Eric M. Larson P.O. Box 5497 Takoma Park, Maryland 20913 (301) 270-3450 larsone@erols.com December 4, 1999 John W. Magaw, Director Bureau of Alcohol, Tobacco and Firearms 650 Massachusetts Avenue, N.W. Washington, D.C. 20226 Dear Director Magaw: This is in response to your letter of November 19, 1999, which addresses my complaint to President Clinton regardiing the failure of the Bureau of Alcohol, Tobacco and Firearms' (ATF) failure to correct serious errors in the National Firearms Registration and Transfer Record (NFRTR). You also deny that your staff misled you and the President about the seriousness of these errors. You make specific reference to my 1997 complaint to the Treasury Department Office of Inspector General (OIG), and to the accuracy and reliability of the NFRTR. Your statements are not supported by best evidence. In fact, your statements contradict the OIG's audit findings, and thus are legally incorrect. I have decided to bring your letter, and this matter, to the attention of the President and to appropriate Congressional committees which have oversight responsibility over ATF. You state: "I have reviewed the OIG reports and agree with my staff that most of your allegations are without merit." Your statement contradicts the OIG findings, which determined that 4 of my 5 allegations are correct, as noted in its October 1998 report. Thus, a correct statement would be that the OIG determined that most of my allegations have merit. Consider these facts: 1. I alleged that ATF has destroyed NFA documents. The OIG determined that persons working at or for ATF destroyed NFA documents, and that the number and type(s) of documents could not be determined. The OIG thus confirmed my allegation.[FN #1] [FN #1] I also made the more serious allegation that "during 1992 to 1996, ATF may have added 119 or more firearms to the NFRTR which were originally registered on Form 1 or Form 4467 during 1934 to 1971, for which ATF lost or deliberately destroyed the original records." I provided a statistical analyses that demonstrates that my allegation may be true, using NFRTR statistics provided by ATF. If my allegation is true, ATF would be required to establish another amnesty period. The OIG apparently did the audit work to address this allegation, but did not report the results in either of its reports. page 2 2. I alleged that ