressional requests for information led to the systematic review of 1996 naturalization cases that was overseen by KPMG. Before the KPMG review, however, INS attempted to avoid undertaking the expansive review requested in Chairman Zeliff's September 17 request, proposing instead to substitute its own more limited review, the "Fingerprint Clearance Process Survey" that examined rap sheet processing since the FCCC began operation. INS had distributed that survey to the Field on September 19, 1996, requesting responses by September 25. The results of that survey as published to the Subcommittee on October 15, 1996, demonstrated that INS could not reliably provide Congress with information that accurately characterized the state of rap sheet processing during CUSA.

Before we turn to the analysis of the Fingerprint Clearance Process Survey and the inaccurate information INS reported to Congress, we first address the allegations made at the intervening congressional hearing. On September 19, the same day that INS distributed its fingerprint survey to the Field, the Subcommittee notified INS that Associate Commissioner Crocetti and CUSA Project Manager Rosenberg were to attend a hearing on CUSA scheduled for September 24, 1996.

      1. Allegations at the September 24, 1996, hearing

The Hearing before the Subcommittee on September 24, 1996, focused in large part on allegations of improper criminal history checking procedures. Among the INS employees who appeared as witnesses, two in particular—one from the Chicago District and the other from Los Angeles—testified about the naturalization of applicants whose disqualifying criminal history reports had not been reviewed.

The Chicago District DAO Joyce Woods testified that approximately 2,000-3,000 rap sheets stored in boxes in the Chicago District naturalization office had not been properly filed or reviewed. She told the Subcommittee that her own cursory review of the boxes revealed that some rap sheets pertained to applicants who had already naturalized while others related to applicants who were scheduled for naturalization. She also testified that some of the rap sheets reflected convictions that would render an applicant ineligible for naturalization.

According to Special Agent James Humble-Sanchez, the Los Angeles District also improperly naturalized persons without reviewing related rap sheets. He testified that after several August 1996 ceremonies in Los Angeles in which 60,000 persons were naturalized, "it was immediately known throughout the Los Angeles District that 5,000 criminal hits had come back."

The contemporaneous media reports about INS' criminal history checking procedures were also mentioned at the September hearing. In a Los Angeles Times article published on September 24, Los Angeles Deputy Assistant District Director Donald Neufeld was paraphrased as saying that after the August ceremonies the FBI had reported on arrest records for 200 of the newly naturalized citizens. The Times also reported that of the 200, INS said that 69 needs further review to determine if denaturalization was warranted. Congressman Mark Souder referred to the Times when he asked Humble-Sanchez to comment on the fact that INS officials had said only 69 cases in Los Angeles would warrant denaturalization. Humble-Sanchez replied that in his opinion the estimate was too low.

      1. INS' response at the September 24 hearing to allegations concerning the failure to review applicant rap sheets during CUSA

The Subcommittee sought answers from INS officials to the allegations of unreviewed criminal history reports. In his prepared statement for the September 24 hearing, Associate Commissioner Crocetti offered specific information concerning the number of "individuals . . . wrongly naturalized [60] due to fingerprint matches with disqualifying convictions reaching the files after naturalization has taken place." At the hearing, he testified that after "surveys to the field," INS "came up with . . . 60 for the entire naturalization program."

Crocetti's statement and testimony inaccurately suggested to the Subcommittee that recent inquiries of the Field had revealed only 60 such cases throughout the country, when in fact INS Headquarters had been advised about more than 100 cases in the Key City Districts alone that were being reviewed for possible denaturalization proceedings. In addition, INS Headquarters was aware that another 700 rap sheets were under review. Examination of Crocetti's testimony in the context of the information then available to INS Headquarters shows that he significantly understated the extent of INS' rap sheet processing problems.

In response to allegations that INS was naturalizing applicants with criminal records, Crocetti's written statement explained:

We are aware of only 60 cases of the nearly 1 million naturalization applicants processed this year where individuals have been wrongly naturalized due to fingerprint matches with disqualifying convictions reaching the files after naturalization has taken place. In each of these cases we are proceeding with de-naturalization proceedings.

During his testimony at the September 24 hearing, Crocetti elaborated on this point, saying:

With regard to the number of records, over 1 million cases—now we've had a lot of allegations in the field that criminals were being naturalized. We conducted two surveys to the field initially. They came in with numbers that were unsubstantiated. We went back to the field; we want to know for sure you have reviewed the file and made a decision with regard to those aliens not being eligible for the benefit? The number we have come up with—it is a few weeks old,412 but we're doing another survey as we speak—was 60 for the entire naturalization program.

By the September 24 hearing, only 13 of the 62 INS offices that were sent the survey—none of them CUSA offices—had responded to the September 19 Fingerprint Clearance Process Survey.413 In light of the fact that Field Operations staff had received reports that suggested higher numbers than those offered by Crocetti at the hearing, and because Crocetti had personally received messages in August 1996 that Los Angeles alone had 61 cases that needed review for possible denaturalization, the OIG asked Crocetti to explain his apparent underestimate of the scope of the problem in his testimony to Congress.

Crocetti acknowledged that when he had first asked Field offices about this issue the number was "definitely higher." He also acknowledged knowing that Los Angeles had reported that many rap sheets (he did not recall the specific number 416 until he was provided a copy of his e-mail message to Los Angeles officials asking how many of the 416 cases had been confirmed as warranting denaturalization) had not been reviewed before the applicants naturalized. He told the OIG that it was high numbers like those provided by Los Angeles in August that were "unsubstantiated," and he wanted to tell Congress only about cases INS knew were improperly naturalized. He explained that he sought to provide Congress with an accurate number of confirmed cases that warranted denaturalization.

Crocetti also told the OIG that when he testified on September 24 he based his remarks on information provided to him in a document prepared by Field Operations rather than on just information provided informally by INS staff. He said he "visually remember[ed] where it sat" on his desk.

However, as the OIG reminded him during his interview, Crocetti later told the Subcommittee, this time at its joint hearing with the Subcommittee on Immigration and Claims of the House Committee on the Judiciary on March 5, 1997, that the number of cases he had offered at the September 1996 hearing had been an estimate based on a rough doubling of the number of cases of which he was then aware that warranted denaturalization. At that March 5 Joint Hearing, when the Subcommittees learned of the initial results of the KPMG-supervised review, Congressman John Shadegg questioned Crocetti about his September testimony. Referencing the latest KPMG statistics, Shadegg asked, "a lot more than 60 people became citizens without complying with the law, correct?" Crocetti responded:

Primarily, I reported the number 60. As the Commissioner pointed out, at that particular time, we were canvassing [sic] the field as to the number of cases where they received a late fingerprint record, where there was a criminal record that resulted in disqualification, and that perhaps would warrant revocation proceedings.

At that particular time, the number that had been reported half-way through the review of the cases was in the 25 area. So being halfway through, I reported that it should not be any more than 60.

The OIG asked Crocetti whether, in light of this record, he had attempted in September 1996 to provide Congress with an estimate or with an exact number of cases of which he was then aware.

Crocetti acknowledged the inconsistency between characterizing his testimony as an attempt to give exact numbers and yet offering an "estimate." He again insisted to the OIG that he "used a number that was given to" him, but he did not recall if the document from which he obtained this statistic had reflected that 60 cases nationwide would require denaturalization, or whether the document had reflected only 25 cases and he used that smaller number to estimate the 60 cases. Referring to the estimate, he added:

I probably did that on my own. May have been discussing it with someone. I have visions of someone in my office talking about that I would do this because we didn't finish the report, and that kind of thing because we rarely got anything on time, and then we rarely got it right the first time. So I am sure I made the call. Whether somebody was there or not I don't know.

Crocetti was unable to produce and the OIG was unable to find any document that reflected the estimate to which he testified in September 1996. Nor did any document provided by INS or otherwise obtained by the OIG reflect that as of September 24, 1996, INS had conclusively determined that any CUSA case warranted denaturalization proceedings. The only evidence available at the time concerned cases that were under consideration for possible denaturalization, with hundreds more rap sheets under review.

The inaccuracies in Crocetti's testimony are troubling because, by his own description, he offered specific numbers to the Subcommittee that appeared to be the product of a survey of field offices together with a follow-up survey when, in fact, they were, at best, a casual estimate or a highly artificial and conservative count. The relatively low estimate also failed to convey to Congress the large number of cases of which he and other officials at INS Headquarters knew were under review across the country.

Crocetti asserted that his statements were offered in good faith. As Crocetti told the OIG, he wanted to "amplify the positive," but he had no intent to mislead. Perhaps the best summary of the confused record created by Crocetti is the one he himself offered. During an OIG interview, he reflected on what he had told the Subcommittee on September 24. He had testified that

[t]he numbers are exaggerated depending on how the people want to use the numbers or spin it in the paper or do whatever they want to do. There are two sides to every story. You can spin it as you want. But I can tell you categorically the numbers are not there and anyone who says it is, I challenge them to produce the numbers.

Although Crocetti denied in his interview with the OIG that he had tried to "spin anything downward," he said he was "countering allegations:"

I was trying my best to give them the closest number so they could appreciate it wasn't the thousands and thousands at that time that were being reported.

As the events of October 1996 unfolded, it became more and more clear that Crocetti's estimates to Congress had been inaccurate. Reports that Crocetti had called unrealistic and exaggerated in congressional testimony, such as the testimony by Joyce Woods from Chicago about the 2,000-3,000 boxes of unreviewed rap sheets or the Los Angeles report that 69 cases from their District would require denaturalization, were corroborated.

      1. INS confirms the allegations of unreviewed rap sheets

In the wake of the September hearings, INS corroborated many of the allegations that had been made by witnesses before the Subcommittee. As described below, INS itself met Crocetti's earlier challenge to "produce the numbers" of ineligible persons who were naturalized. A review team from the Chicago District and INS Headquarters confirmed that thousands of rap sheets were being stored in boxes in that District and had not been reviewed, as alleged by DAO Woods. More than a thousand of those records pertained to applicants who had already become citizens. In Los Angeles, even before the testimony of Special Agent Humble-Sanchez, Crocetti and Los Angeles officials were aware that at least 1,000 rap sheets had not been reviewed, 400 of which related to applicants who had already naturalized. Humble-Sanchez's specific testimony concerning "5,000 hits" arriving in the District after August ceremonies was not confirmed, but INS officials quickly learned that thousands of applicant rap sheets had not been reviewed before ceremony. Finally, in the Miami District, 900 rap sheets went unreviewed before the applicants became citizens.

        (1) Unreviewed rap sheets in the Chicago District

In response to DAO Woods' allegations, the Chicago District reviewed the boxes of rap sheets stored in its naturalization office. By October 1, 1996, the Chicago District confirmed that 1,378 of the 2,493 unreviewed rap sheets belonged to applicants who had already naturalized. Of the 1,378 persons whose rap sheets had not been reviewed before naturalization, the Chicago staff found that 301 required further review for possible denaturalization proceedings. A Headquarters "review team" led by then-Acting Assistant Commissioner for Naturalization Terrance O'Reilly traveled to the Chicago District on October 1, 1996, and confirmed the District review team's findings.414

The supervisor of Chicago's naturalization program summarized the District's problem as having been one of competing priorities during CUSA. Even before the revelations of September 1996, the supervisor told the OIG that the District had created a special team to work solely on fingerprint and rap sheet processing. They had similarly formed a team to focus on the processing of naturalization certificates, another part of the process in which they had fallen behind. However, Chicago managers had deemed the processing of certificates a higher priority after problems arose during a large naturalization ceremony at Soldier Field in May 1996.

        (2) Unreviewed rap sheets in the Los Angeles District

In addition to the more than 1,000 rap sheets that had not been reviewed at the time of the District's response to the August 1996 survey, Los Angeles continued to receive rap sheets in large numbers throughout August and September 1996. According to the District's own review in early October, staff had naturalized 69 persons during the last months of CUSA whose rap sheets had not been timely reviewed and who were potentially ineligible to naturalize because of their criminal histories. This finding was provided to INS Headquarters' Benefits Division by e-mail on October 16, 1996.

Additional reviews revealed numerous other fingerprint processing errors in the Los Angeles District during CUSA. For example, in September 1997 district officials identified 3,500 rap sheets that had sat untouched even during the months of case review that followed CUSA. Staff found that 1,500 of these rap sheets pertained to applicants who had naturalized during fiscal year 1996. Almost one-third of these records reflected felony arrests or criminal charges that could be felonies but disposition information was not specifically noted in the record. 415

        (3) Unreviewed rap sheets in the Miami District

The problems that surfaced in the Miami District, unlike those in Los Angeles and Chicago, were not ones that had been expanding over many years, but were ones that developed during CUSA. Given the volume and pace of naturalization production in the summer of 1996, the Miami District could not keep up with the incoming criminal history reports from the FBI. Although the naturalization manager increased the staff devoted to the sorting and review of incoming records from the FBI, some records arrived in Miami after the applicant's naturalization interview and some after naturalization. In addition, even the augmented staff was not sufficient to keep up with the FBI records they received before the interview or ceremony.

We found that rap sheets from the FBI that either had not been timely placed in the applicant's file or had been received after the applicant's naturalization ceremony were stored in the Miami naturalization office. During a September 1996 review, Miami District officials discovered 908 rap sheets in this office that had either arrived late in the District or had simply not been reviewed.416

    1. The Fingerprint Clearance Process Survey

We now examine in greater detail the Fingerprint Clearance Process Survey that INS undertook just after Chairman Zeliff's first request for information about the number of persons naturalized during CUSA who had criminal records.

As previously noted, the survey was issued on September 19 and responses were requested by September 25. In the wake of the September 24 hearing, as both the House and Senate pressed for information about newly naturalized citizens' criminal records, INS Headquarters officials began to view the results of the survey as the specific response it would offer to congressional inquiries. Therefore, Headquarters staff made follow-up inquiries of the Field after September 25 to confirm their reports. However, the survey was not completely responsive to the questions Chairman Zeliff was seeking to answer.

First, the survey was limited to rap sheets processed through the FCCC, which had only begun to forward records to the Field in July 1996. Second, the survey framed the question in terms of how many records INS had received after the benefit applied for, i.e., naturalization, had been granted. This shifted the focus of any blame about INS' failure to review rap sheets to the fact that CUSA naturalization had been moving swiftly—too swiftly, apparently, for the FBI to keep pace. However, asked this way, the survey question did not address the degree to which INS offices had failed to appropriately review those records that had arrived before interview and before ceremony. Finally, INS did not properly gather the data concerning the number of late-arriving "hits." As a result, INS provided Congress with inaccurate and misleading information about how it handled applicant criminal history reports during CUSA.

      1. The questions posed by the survey

The Fingerprint Clearance Process Survey was prepared by the Benefits Division (under Crocetti) and was signed and distributed by William Slattery, the EAC for Field Operations, on September 19. The survey, sent to all district offices, sub-offices, and service centers, posed three questions:

  1. Since July 1, when "hits" were forwarded to you by the [FCCC], how many matched records did you receive after the benefit had been granted?

  2. Of the cases identified in No. 1 above, how many warrant the initiation of proceedings to remove the benefit?

  3. Of the cases in No. 1 above, how many have yet to be reviewed so that a determination can be made as to whether benefit removal proceedings are warranted?

The survey noted that in terms of naturalization cases (the survey addressed fingerprint clearances for both naturalization and adjustments of status cases), the date on which the benefit was granted was the date when the applicant "actually naturalized." In other words, the survey did not attempt to capture the number of rap sheets received after an applicant had been interviewed as long as that rap sheet arrived before the applicant was sworn in.

Neither the documentary record nor INS officials involved in administering the survey could offer many details to explain the rationale for limiting the survey to only records processed through the FCCC. An earlier draft prepared by a staff officer within the Benefits Division did not include this restriction. That draft was addressed to Deputy Assistant Commissioner for Benefits Thomas Cook and was finalized by Crocetti, according to the staff officer who wrote the draft. When interviewed by the OIG, Crocetti and the staff officer who had prepared the draft and had been tasked with collecting the responses did not recall the specific reasons for the limitation, but surmised that the Field would not have been able to answer the question for records processed within their districts before the FCCC because INS had no Service-wide tracking system to reflect the date on which INS offices received responses from the FBI.

Our investigation confirmed that none of the Key City Districts during CUSA maintained a log listing the date on which rap sheets or other records were received from the FBI, even though the failure to date-stamp incoming records had been critically noted in the OIG's 1994 inspection report. Under the FCCC, however, the "fingerprint specialist" had been directed to maintain a log tracking the receipt of FBI responses, and Headquarters officials assumed that these logs could be consulted to produce the required answers to the survey.

If the survey is considered in the context of Crocetti's September congressional testimony, INS' restricted inquiry can be viewed as a way to produce positive results for INS. Implementing the FCCC had been a major change in how INS processed fingerprints, and this change was clearly expected to improve INS' performance. Commissioner Meissner, EAC Aleinikoff, and Crocetti repeatedly pointed to this innovation when addressing steps INS had taken to improve naturalization processing integrity under CUSA. The record suggests that the survey was designed, in part, to provide INS officials with data that would tend to rebut allegations of having failed to repair a system whose weaknesses had been apparent for several years.

As discussed earlier in this chapter, however, the FCCC did not bring the new level of order to fingerprint processing as Headquarters officials had hoped. By September 1996, fingerprint specialists in the Field still did not fully understand their role and "hits" were delayed en route from the FCCC to district offices. Nor had the Key City Districts maintained detailed logs that accurately tracked when "hits" were received in relation to the date on which the applicants became citizens. Accordingly, field managers could not simply turn to their records to quickly answer Headquarters' survey questions. As a result, the responses submitted by the Field were the product of various efforts to quickly count known "late hits." This ad hoc approach toward data collection thus contributed to the inaccuracy of the information subsequently reported to Congress.

      1. Reliance on the survey to answer congressional questions

INS proffered the results of its Fingerprint Clearance Process Survey to House and Senate subcommittees as Congress continued to investigate naturalization processing under CUSA. In a letter dated October 1, 1996, Commissioner Meissner informed Chairman Zeliff that producing the information requested in the House subpoena concerning all felony records of citizens naturalized during FY 1996 would be extremely burdensome and "would divert a significant amount of resources away from necessary [INS] operations." She informed the Chairman that INS Headquarters had directed its field offices "to review all naturalization cases in which they received a matched FBI record after the person was naturalized." She specified that the review would cover the period from July 1, 1996, through the end of the fiscal year (a 3-month period), although she did not explain the reason for this limitation other than to say that this was "when we centralized receipt of FBI records in our Nebraska Service Center." In addition, Commissioner Meissner advised the Chairman that INS field offices were reviewing the cases to determine whether denaturalization proceedings were warranted and whether such proceedings had been initiated. She then volunteered to provide the Subcommittee with "a summary of the results of this review."

During this time, the Senate also joined the search for information about the naturalization of persons with criminal records. In anticipation of an upcoming hearing before the Subcommittee on Immigration of the Senate Judiciary Committee (hereinafter the Subcommittee on Immigration), INS was asked to respond to questions from Senator Simpson's office about, among other things, its fingerprint processing procedures. Senator Simpson's staff posed questions about the number of persons naturalized before a criminal history report was received, the number of persons naturalized for whom the criminal history report had never been received, and the number of persons naturalized who had a disqualifying criminal record. In addition, a letter written to the Attorney General by Senator William Roth dated October 3 was forwarded to INS for response. Senator Roth expressed concern about the Citizenship USA program and specifically about allegations that applicants had been naturalized without adequate background checks.

In response to the Senate's inquiries, INS again referenced its Fingerprint Clearance Process Survey. In his prepared statement for the October 9 hearing before the Subcommittee on Immigration, EAC Aleinikoff noted:

We are currently conducting a survey o[f] our field offices to determine the precise number of cases in which the district has become aware of a disqualifying conviction after the individual has been naturalized. Preliminary indications are that only several dozen individuals out of the more than 1.2 million naturalization applicants processed this year have been wrongly naturalized.

EAC Aleinikoff did not specify that the survey was limited to the 3-month, post-FCCC time period.

Like Crocetti before him, Aleinikoff emphasized the positive view of INS' fingerprint processing procedures in his statement to the Subcommittee on Immigration. Although he noted that "in a very small number of cases, both in the past and recently [emphasis added], FBI record matches have been received after a benefit has been granted," he noted that the problem had been solved since implementation of the FCCC. He wrote, "through this centralization, we now know that 'hits' are reaching the district offices before the adjudication is finalized."

Aleinikoff was not asked any questions about INS' fingerprint processing at the October hearing before the Subcommittee on Immigration. However, Aleinikoff met with INS officials the morning of the hearing and directed that staff conduct teleconferences with the Field in order to ensure that the information requested in the Fingerprint Clearance Process Survey was produced along with supporting documents concerning affected cases. A letter dated October 7 from Chairman Zeliff reminding INS that the failure to respond to his earlier subpoena could result in the Subcommittee finding Commissioner Meissner in contempt had prompted this meeting.

      1. The results of the Fingerprint Clearance Process Survey

On the morning of October 10, 1996, the day after the Senate hearing, Deputy Commissioner Sale held two teleconferences with representatives of all district and regional offices to stress the importance of the Fingerprint Clearance Process Survey. Later the same day, Crocetti and Deputy General Counsel Paul Virtue met with Subcommittee staff who advised them of congressional concern that INS had naturalized "tens of thousands of criminals" because of its failure to review rap sheets. Subcommittee staff specifically directed INS to obtain the information requested in the September subpoena by electronically extracting from INS computer systems information that could be electronically compared to data maintained by the FBI.

On October 15, 1996, Commissioner Meissner wrote to Chairman Zeliff and advised him that INS had delivered to the FBI tapes containing the electronically extracted information. However, in her letter she attempted to assuage the Subcommittee's concerns by stating that the electronic comparison would not corroborate allegations that thousands of ineligible applicants had been naturalized. In addition, she provided the Subcommittee with the results then available from the Fingerprint Clearance Process Survey, saying that she was providing an "update of INS' ongoing effort to identify cases where a criminal record ('hit') was received after a person was naturalized." Commissioner Meissner advised that INS "field offices reported the receipt of 415 late hits" out of 30,422 total rap sheets received since the FCCC became operational. Commissioner Meissner promised that INS would provide the Subcommittee with the results of its review of the 415 cases but advised that "most of these cases" would not require denaturalization.

The Commissioner's letter was accurate in that 415 was the total number of late-arriving rap sheets that Headquarters staff had compiled based on reports from the Field. Her letter to Congress also specifically pointed out that this number was not the total number of unreviewed CUSA rap sheets, but rather was only the number for those "hits" that arrived post-naturalization since implementation of the FCCC. However, the Field had made numerous mistakes in compiling the survey responses, many of which should have been obvious to INS Headquarters officials who reviewed the reports. Furthermore, given what Headquarters then knew about failures to review applicant rap sheets over the course of the entire CUSA program, Commissioner Meissner's letter omitted a fact that would have been responsive to congressional inquiries: that thousands of applicant rap sheets had sat unexamined in INS offices throughout the Key City Districts during CUSA, and INS had no method for determining whether they had arrived late or had simply been ignored.

In the discussion that follows, we highlight some of the errors made by the Field and by Headquarters in compiling the Fingerprint Clearance Process Survey. We end our discussion of this issue by reviewing INS officials' response to questions posed by the OIG about this survey.

(1) Errors in the Fingerprint Clearance Process Survey

The chart below is reproduced from the INS document recording the response to the Fingerprint Clearance Process Survey.

Name Fingerprint Clearance Process Survey
Office Late Idents Proceedings Justified Removal Not Warranted Decision Pending
ATLANTA 0 0 0 0
Charlotte 0 0 0 0
BOSTON 0 0 0 0
Hartford 0 0 0 0
Providence 0 0 0 0
BUFFALO 0 0 0 0
Albany 6 1 5 0
CSC 8 N/A 8 N/A
CHICAGO 0 0 0 0
Indianapolis 2 0 2 0
Milwaukee 0 0 0 0
Cincinatti 0 0 0 0
DALLAS 0 0 0 0
Oklahoma City 0 0 0 0
DENVER 6 0 6 0
Salt Lake City 7 0 7 0
DETROIT 3 0 0 3
EL PASO 50 0 0 50
Albuquerque 0 0 0 0
HELENA 0 0 0 0
HONOLULU 1 0 1 0
Agana 1 0 0 1
HOUSTON 0 0 0 0
St. Louis 0 0 0 0
LOS ANGELES 69 9 0 60
MIAMI 63 0 26 37
Jacksonville 0 0 0 0
Orlando 0 0 0 0
Tampa 9 0 7 2
West Palm Beach 0 0 0 0
Louisville 0 0 0 0
Memphis 10 0 10 0
NEW YORK 64 10 18 36
NEWARK 31 1 30 0
OMAHA 0 0 0 0
Pittsburgh 0 0 0 0
PHOENIX 3 0 0 3
Las Vegas 0 0 0 0
Reno 1 0 1 0
Tucson 0 0 0 0
PORTLAND, ME 1 0 1 0
St. Albans 3 0 3 0
PORTLAND, OR 1 0 1 0
SAN ANTONIO 12 0 12 0
SAN DIEGO 7 1 5 1
Fresno 15 0 15 0
Sacramento 7 0 4 3
San Jose 0 0 0 0
SAN JUAN 0 0 0 0
Charlotte Amalie 0 0 0 0
SEATTLE 1 1 0 0
Spokane 2 1 0 1
ST. PAUL 0 0 0 0
TSC 10 N/A 10 N/A
WASHINGTON 22 0 22 0
Norfolk 0 0 0 0
TOTALS 415 24 194 197

We point out some of the errors listed in this chart and reported by Commissioner Meissner in her October 15 letter.

        1. Errors concerning the Chicago District statistics

First, the 415 cases reported by Commissioner Meissner in her October 15 letter omitted any records for the Chicago District. Given what INS Headquarters had already discovered about rap sheet processing in Chicago, in particular after the testimony of DAO Woods in September, this constituted INS' greatest error of omission in the letter to the Subcommittee.

As previously discussed, INS Headquarters had confirmed DAO Woods' allegations and found 1,378 unreviewed rap sheets that pertained to applicants who had already naturalized. Because the review team had described these rap sheets as relating to cases filed early in the CUSA year, we presume that they had not been processed through the FCCC. Therefore, while it was not literally inaccurate to omit news of the 1,378 Chicago cases in the report of the Fingerprint Clearance Process Survey (because the survey examined only cases processed through the FCCC), the omission is remarkable given the efforts Congress was then making to obtain full disclosure from INS.

In addition, the report INS Headquarters received from the Chicago District about the survey suffered from misinterpretation on the part of Chicago officials. They believed that the survey requested information about cases processed through the Nebraska Service Center under the Direct Mail program. This error Chicago officials were making was clear from the face of the document submitted to INS Headquarters (it specified that the numbers referred to Direct Mail cases), but no efforts were made to clarify it. Because the Chicago District had only begun to work on such cases the month before the survey, they reported receiving fewer than 20 "hits" after the interview had occurred but prior to naturalization.

        1. Errors concerning the San Francisco District statistics

San Francisco's response to the survey was inaccurate because it was rushed, District officials told the OIG. ADDA Still remembered that he checked with the DADDA and with the supervisory applications clerk before reporting to INS Headquarters that no late-hits had been received in the San Francisco District. Neither the DADDA nor the supervisory clerk told the OIG that they remember Still making such an inquiry, and both said they were aware in September 1996 that their office had received some records from the FBI after applicants' naturalization.

The San Francisco District Office (on behalf of San Francisco and its satellite office in Oakland) reported to Headquarters that it had not received any "hits" after naturalization. Meanwhile, at the Oakland satellite office, officers and clerks had been segregating late-arriving rap sheets as they arrived. These records were reviewed nine days after Commissioner Meissner's letter to the Subcommittee and showed that 45 records arrived late for applicants naturalized in August and September 1996, in the Oakland office alone.417

INS Headquarters also separately listed survey responses from the San Francisco District Sub-offices. San Jose was listed as reporting no "hits," although the OIG found no evidence that INS officials in San Jose had undertaken a review to answer the survey questions. Fresno and Sacramento reported 15 and 7 hits, respectively.

The fact that the two smallest sub-offices in the San Francisco District reported receiving late hits while the larger two offices reported none was not investigated or clarified by INS Headquarters.

        1. Errors concerning the Miami District statistics

In the Miami District Office late-arriving rap sheets were date-stamped with a naturalization-specific date-stamp and were segregated, not interfiled.418 Miami reported to INS Headquarters that since July 1, 1996, 116 rap sheets had arrived after the applicants naturalized, and of those cases 63 did not require further review for possible denaturalization proceedings.

Miami's 116 cases was the largest number of late-arriving hits reported, despite the fact that the New York and Los Angeles Districts had processed tens of thousands more cases. The Headquarters' compilation of the survey results mistakenly listed only 63 late-arriving hits for Miami instead of 116.

        1. Errors concerning the New York District statistics

Just as INS failed to reveal in the information provided to Congress significant problems it had previously confirmed about rap sheet processing in Chicago, INS Headquarters also failed to account for the fingerprint processing problems at the Vermont Service Center that had surfaced earlier in the summer. On several occasions, as discussed above, the New York District encountered processing errors by the VSC that resulted in applicants being naturalized without the rap sheets being available to the adjudicator.

In its calculation of the Service-wide total of late-arriving rap sheets, INS Headquarters staff used 64 cases for the New York District, although the New York District had actually reported 67 cases in which the rap sheet had been received after naturalization.419 This figure was reported by the SDAO in Garden City to whom clerks forwarded all rap sheets that arrived after applicants were naturalized for tracking and reviewing for possible denaturalization.420 District managers assigned this task to the SDAO after they discovered the fingerprint processing problems at the VSC earlier in the summer. According to the SDAO who received the rap sheets, he prepared a list with identifying information about the case, but he only tracked the case if, after pulling the file, he determined that the applicant had failed to disclose the arrest to the adjudicator. Since not all applicants denied having arrests, the SDAO necessarily underreported the actual number of late-arriving rap sheets for New York.421

        1. Errors concerning the Los Angeles District statistics

The Los Angeles District had maintained rap sheets that staff had not reviewed before the applicants naturalized. Of those, largely consistent with the number first reported to Crocetti in August 1996, Los Angeles officials determined that 69 cases warranted review for possible denaturalization proceedings.

However, Los Angeles managers and staff made a mistake in responding to the Fingerprint Clearance Process Survey by not focusing on the fact that the survey was limited only to records processed through the FCCC. These officials also did not recognize that the survey first asked how many "hits" arrived late and then asked, of those, how many might warrant denaturalization proceedings. Instead, Los Angeles officials responded to the survey's first question, "how many hits arrived after the benefit was granted?", with the number that represented cases that might warrant denaturalization.422

While Los Angeles officials made significant errors in responding to the Fingerprint Clearance Process Survey, staff at INS Headquarters should have quickly detected these errors. First, Los Angeles officials had told Crocetti and officials at the Western Region in August 1996 that since March 1996, at least 416 rap sheets had not been reviewed before applicants naturalized and that another 686 cases were then under review. Although the survey was limited to the post July 1 time period, the disparity between these original numbers and the Fingerprint Clearance Process Survey answers should have prompted further inquiry.

Second, on October 5, 1996, Los Angeles DADDA Neufeld sent Stella Jarina of the Office of Field Operations—the same person who served as the original point-of-contact for the survey conducted in August in response to The Washington Post inquiry—a "CUSA Fact Sheet" in response to Special Agent James Humble-Sanchez's testimony before Congress in September. In that fact sheet, Neufeld wrote, "of the 60,000 persons naturalized in the Los Angeles District in August of 1996, less than 200 had arrest records which were reported by the FBI subsequent to the naturalization ceremony." DADDA Neufeld had previously offered the same statistics to the Los Angeles Times, where his explanation was paraphrased that of the 200 "only 69 appear to have been involved in crimes serious enough to perhaps disqualify them as U.S. citizens."

Any one of these other reports should have alerted Headquarters officials to the mistakes Los Angeles District officials made in their response to the Fingerprint Clearance Process Survey.

When the OIG interviewed Los Angeles officials about rap sheets, we asked specifically how often rap sheets arrived late or simply went unreviewed until after the applicant's naturalization, regardless of their impact on the applicant's eligibility. DADDA Neufeld told us that no one had kept track of this information. He conceded that if we included among the "hits" rap sheets that only showed immigration violations, the District might have received as many as 5,000 "hits" with respect to applicants naturalized in the August ceremonies alone, as alleged by Special Agent Humble-Sanchez, but he did not believe that all of these rap sheets had been received after the ceremonies. ADDA Arellano disagreed with Special Agent Humble-Sanchez's testimony, calling it a "gross exaggeration," but nevertheless told the OIG that the more accurate number of "late" hits was approximately 2,000.

        (2) Headquarters officials did not recall the details of the Fingerprint Clearance Process Survey

By the time of the OIG investigation, the Fingerprint Clearance Process Survey results had been overshadowed by the intervening reports from the KPMG-supervised reviews. Every Headquarters official we interviewed who had a significant role in CUSA remembered that in September and October 1996 they had sought to provide Congress with statistics from the Field, but no one recalled the specifics of the survey or any efforts made to review its conclusions before the information was provided to Congress. Commissioner Meissner remembered that it had been a "collective activity," and that she had not specifically assigned responsibility to any one person. Deputy Commissioner Sale recalled that the numbers kept changing and that the survey was confusing. David Rosenberg did not remember any particular INS official who worked on the survey.

Because EAC Aleinikoff had been the INS official who had specifically sought follow-up information concerning rap sheets in the Chicago District, the OIG asked him about the apparent inconsistency of reporting only 415 late "hits" nationwide since implementation of the FCCC in light of information that thousands of rap sheets had been ignored in the Chicago District alone. His answer was "that's a good question." He then noted that this was a question that he could not answer. In explaining INS' failure to take into account the evidence from Chicago in compiling the survey results, EAC Aleinikoff said that "the numbers kept shifting" as INS "tried to get the best information" it could.

  1. Conclusion

Congress did not rely solely on information provided by INS Headquarters to evaluate the degree to which CUSA adjudications had been compromised by failures to review applicant criminal histories. By continuing to press for more precise information, Congress eventually received the more comprehensive picture provided by the KPMG-supervised reviews. Once those reviews became public, as explained at the beginning of this chapter, INS began to more fully admit that criminal history processing procedures during CUSA had been poorly administered. However, INS' explanations were incomplete even then. Instead of recognizing the multiple ways in which INS itself had poorly administered virtually every aspect of its criminal history checking procedures, INS pointed to its accelerated processing times and its flawed presumptive policy as the source of the mistakes.

As the record makes clear, INS' mistakes in properly administering applicant background checks go well beyond the pace of processing and the presumptive policy. Virtually every aspect of INS' fingerprint and bio-check systems was compromised during CUSA. INS had been warned by the GAO and the OIG about its vulnerable procedures, but failed to address the problems. The one belated innovation—the FCCC—was well-intentioned but undermined by its awkward and rushed implementation. Then, when called upon by Congress to explain its performance, INS failed to provide an accurate representation about the extent to which it had risked adjudication integrity by failing to properly administer its criminal history checking procedures. The evidence shows that INS' representations concerning CUSA processing, like the fingerprint checking procedures themselves, were wholly unreliable.

  1. Preliminary results revealed 113,126 unclassifiable cases (i.e., cases in which the fingerprints were not readable by the FBI) and 66,398 cases in which the FBI did not have a record of receiving a fingerprint card (referred to as "not found"). KPMG's final report on this review, discussed below, showed a slightly higher total of 186,077, or 18 percent, of the total number of naturalized persons (124,711 unclassifiables and 61,366 not found).

  2. A rap sheet is a response to a fingerprint check that shows a criminal history. These positive responses were also referred to as "hits" and "kickbacks."

  3. Since 1952 the Immigration and Nationality Act has specifically included a provision for the "personal investigation of the person applying for naturalization in the vicinity or vicinities" of the applicant's residence and workplace. That investigation could be waived by the Attorney General for certain individuals or classes of individuals. As of 1991, the Code of Federal Regulations provided that the district director could waive the neighborhood investigation of naturalization applicants. This regulation still requires INS "to conduct an investigation" and states that the investigation must include, at a minimum, "a review of all pertinent records" and "police department checks."

  4. The Code of Federal Regulations requires an applicant for naturalization to submit his or her fingerprints on Form FD-258 (a fingerprint card), and this requirement is incorporated into the instructions on the N-400.

  5. As of April 1997, the FBI had over 219 million fingerprint cards on file, which included the 132 million fingerprint cards of persons who had been arrested ("criminal cards") and nearly 87 million fingerprint cards of persons applying for licenses and employment ("civil cards").

  6. In addition to fingerprint impressions, FD-258s also contain name, date of birth, and other biographical data written or typed in the blocks provided in the upper portion, or "masthead," of the fingerprint card. This data was also used in the fingerprint process and could be submitted to the FBI manually on the FD-258 or via magnetic tape format (called "machine readable data" or "MRD"). The user fee charged by the FBI was less if the agency submitted the biographical data via magnetic tape format. Before and during CUSA, INS did not submit the biographical data via magnetic tape format and thus was charged the full user fee.

  7. The cost of the fingerprint check was incorporated into the fee paid by the applicant to INS and was $95.00 before and during CUSA. INS is authorized by Section 286(m) of the INA to charge an application fee.

  8. Before and during CUSA, INS directed applicants to entities outside of INS for fingerprinting. Local private vendors conducted the majority of applicant fingerprinting for a fee. Applicants could also be fingerprinted by local law enforcement agencies.

  9. As of August 1996, submitting agencies were directed to send cards to CJIS at its new location in Clarksburg, West Virginia. Until that time, CJIS had been located in Washington, D.C. During the spring of 1996, the FBI had begun to transfer its personnel from the Washington location to West Virginia and some INS officials believed this contributed to the increasing processing times at the FBI. See our discussion of FBI processing times and INS' response, below.

  10. Fingerprint cards that were missing an ORI code or had an invalid ORI code were returned to INS Headquarters and then sent to the regional offices for distribution to the Field.

  11. This name-check process should not be confused with the bio-check process introduced above and discussed later in this chapter even though INS referred to both checks as "name-checks." The name-check process was a search for the applicant's fingerprints in the FBI's criminal history record database to determine if the applicant had a criminal arrest history based on the applicant's name and other identifying information contained in the masthead of the FD-258. For the bio-check process for naturalization applicants, INS submitted information to the FBI's Information Resources Division (a different division from CJIS) and to the CIA. Bio-checks consisted of a search of the FBI database known as the Central Records System, which contained information from FBI and other law enforcement agency investigative files, for such information as terrorist and organized crime activities. INS also had an interagency agreement with the CIA to conduct searches in a similar database that contained information on investigations of counterintelligence activities and other matters involving national security. The bio-check searches were based on biographical information provided by the applicant on the N-400 that was then generated automatically on tapes, or the information was manually provided to the agency on a separate form known as the G-325A.

  12. In 1989 the FBI limited manual comparison to the top two candidates because they had found that the chances of finding a match with the third or subsequent candidate were remote.

  13. Many INS officials throughout the Key Cities told the OIG that they believed that the FBI's notation, "no record based on name-check" on an unclassifiable fingerprint card meant that the applicant whose name was on the submitted card had been checked in the FBI database and no criminal history had been found for that applicant. This was a misunderstanding of the "name-check" process. For example, if the name-check process had revealed several potential matches for that applicant, but the applicant's card was of insufficient quality for comparison to the fingerprint cards on file, the FBI would return the card to INS with the notation "no record based on name-check," even though the applicant might have a criminal history. The FBI's notation "no record based on name-check" meant only that as a result of the name-check process the FBI could not confirm any match between the submitted card and the fingerprints the FBI had on file.

  14. According to Commissioner Meissner's testimony at the March 5, 1997, Joint Hearing before the Subcommittee on National Security, International Affairs, and Criminal Justice (of the House Committee on Government Reform and Oversight) and before the Subcommittee on Immigration and Claims (of the House Committee on the Judiciary), the designated waiting period during CUSA was 60 days. However, as described below, we found a lack of unanimity in the Field concerning the length of the presumptive period.

  15. The Code of Federal Regulations was subsequently modified to require INS to "notify applicants for naturalization to appear . . . for initial examination on the naturalization application only after the Service has received a definitive response from the Federal Bureau of Investigation that a full criminal background check of an applicant has been completed."

  16. Assistant Attorney General Stephen Colgate testified before Congress in March 1997 that the preliminary results of the KPMG-supervised review showed that INS had naturalized 179,524 persons whose fingerprint cards had been deemed unclassifiable by the FBI or for whom the FBI had no record of having received a fingerprint card. In addition, thousands of applicants were naturalized without the adjudicator having the benefit of the rap sheet in the file at the time of the interview because the rap sheet had never been filed or because it had not yet been received from the FBI. Although the KPMG-supervised review determined, for the most part, that denaturalization was not warranted, failure to have the rap sheet in the file at the time of the interview is significant. Later in this chapter we address more fully the thousands of rap sheets that were not in applicants' files when they were interviewed and INS' failure to accurately disclose the extent of this problem to Congress in October 1996.

  17. Instruction 245.2(d)(1) specified a 60-day presumptive period for a bio-check conducted by consular offices in conjunction with applicants in certain kinds of adjustment of status cases.

  18. Because the instruction is poorly worded, it could easily be misunderstood to apply only to bio-checks and not fingerprints. The confusion arises out of the fact that instead of referring to form FD-258 used in connection with fingerprint checks, it refers instead to Form G-325, which was the routing slip formerly attached to the FD-258. Because Form G-325A, Form G-325B, and Form G-325C are also discussed in the instruction and are used in connection with bio-check requests, the reference to Form G-325 instead of form FD-258 could lead one to conclude that the instruction applies to bio-checks only, especially since Form G-325 had not been used in connection with the FD-258 for several years. This confusion was apparent in Commissioner Meissner's November 1996 memorandum to the Attorney General explaining the history of the fingerprint process. In that memorandum, Meissner mistakenly explained that the language of the instructions indicated that they applied only to bio-checks when, in fact, the instructions were written to apply to bio-checks and fingerprint checks.

  19. If an officer requested new fingerprints from an applicant in an interview, once that card was sent to the FBI (or the newly completed card was resubmitted), the officer was to indicate "Neg 40" on the application and continue the case. If no record was received within 40 days, the applicant's record was presumed clear. The evidence indicates that this understanding of the presumptive period persisted well into CUSA, as demonstrated by a memorandum issued to adjudicators in December 1995. If an applicant was submitting a new fingerprint card at the time of the interview, the adjudicator was to mark the application "with the notation 'NEG-40' made on the I-468 and immediately send the chart to the FBI." One version of a processing cover sheet used for naturalization adjudications in Los Angeles during CUSA confirmed this instruction by a block clearly captioned "NEG 40," offering the adjudicator the choice of checking that box and continuing the case.

  20. Similar to the procedure used in the Los Angeles District, in instances in which the applicant had to provide a fingerprint card at the initial interview, Miami officers would complete the interview, approve the applicant and then hold the case and not pass it along for further processing until the 45-day presumptive period had passed. If no rap sheet was received during that period, the case was then scheduled for a naturalization ceremony.

  21. The Naturalization Section Chief told the OIG that this was an understatement of waiting times naturalization applicants were then experiencing.

  22. He specifically remembered that adjudicators had to make sure that the 60-day "clock" had run before adjudicating an application. Section Chief Neufeld had not been among those in Los Angeles, discussed above, who had believed the presumptive period was only 40 days.

  23. The document does not distinguish between cards rejected for masthead errors and those rejected for unclassifiability; however, it did encourage the Field to pay closer attention to the quality of the fingerprint card submissions, discussed below.

  24. As discussed in the previous chapter (and see our additional discussion, below), in March 1999, INS reported to Congress that in rechecking its data it found an additional 71,413 persons naturalized in FY 1996 that had not been previously reported. For these 71,413 cases, INS did not obtain the status of the fingerprint check. Rather, INS sought to determine only whether the person had a criminal history. Based upon matching of electronic data, INS in conjunction with the FBI determined that 3,656 of the 71,413 persons had a criminal history. INS reported to Congress on June 6, 2000, that it had obtained the rap sheets for these 3,656 individuals and that it planned to conduct a file review to determine whether any of these individuals were improperly naturalized.

  25. It is also possible that these cases were not found in the FBI's databases because INS never submitted a fingerprint card or because the FBI and INS databases reflected different identifying information for the same case. For a discussion of the inconsistencies between the INS databases and the FBI's billing records, see KPMG's Case Stratification Final Report, July 28, 1997, pages 3-4, discussed below.

  26. Instruction 105.10(c) provided in pertinent part: "When the FBI or CIA furnishes a relating record, advises that one exists or may exist, or returns a fingerprint card with the notation 'Fingerprints illegible', the material shall be stamped on the reverse by the field office to show date of receipt and shall be immediately sent to the operating branch for immediate attachment to the file. The operating branch shall also stamp the reverse to show date received."

  27. Because the fingerprint check generally would be the only step in the adjudication process that could qualify as the "police department check" required by federal regulation, failure to ensure that the applicant had a full fingerprint check was arguably in violation of the law. For unclassifiable fingerprints, the name-check by the FBI might be considered a "police department check" for the purpose of 8 CFR § 335.1. For fingerprints rejected for masthead errors, however, the FBI performed no check of any kind.

  28. When asked how an officer was to use his or her discretion in deciding whether to request a new fingerprint card, the San Francisco District ADDA told the OIG that the officer would consider the name, age, and socio-economic status of the applicant; a young, adult male of lower socio-economic status might trigger further inquiry by the officer. As another supervisor put it, if the applicant was "an old lady," it would be unlikely that he would run a further check. This evaluation was recognized by officers as a kind of "profiling" of the applicant.

    The Acting Officer-in-Charge (OIC) of the San Jose Sub-office told the OIG that profiling was used in the San Jose Sub-office since the time of the previous OIC, but only when an officer was deciding whether an expired fingerprint card or check should be supplemented by a new one. He said that nationality was not a factor in such an evaluation, but method of entry to the United States, age, occupation, and gender were. He pointed out that even if he decided not to request a new set of fingerprints, he would nevertheless run a local law enforcement check to ensure that there was no reported criminal activity by the applicant.

    During the summer of 1996, this idea of profiling expanded in San Jose. In an effort to ensure that they had enough applicants ready to naturalize for the large ceremony in September, San Jose CUSA officials increased their rate of post-auditing temporary files. At the same time, their confidence that they had received any related "hits" in a timely fashion had begun to erode. They therefore wanted to check applicants whose files were being post-audited in the local law enforcement database before making final approval. The supervisor said that they used "profiling" to determine which applicants' criminal histories should be checked in the local law enforcement database. This type of profiling was based on the assumption that applicants from certain countries were often involved in particular crimes. A permanent DAO in the San Jose office said that using such criteria was an unwritten rule, an understanding to which they were led in training and in practice by supervisors who instructed adjudicators to look for particular kinds of crimes in the backgrounds of persons from certain countries. This DAO declined to discuss the profiling conventions because she did not want to "sound prejudiced."

  29. At least one SDAO reported that the New York District followed INS Headquarters' rule of allowing the adjudicator to determine whether to resubmit the fingerprint card at the time of interview.

  30. Once the case was data-entered by clerks, both fingerprint cards were stripped from the application; one was sent to the FBI and the other was placed in the applicant's file. When the fingerprint card was sent to the FBI, the N-400 was stamped with the date and a code indicating whether the case was for the Manhattan or the Brooklyn office.

  31. The March 1994 memorandum had asked that the districts report, among other things, local procedures that were used when fingerprints were unclassifiable.

  32. If the computer systems indicated that the A-file was located in another section of the district and that the application pending was for naturalization, clerks forwarded the rejected card to the naturalization section.

  33. According to one long-time clerk in the main file room, if an applicant's A-number was incorrect and the file could not be located in the computer systems, the rejected fingerprint card was destroyed.

  34. One SDAO, however, reported that the clerks simply routed the rejected fingerprint cards to the files and it was the responsibility of the DAO to resubmit the card.

  35. For a discussion of the problems that Miami and other districts were experiencing with clerical and records resources in the years leading up to CUSA, see our discussion of records resources in our previous chapter on file policy.

  36. Applicants are required to answer the question "have you ever been arrested?" on the N-400, and a rap sheet could either corroborate or contradict the applicant's written answers or explanations offered at interview. Providing "false testimony to obtain any benefit from the [Immigration and Nationality] Act" precludes the applicant from establishing good moral character and thus makes him or her ineligible for naturalization.

  37. For a discussion of these rap sheets, see section F of this chapter below.

  38. For a discussion of Los Angeles District's use of temporary files before the CUSA program, see our chapter on A-file policy and practices.

  39. During congressional hearings in September and October 1996, specific allegations arose concerning the fingerprint and rap sheet processing practices of the Los Angeles District. In addition, the two employees who made allegations concerning those practices were among the six INS employees who alleged that INS had retaliated against them for having cooperated with the congressional investigation (see our chapter on allegations of retaliation). Accordingly, we investigated Los Angeles practices in regard to fingerprint and rap sheet processing in considerable detail. In addition to the discussion we offer in this chapter, we address our findings concerning additional allegations about Los Angeles procedures in a separate appendix to this report.

  40. The OIG issued a report in February 1994 titled, "Alien Fingerprint Requirements in the Immigration and Naturalization Service." The GAO issued its report, "INS Fingerprinting of Aliens: Efforts to Ensure Authenticity of Aliens' Fingerprints," in December 1994.

  41. The KPMG Case Stratification Report shows that among CUSA applicants, the number of persons for whom FBI found a matching criminal history was quite large: 76,678 of the 844,044 classifiable fingerprint cards checked, or 9 percent. The OIG inspection report of February 1994 found that the rate of "idents" was 5.2 percent.

  42. Despite INS' assertion that interviews were postponed when a fingerprint card was resubmitted, as previously discussed, we found that it was common for adjudicators to interview the applicant without any delay and to assume that any rap sheet would be received prior to the ceremony.

  43. The inspection team's file review showed that in 170 cases where applicants had criminal records, applicants in 118 cases (69 percent) initially denied having been arrested.

  44. After CUSA, as a result of congressional inquiries and hearings, INS no longer downplayed the importance of FBI checks. In October 1996, Commissioner Meissner submitted a document to Congress noting that "local systems do not provide the comprehensive FBI information required by INS."

  45. The records sent to the Field were FBI billing records, discussed in greater detail below. Despite INS Headquarters' indication that billing records were a useful method to check the reliability of fingerprint procedures, the March 29 memorandum indicated that the billing records would not be provided to the Field routinely because to do so would be "unduly cumbersome."

  46. The memorandum directed the Field to respond within 24 hours. The record does not indicate the reason such a short response time was imposed.

  47. INS later asserted that the on-site inspections were conducted, referring to the "INSpect" program funded in late 1995. See our discussion of INSpect, below. Because the first inspections under this program did not occur until approximately two years after the release of the OIG's inspection report, such inspections do not seem to have immediately resulted from INS' promise in March 1994 to monitor practices in the Field.

  48. According to then Deputy Commissioner Sale, who had vetoed the previous recommendation to eliminate fingerprint checks in late 1993 while serving as Acting Commissioner, she was out of town when the recommendation was raised again in March 1994 and was taken to Commissioner Meissner for approval. Sale told the OIG that she did not become aware of the decision until she saw opposition in the media and Congress. Commissioner Meissner told the OIG that she did not recall that INS had actually ordered an elimination of the fingerprint check requirement but instead recalled that INS managers had proposed to take such action. While Commissioner Meissner said she did not recall implementing a definitive decision to eliminate fingerprint checks, she recalled that whatever decision had been made, it had been reached without debate; the matter had been viewed as a budget issue and not a policy issue. Commissioner Meissner described INS' actions in connection with this decision as "sloppy work" and conceded that it was not well thought out.

  49. The fact that INS planned to continue bio-checks did not mitigate the harm in eliminating the fingerprint check because, as discussed further below, the bio-check process involved a search of an altogether different set of records than the fingerprint check. The bio-check was a search of information about an applicant's connection to a federal investigation or about arrest records in a foreign country. The fingerprint check was a search of criminal arrest records in the United States.

  50. Former Associate Commissioner Weinig told the OIG that the reason for elimination of the fingerprint check was directly related to its increased cost. Faced with the decision to reduce staff or cut back on fingerprint checking costs, Weinig chose the latter. He reasoned that INS did not seem to care much about the fingerprint check, so he decided it could be eliminated.

  51. On April 11, 1994, Headquarters had solicited suggestions from the Field for "practical and immediate improvements" to the fingerprint process.

  52. The memorandum prepared by Deputy Commissioner Sale officially creating the FEWG was signed by Crocetti, then Acting Associate Commissioner, as the "Designated Process Owner."

  53. The FEWG considered whether "selective screening techniques" should be used to reduce the number of fingerprint cards submitted to the FBI. Based on the fact that INS spent $15 million annually for close to a million fingerprint checks but received a "hit" in less than 10 percent of the cases, the group believed that INS needed to consider "more cost effective means." At the same time, the FEWG also found that while INS' Adjudications/Examinations Program may have determined through experience that certain groups of cases did not warrant a fingerprint check, INS needed to "establish a sound statistical means for demonstrating this trend to itself." Accordingly, the FEWG recommended that INS develop "a statistical system for verifying low risk categoires [sic] of applicants" and that once this was accomplished, INS would be able to "carefully design screening techniques that would stand scrutiny by Congress, the law enforcement community and legal interests."

  54. INS had also been directed by the Senate Appropriations Committee, in its report of July 14, 1994, to implement a fingerprint process in which INS would accept fingerprints from applicants only if they were taken by authorized entities, including trained INS employees, recognized law enforcement agencies, or certified outside entities.

  55. As a result of this legislation, INS established new INS offices known as "Application Support Centers," or "ASCs," specifically for fingerprinting applicants.

  56. A notation of an "ident" or an "unclassifiable" would alert the adjudications officer that a hard-copy response (either a rap sheet or a confirmation of "no record based on name-check") needed to be located if it were not already in the file.

  57. An applicant's record may not have appeared in the billing records for other reasons such as failure by the FBI to record or receive the card or if the FBI and INS had different identifying information for the same applicant.

  58. According to the FEWG report, the FBI projected that in 1998 it would bring on-line its "Integrated Automated Fingerprint Information System" (IAFIS), which would allow for automated processing of digitized fingerprints transmitted to the FBI by scanning of the fingerprint cards or "live-capture" of fingerprints. According to media accounts, the IAFIS program became fully operational in August 1999.

  59. For example, INS could have used the billing records to track statistics about the percentage of fingerprint checks resulting in rap sheets and thereby confirm or dispel the notion that such checks were relevant in only a small number of naturalization cases. Also, INS could use the data to determine which districts were submitting the highest number of unclassifiable cards, information that INS could then use to target specific offices for reminders to resubmit new fingerprint cards with the unclassifiable card in order to avoid incurring a new FBI charge for the same applicant.

  60. INS was required to respond to the GAO report by sending a letter to the House Committee on Government Reform and Oversight. This process was coordinated by INS' Office of Internal Audit (OIA), which obtained responses from the Office of Programs and the Office of Field Operations. Using these responses, a letter for Commissioner Meissner's signature was prepared for Congress.

  61. In the Office of Field Operations' recommended response, EAC Slattery recommended that INS advise GAO that it would be able to obtain all fingerprint responses once the FBI had implemented the Integrated Automated Fingerprint Identification System.

  62. The reengineered CLAIMS was eventually ready for testing in the summer of 1997 and was introduced in the Field at the Nebraska Service Center in December 1997. Field installation continued in 1998 and 1999.

  63. Although the practice of marking the application continued sporadically at NDEC, it was no longer considered the governing indicator of whether the applicant's fingerprint check had been initiated.

  64. Although ensuring that the fingerprint check was completed before naturalization was better than not conducting any check, the efficacy of the system depended on having in place a mechanism to prevent from naturalizing "approved" applicants whose criminal histories arrived after interview but before the ceremony. As discussed above, not only did Los Angeles District not get criminal history records to adjudicators in time for interview, it also did not have a reliable system during the height of CUSA for pre-ceremony screening.

  65. O'Reilly had understood that the presumptive period was 45 days. He told the OIG during his interview that he still believed the presumptive period during CUSA had been 45 days. He said that while this was its length in the mid-80s when he last worked in naturalization, he conceded that it might have changed and he "just missed it." The 1994 OIG inspection report described INS' presumptive period as 60 days, but O'Reilly told the OIG that he had received but never read the report.

  66. The theory behind this NACS function suggests that it was designed to prevent the administration of the naturalization oath until a presumptive period had passed. That design—waiting only 45 days and using the ceremony date rather than the interview date as the deadline by which INS would have to receive an applicant's criminal record—was consistent with what many long-time INS employees told the OIG was their understanding of INS presumptive policy.

  67. The OIG asked site manager O'Reilly about Los Angeles' efforts to ensure that fingerprint cards from NDEC applications were checked before interview. O'Reilly said he was not certain, but he assumed ("from logic") that those 60,000 applications that had been stripped before NDEC began were actually data-entered first, by employees at Los Angeles, before the applications whose cards were stripped at NDEC. Because NACS was designed to schedule cases for interview in order of data-entry, the approximately 60,000 cases would therefore be scheduled before those that had their cards stripped more recently. At a rate of interviewing only 500 cases per day in Los Angeles at that time, it would have taken a long time to reach the cases that had only recently been stripped.

    Although O'Reilly's theory does offer a scenario of how the Los Angeles CUSA project could have properly administered the presumptive period for NDEC cases, there is no evidence that the data-entry process distinguished between those applications whose fingerprint cards had been stripped sometime before the project and those applications whose cards were stripped at NDEC. We found no witness or document indicating that Los Angeles had segregated the stripped applications from those that had not been stripped. Most of the Los Angeles officials we interviewed, including O'Reilly, remembered that cases were stored in boxes according to the date on which they were filed. The cases that had already been stripped were not in chronological order—in other words, Los Angeles had not stripped only the "oldest" or "newest" cases, but had stripped some in some months and not in others. O'Reilly himself did not pay attention to which of the cases had been stripped and which had not, and did not pay attention to the order in which the cases were data-entered. Other INS officials in Los Angeles were similarly inattentive to processing details concerning fingerprints.

    Furthermore, even if O'Reilly was correct in his recollection about such a processing step, the computerized distinction between the two types of cases would have been overridden by the frequent alterations that INS made to schedule cases at certain times in certain locations—for example, at the new Laguna Niguel site instead of downtown Los Angeles—for applicants who lived closer to a particular office. By providing applicants this convenience, INS could not ensure chronological processing district-wide. In addition, as previously discussed in our chapter on interviews and adjudications, INS' computer system did not guarantee the chronological processing of cases. Finally, as noted below, evidence concerning scheduling history provided to the OIG by INS as part of the current investigation shows that cases data-entered at NDEC as late as September 1995 were scheduled for interview at Laguna Niguel as early as the following month.

  68. The details of this "hold" procedure as envisioned by NDEC officials are described in our appendix that specifically addresses practices in Los Angeles District during CUSA.

  69. The computer only protected against making such cases "available" for automatic scheduling of interviews. Once "available," the case could be scheduled at any time.

  70. This computer entry had not been made for these cases because no one had told the data-entry staff that it was required.

  71. We infer that Cook took action because of complaints he received from Los Angeles project managers, but none of the witnesses involved remembered the details.

  72. Cook believed that by removing the security flag protection, cases previously data-entered would become available for scheduling. He was mistaken because this programming change would only affect cases data-entered in the Los Angeles District after the flag was removed. IRM understood what he was trying to achieve, however, and so not only removed the flag for future cases, but made adjustments in the computer system so that all cases entered previously also would be available for scheduling.

  73. The OIG did not discuss this matter with Alice Smith, so it is unknown what consideration, if any, she gave Stanley's memorandum. Smith's role during CUSA was to focus on building the districts' partnerships with CBOs and did not include any supervision of fingerprint processing or other data-entry tasks. Any effective review of "alternate controls" at NDEC would have had to include either O'Reilly, Neufeld, or Rosete, none of whom was aware of any need to double-check Los Angeles' fingerprint procedures.

  74. On October 9, 1996, Rosemary Jenks from the Center for Immigration Studies, a non-profit research institution, made numerous allegations about abuses in the CUSA program during her testimony before the Senate Judiciary Committee's Subcommittee on Immigration. She submitted a lengthy prepared statement with more than 100 pages of attachments that were made part of the hearing record. In her prepared statement, Jenks alleged that after learning that a data-entry processing error concerning fingerprint checks was preventing approximately 84,000 cases (correct number was approximately 85,000) from being scheduled for interviews, INS initiated an automated solution to get the cases scheduled immediately and thus bypassed the presumptive period.

    Although the allegations made by Jenks implied that INS' processing error had affected cases in all districts, our investigation determined that it was accurate only in regard to cases in New York and Miami. However, as explained below, 70 percent of the 85,000 cases were processed in these two districts.

  75. Among the allegations raised by Rosemary Jenks of the Center for Immigration Studies was that service center personnel were insufficiently trained in regard to fingerprint processing before implementation of Direct Mail and that this lack of training contributed to later processing errors. Our investigation has sustained this allegation.

  76. What guidance did exist also was contradictory. For example, the SOPs included a processing sheet that was to be used to note each step as it was completed that mistakenly indicated that fingerprints were required for applicants younger than 14. Of course, as written on the N-400 instructions form and in the Operations Instructions, fingerprints were required for applicants older than 14.

    We also note, however, that the SOPs were not the only text that gave inaccurate guidance concerning fingerprint processing. The N-400 instructions noted that a fingerprint card was required for all applicants older than 14 and younger than 79, while the Operations Instructions said that a card was required for all applicants older than fourteen but younger than 75.

  77. As discussed in other chapters of this report, while the service centers already used the CLAIMS system to data-enter information from other applications, additional software was required to accommodate data-entry of naturalization applications.

  78. This data-entry step was the one omitted at NDEC that resulted in the scheduling problems INS experienced in October 1995, as discussed above.

  79. In January 1994 INS had adopted regulations addressing applicants who never submitted required evidence, such as fingerprint cards, with their applications, and the regulations imposed a maximum time of 12 weeks or 84 days for an applicant to submit what had been requested by INS. Notices sent to applicants indicated that they were given an additional three days' mailing time for a total of 87 days. The regulations specified that applicants who failed to comply in this amount of time could have their applications denied for failing to do so.

  80. If the data-entry clerk indicated in the "Evidence" screen that the fingerprint card was not submitted, the computer automatically generated a letter to the applicant indicating that the applicant was required to submit a fingerprint card in order for processing of his/her application to continue. A similar procedure was followed if the applicant failed to submit photographs. In such cases, the data-entry clerk would continue to process the case, but while the case would upload to the CLAIMS mainframe, it would not upload to the NACS mainframe. If the fingerprint card was not submitted and the case was not updated within 87 days, the case uploaded to the NACS mainframe and became available for scheduling so that INS could bring in the applicant and deny the case. Ironically, however, there was no automated distinction between cases sent to the district office that were to be scheduled and denied as opposed to simply adjudicated. For this information, INS field offices still relied on manual indications on naturalization processing sheets.

  81. The CLAIMS software originally developed for Direct Mail permitted the scheduling of a case as soon as this second data-entry step was performed. When INS added file-ordering and the processing of fingerprint cards to the service centers' responsibilities under Direct Mail (for a discussion of this change in the role of the service center in the theory of Direct Mail, see our previous discussion in our chapter on A-file policy, above), CLAIMS was revised to include the 60-day clock.

  82. Fernanda Young, the Assistant Commissioner for IRM for benefits computer systems, told the OIG that it was not within IRM's authority to decide whether training was needed in the service centers and that any such instruction for IRM to provide training would had to have come from Service Center Operations within the Benefits Division.

  83. The SOPs did not provide any information concerning the data-entry requirements.

  84. IRM reported that 84,857 cases were affected: 37,694 for New York, 22,030 for Miami, 15,031 for Chicago, and 10,102 for Los Angeles.

  85. According to the FBI, INS was the only agency it worked with that did not keep track of its submissions.

  86. In addition to the random sample discussed above, Aytes had originally considered auditing the affected cases. The audit would not have impeded their adjudication. It would have involved retrieving the affected cases and comparing them to the fingerprint processing results once available from the FBI. However, as Aytes explained to the OIG, based upon the representations of the service centers that fingerprint cards were being submitted regularly to the FBI, he determined that the audit was unnecessary.

    In her October 18, 1996, letter to Senator Simpson in response to Rosemary Jenks' allegations, Commissioner Meissner wrote that "audits were requested" of the "cases that were not matched," presumably after comparison to the FBI data. In fact, though Aytes had entertained the audit idea and had faxed case lists to the service centers so that certain cases among the 85,000 could be audited,he later dropped this idea. Accordingly, Commissioner Meissner's letter precisely noting that "audits were requested" should not be read to mean that audits actually were conducted.

  87. In the second random sample conducted on August 27, 1996, 500 cases were checked for the CSC and 36.6 percent were matched; 500 cases were checked for the NSC and 27.8 percent were matched; 1,210 cases were checked for the TSC and 43.9 percent were matched; and 2,604 cases were checked for the VSC and 37.0 percent were matched.

  88. The IRM documents about this systems change do not specify by name the Benefits Division official who directed it.

  89. Cases for which the "Evidence" screen indicated that no fingerprint card had been submitted were included in the sweep but were to be placed on hold while a request for a fingerprint card was generated.

  90. Beginning in May 1996, the Miami District sent these cases to the TSC for data-entry by TSC staff on an overtime basis over a number of weeks. The TSC completed data-entry by mid-June. These 25,000 cases are referred to as Miami's "second frontlog."

  91. A "case release" refers to cases already data-entered; turning the clock or "flag" off would affect cases data-entered in the future.

  92. The INS Headquarters IRM employee who participated in the discussions about releasing cases and removing the flag said that it was not possible to manipulate particular cases. Cases could be released or a flag removed only for a particular office on particular days.

  93. Miami DAOs were instructed that, ideally, the 45-day presumptive period would have run by the time of interview. For cases data-entered in the Miami District, the DAO looked for a date-stamp on the N-400 that indicated "MIA-CITZ-[date]." Because the TSC did not use a date-stamp or indicate the date that fingerprints were sent to the FBI, DADDN Watson contacted the TSC and was told that the fingerprint cards were sent out the same day that the N-400s were data-entered and that it was reasonable to use the "clock-in date" in NACS as the first day of the presumptive period. Consequently, DAOs were instructed to check NACS for the "clock-in date" and to use that date in calculating the presumptive period. If the presumptive period had not lapsed at the time of the interview, DAOs were instructed to interview the applicant, continue the case, and place a note on the outside of the file that indicated "hold for prints" or that otherwise indicated that the presumptive period had not been completed. DAOs then placed these files in a box that was kept separate from the cases continued for other reasons, such as for reexamination or for receipt of the A-file. Some DAOs told the OIG that they would put their approval stamp on the application but would place the file in the "continued" box. Other DAOs reported that they interviewed these applicants and continued the cases but did not approve them until after the presumptive period had passed. One DAO recalled that when she first began in the spring 1996, the policy was to approve the cases and place them in the "hold for prints" box but that policy later changed during CUSA to simply continue the case. DADDN Watson recalled that prior to CUSA experienced DAOs would likely put their approval stamp on a case that had been continued in order to wait for the fingerprint check and that this may have continued for a while during CUSA. She stated that DAOs were later instructed to continue the case without approving it. While some supervisors reported that clerical staff were responsible for determining whether the presumptive period had expired before scheduling the cases for final ceremony, other supervisors reported that DAOs were responsible for reviewing the files before ceremonies to determine whether the approval was appropriate. We found that not all DAOs were aware of or followed these instructions concerning calculating the presumptive period before completing an interview. Of the 95 DAOs interviewed in the Miami District, nine DAOs stated that when they interviewed an applicant they presumed that they were adjudicating the case with a file that was ready to be adjudicated.

  94. The Miami District did not maintain statistics on the number of persons interviewed, however, documents show that the District Office was scheduling 1,100 interviews per day, 6 days per week beginning in mid-June. Based on these figures, the Miami District Office interviewed approximately 100,000 applicants between June and September 1996.

  95. We note that the fingerprint cards for some of the 10,905 applicants (INS cannot say how many) interviewed in less than 60 days had been sent to the FBI a few weeks before data-entry, thus allowing the FBI 60 days or more to complete the fingerprint check in these cases. This happened because in May and June 1996 the TSC data-entered Miami's second frontlog of applications whose fingerprint cards had already been sent to the FBI. However, INS did not attempt to ensure that only frontlog cases were subject to the removal of the automated 60-day protection. INS Headquarters had directed IRM to release all cases for scheduling.

  96. A post-CUSA review of Miami's unmatched rap sheets by District staff revised this number to 1,400 for applicants who had already naturalized. For a more complete discussion of how these rap sheets were handled during and after CUSA, see the section of this chapter entitled "Unreviewed rap sheets in Miami District," below.

  97. The OIG was unable to determine the ultimate disposition of these cards.

  98. These fingerprint cards were not the rejected fingerprint cards discussed earlier that the VSC had failed to properly process. We were unable to determine how many fingerprint cards were received in this shipment to the New York District.

  99. On the same day, Chapman learned from visiting EDS contractors who had just come from the VSC that service center employees referred to unclassifiable fingerprint cards returned by the FBI as "negative responses," implying that VSC was not taking any additional action with respect to these cards.

  100. It is unknown what happened to the crates of fingerprint cards sent to the New York District. Given Chapman's understanding that they should have been sent to the FBI, it is presumed that New York District forwarded the cards to the FBI.

  101. Chapman said that Yates later assured her that the fingerprint cards she had received were duplicate copies to be included in the file and that the other cards had, in fact, been sent to the FBI.

  102. At the Joint Hearing on March 5, 1997, Congressman Shadegg questioned Crocetti about INS efforts to maintain the integrity of the fingerprint process during CUSA. In particular, Congressman Shadegg asked Crocetti about a July 1996 e-mail message he had sent to Assistant Commissioner Aytes asking, "What in the world is going on here? This is a time bomb." Crocetti mistakenly believed that the e-mail quoted by Congressman Shadegg had actually been about the bio-check process and was unrelated to fingerprint processing.

  103. VSC Director Yates told the OIG that original estimates were that the VSC would receive 10-15,000 applications per month but that the VSC actually received approximately 35,000 monthly. Similarly, a manager for the contractor told the OIG that they had originally been told by INS to expect 500-700 applications daily, but that this amount had increased to 4,000 daily shortly after CUSA.

  104. As noted earlier in this chapter, positive responses to fingerprint checks and bio-check submissions were sometimes referred to as "hits."

  105. The NSC was staffed by Technology Systems Associates, Inc. ("TechSys"), a sub-contractor of Labat Anderson, Inc.

  106. The FCCC's Standard Operating Procedures were never finalized.

  107. The memorandum provided no instructions as to the purpose of the log or how to create a tracking system.

  108. According to records used by the FCCC, the total number of fingerprints processed by the FBI for INS adjudications-related cases in 1995 was 1,259,111. The total number of rap sheets, unclassifiables, and rejects reported was 294,523. Consequently, a 10 percent increase resulted in an estimate of 330,000 fingerprint responses.

  109. The document actually indicates 28 employees, but this is a mathematical error as it specifies 24 mail/file clerks, 2 data-entry clerks, 1 quality control inspector, and 2 supervisors, for a total of 29.

  110. The e-mail message actually states the incorrect figure of 28.

  111. In the meantime, Crocetti asserted in his September 24, 1996, prepared statement to Congress that the FCCC had been one of the "important improvements to the fingerprint process" that INS had made during CUSA.

  112. In most offices, staff put copies of the fax notifications in the files affected to alert adjudicators that a rap sheet would be forthcoming or that the applicant's fingerprints had been rejected. The Miami fingerprint specialist explained that because naturalization files were segregated from other files, he first had to determine the type of benefit that the applicant was applying for in order to know where within the District to route the notification.

  113. Section Chief Rose Chapman, whose office was in Manhattan, told the OIG that she understood that copies of the fax notifications were placed in the related files. When told by the OIG that SDAOs in Garden City and Brooklyn had reported that they simply posted the lists on the walls, Chapman indicated that she thought such a procedure would be pointless because adjudicators would not have had enough time to consult the lists before conducting interviews.

  114. The fingerprint specialist duties for the Los Angeles District were later reassigned to the DAO responsible for processing rap sheets and rejected fingerprint cards for the district. From witness accounts, this assignment likely occurred in August or September 1996. This DAO denied that he was the fingerprint specialist for the FCCC. He recalled that the previous fingerprint specialist simply gave him the duties associated with responding to the lists from the FCCC. He indicated that he did not act on the faxed lists and instead waited for the rap sheets and other responses to arrive and routed them accordingly.

  115. In the Sacramento Sub-office of San Francisco, the position was assigned to a DAO who immediately delegated the responsibility to a temporary applications clerk who had been working at the office for only six months.

  116. In fact, the contractor staff at TSC was unaware that cases could be put on "hold" in NACS to prevent them from being scheduled. Meanwhile, FCCC personnel interviewed by the OIG confirmed that FCCC staff did not place "holds" on any cases.

  117. INS was dependent on FBI estimates because, as noted previously, it did not track when it sent its cards to the FBI or when they were returned—data necessary if INS was to independently calculate processing times.

  118. Gilbert Kleinknecht, former Associate Commissioner for Enforcement, was the INS official who provided this information to Aleinikoff. Kleinknecht told the OIG that he had attended a meeting sometime earlier that year at which representatives of the Benefits Division were complaining about the length of time it took the FBI to process an applicant's fingerprint check. Kleinknecht suggested that they, INS, inquire of the FBI about processing times. Aleinikoff concurred, and asked Kleinknecht to follow up. Kleinknecht did not specifically recall the name of the person with whom he spoke about processing times, but he told the OIG he would have inquired through senior officials at the FBI. The INS Benefits Division officer, on the other hand, obtained his information directly from Donald Spitzer, the Operations Manager of the FBI's Criminal Justice Information Services (CJIS) Division, the division that conducted fingerprint checks.

  119. It is not clear to the OIG why the FBI's reported daily output in the April 18 e-mail was 47,000 cards while notes from Chang's late-March contact with the FBI indicate their reported daily output as 36,000 cards. The March data was provided by Donald Spitzer. The information reported in this April 18 e-mail came from someone other than Spitzer—"an FBI analyst," according to the documents.

  120. Although the OIG did not examine the degree to which INS' own submissions to the FBI caused the increasing backlogs at the processing facility, we asked INS officials what steps they took, if any, to alert the FBI to the anticipated increase in submissions INS would offer during fiscal year 1996. FBI officials told Congress and the OIG that they never received any such warnings.

    Assistant Commissioner for Benefits Michael Aytes, without offering specific details, told the OIG that "we went to the FBI and asked if they could handle the volume and we relied on that." No other INS witness recalled any specific communication with the FBI about this matter.

  121. According to FBI officials, INS could have requested specific information about the FBI's processing times, including the number of cases processed within given time periods, the percentage of cases processed within those time periods, the number of cases still in process, and the length of time that each card had been in process. INS did not request this information until September 1996.

  122. The statistics about fingerprint checks reported by Cook were in fact incorrect. The evidence shows that the processing times Cook was referring to were actually those for bio-checks, which were also then presumed to be processed within 60 days.

  123. There is evidence indicating that extending the presumptive period to 120 days was not immediately effective. The FCCC conducted a survey in early September of 128 returned cards and rap sheets and found that 25 percent took between 110 and 166 days to process. We also found evidence that the Field was confused by the policy change articulated in the September 27 memorandum. Field offices questioned whether the 120-day rule applied only to cases filed after September 27 or also to cases already interviewed and scheduled for interviews.

  124. In this same time period, 250,000 people were naturalized.

  125. For applicants seeking to adjust their status, INS also submitted bio-check requests to the Department of State for a check against its nonimmigrant visa application records.

  126. The fact that INS destroyed some bio-check responses from the FBI (discussed below) was only indirectly brought to the attention of Congress. One document submitted by witness Rosemary Jenks at the October 1996 hearings before the Senate Judiciary Committee's Subcommittee on Immigration was an INS e-mail indicating that INS had been "burning hits" for "8-10 years." Although the message was entitled "fingerprint checks," the "hits" to which it referred were actually bio-check responses.

  127. INS does not require a definitive response to the bio-check.

  128. As discussed in our chapter on interviews and adjudications, the Los Angeles District conducted an aggressive "off-site" or "outreach" processing program in which applications were submitted outside the usual data-entry procedures. Los Angeles outreach cases were scheduled manually by INS staff, and data-entry for these cases did not occur until after the applicant had been interviewed and either approved or denied. For these cases, therefore, NACS could not generate a bio-check request until after the interview. This failure to process outreach cases in NACS continued throughout CUSA and was not corrected by Los Angeles INS until the error was brought to its attention during an NQP audit in 1997.

  129. Offices without NACS were directed to submit the G-325As to the FBI via the INS Regional Offices, where the forms were checked for completeness and legibility to ensure that the FBI could process them.

  130. If an automated request did not include sufficient biographical data about the applicant, the request would be processed manually and, therefore, would be charged at the more expensive manual rate.

  131. This estimate also assumes that the automated requests submitted by INS were suitable for automated processing and thus cost only the lower amount. As is discussed below, many automated requests submitted by INS contained insufficient information to permit the FBI to conduct automated searches. For these deficient automated requests, INS incurred the higher cost of the manual search.

    It is also worth noting that INS records showed that it had submitted twice as many automated bio-check requests through NACS as were reflected in FBI records. INS reported that it had submitted 1,870,655 such requests, while the contractor, Maxima Corporation, reported 1,697,014.

  132. "Duplicates" were either names that appeared more than once on a tape, or names that had been submitted within the previous six months.

  133. Seventy-six percent of INS' requests for bio-checks were reported back to INS (on tape and on paper) within two to three weeks. These responses were typically "no record" responses. Since INS presumed after a certain time that an applicant did not have a record if a "hit" or "possible hit" had not been received and therefore did not use "no record" responses in the adjudication process, the FBI's processing times on these cases were irrelevant.

  134. As discussed below, each time INS failed to submit sufficient information to permit an automated bio-check, the FBI attempted to conduct a manual search. Each manual search cost INS $10.60 more than an automated check. In addition, INS submitted tens of thousands of duplicate requests. In FY 1996 alone, INS paid the FBI approximately $195,000 for duplicate checks of naturalization applicants.

  135. Under INS' presumptive policy, "no record" responses could be destroyed.

  136. An e-mail message submitted to Congress by Rosemary Jenks as part of the hearings into CUSA noted: "we are essentially destroying the hits on 1/2 our name checks (at a cost of approximately 1,000,000 a year) and naturalizing an assortment of scum." The message had been sent to Deputy Assistant Commissioner Thomas Cook by the staff officer who had been contacted by the Records Division regarding the destruction.

    As discussed below, the burning of the "hits" sent to the Records Division did not mean that the records of CUSA applicants were being destroyed. The Records Division was primarily destroying responses to hard-copy G-325A requests and other material unrelated to naturalization. For CUSA naturalization applicants for whom bio-check requests were primarily submitted by automated tape, INS failed to conduct timely bio-checks for reasons unrelated to this "burning" practice at the Records Division.

  137. We did not obtain the figures for duplicate submissions of bio-checks. However, the OIG's 1997 audit of fingerprint check and bio-check processes reported that during FY 1996, the FBI identified a total of 139,291 duplicate submissions related to INS' naturalization program. According to the report, INS was charged $195,007 (as noted above) for the processing of these duplicate submissions, assuming that INS was charged the lower amount (the price of an automated check, or $1.40) for processing each request.

  138. INS considered whether to incur the cost for the bio-checks when it was likely that applicants who had applied several months earlier had already naturalized.

  139. The OIG was unable to determine when INS actually submitted the tapes to the FBI.

  140. As discussed, INS eventually opened the FCCC in late June 1996 to receive and distribute all FBI responses for fingerprint and bio-check submissions. Also, OIA began the Field Assessment Program in the spring of 1996, but that review was broader than originally contemplated (and thus did not focus only on criminal history checking procedures) and did not result in a completed report until after CUSA.

  141. The officer, who was better informed about INS' bio-check procedures than he had been in December 1995, also noted that because of the lengthy processing time for bio-checks, it was "improbable" that INS was making the derogatory information available to the adjudicator in time to be considered. The report acknowledged that INS had no procedures for notifying the adjudicator about possible derogatory information about an applicant or for putting an adjudication "on hold" until such information could be received.

  142. As previously discussed, INS did not seek to determine whether the FBI conducted a fingerprint check of the additional 71,413 persons identified as having naturalized in FY 1996. INS in conjunction with the FBI determined that 3,656 of these individuals had a criminal history. As of June 2000 INS was planning to conduct a file review of those 3,656 cases.

  143. In July 1998, as part of the class action litigation in Gorbach v. Reno, the District Court in the Western District of Washington enjoined INS from initiating or continuing with the administrative denaturalization process. As of that date, INS General Counsel had reviewed 4,269 cases for possible revocation. Of those, 2,686 had been deemed appropriate for revocation. The Ninth Circuit Court of Appeals lifted the injunction in June 1999 but withdrew that decision in October 1999 when it decided to rehear the case en banc. As of June 2000, no further action had been taken in the case.

    In July 1998, because of the injunction, INS had directed its resources to the filing of appropriate denaturalization cases in federal court. A working group comprised of staff from the General Counsel's Office, from the Office of Immigration Litigation (OIL) of the Department's Civil Division, and from U.S. Attorneys' offices identified which cases should be brought first. As of July 1999, the group anticipated referring approximately 600 cases to OIL for denaturalization litigation.

    As of May 2000, the working group has completed its review of 400 of those cases. They have referred 110 cases to OIL for litigation. Of the 110 cases that have been referred, OIL has in turn referred 68 cases to the appropriate U.S. Attorneys' offices for litigation. The U.S. Attorneys have authorized litigation in 27 of the 68 cases. Four people have been judicially denaturalized.

  144. The discussion here is based on the analysis conducted by the KPMG-supervised review team of the cases originally identified as the "universe" of CUSA cases. That universe, as we explained in the previous chapter, was later determined to be based on an inaccurate count of cases. The total number of CUSA cases was revised from 1,049,867 to 1,109,059.

  145. This classification refers to unprosecuted immigration violations.

  146. Several possibilities could account for the cases in this category: INS never submitted a fingerprint card to the FBI, the card was rejected by the FBI because the masthead data was incomplete and the card was not resubmitted, or there were discrepancies between INS and FBI records.

  147. When the KPMG-supervised review first began, rap sheets were sorted into three categories—felony arrests, misdemeanors, and immigration violations—and files were requested only for cases in which the rap sheet indicated a felony arrest. Later, the team reviewed all misdemeanor rap sheets for potentially disqualifying arrests for "crimes involving moral turpitude" (CIMT). Based upon this review of misdemeanor rap sheets, the team determined that there were 301 additional cases for which the rap sheets indicated possibly disqualifying "CIMT conditions." The 17,257 files that were requested for the criminal history case review include the 301 additional misdemeanor CIMT cases.

  148. As discussed in a previous chapter, applicants were asked to indicate on the N-400 whether they had ever been arrested or convicted of a crime. Adjudicators typically also asked this question in the interview.

  149. In KPMG's "Supplemental Case Review" report, these additional rap sheets are described as having been discovered by the OIG. In fact, as noted in our previous chapter, the rap sheets were gathered by INS officials and it was INS officials who delivered them to the review team. Among the additional rap sheets reviewed were those discovered in Los Angeles in 1997 as described in this chapter, below (see "unreviewed rap sheets in the Los Angeles District"), and in our appendix on Los Angeles Criminal History Checking Procedures.

  150. In five of the six remaining cases the individuals identified on the rap sheet did not match the individuals noted in the case file. The one remaining case involved an INS administrative violation and therefore was outside the scope of the review.

  151. As discussed below, by October 1996, the San Francisco District employees confirmed that offices within the District had naturalized at least 65 persons before receiving or reviewing their criminal history reports.

  152. Although not specified in the memorandum, we presume that these criminal events were not disqualifying.

  153. It is unknown what overlap, if any, exists between the group of naturalized applicants with criminal records referred to by Chapman and the 71 persons identified by the VSC.

  154. Crocetti told the OIG that the review he referred to as "a few weeks old" during the hearing was the canvass coordinated by Elwood that he had requested from Field Operations and Benefits in August 1996.

  155. The 13 offices responding at that point had identified only two cases warranting denaturalization.

  156. According to the KPMG-supervised review, there were 4,285 rap sheets for cases adjudicated by the Chicago District during CUSA. The 2,493 unreviewed rap sheets represented more than half of the District's total for the year.

  157. For additional details concerning fingerprint and rap sheet processing in Los Angeles District during CUSA, see our appendix on Los Angeles criminal history checking procedures.

  158. As discussed in relation to the Fingerprint Clearance Process Survey, below, because INS offices often failed to date-stamp incoming rap sheets, it was often difficult to distinguish between those rap sheets that an office received late and those that an office simply reviewed late. Miami District was an exception and date-stamped incoming rap sheets as they arrived in the naturalization section.

    Miami officials eventually provided to INS Headquarters copies of the rap sheets that had accumulated in the Miami District Office and they were made part of KPMG's supplemental review, discussed above. The OIG randomly selected 100 of these 908 rap sheets for review and obtained a NACS printout for these cases. The OIG review shows that the records set aside in Miami included rap sheets that Miami received late and ones that Miami simply reviewed late. Of the 100 files reviewed, 97 of which were unique records, 62 had naturalization dates in July, August, or September 1996. Of those 62 cases, 43 showed a Miami date-stamp to indicate the date of receipt. Of those 43 cases, 16 had receipt date-stamps prior to the naturalization date. The remaining 27 cases had date-stamps after the naturalization date.

  159. Two of the Oakland cases (A27 756 155 & A91 738 648) were later identified by KPMG-supervised reviewers as applicants who were presumptively ineligible for naturalization.

  160. While the OIG does not question the diligence of the Miami District Office officials' efforts to accurately determine the number of cases with late-arriving rap sheets, we point out that in the OIG's review of 100 rap sheets, as noted above we found that 26 did not have a date-stamp.

  161. In our review of documents produced for this investigation, we found no indication that the number originally reported was amended by the New York District.

  162. According to Section Chief Rose Chapman, she inquired about late-arriving rap sheets to the SDAOs at both the New York District Office and Brooklyn citizenship sites and was informed that there were not any.

  163. The SDAO told the OIG that he could not recall how many late-arriving rap sheets he handled in which the applicant had disclosed the arrest.

  164. In response to the second question, "of the late 'hits,' how many needed to be reviewed for possible revocation of the benefit granted?", Los Angeles reported "9." However, "9" was the number of cases Los Angeles had conclusively decided warranted revocation, while the remaining 60 cases were still under review.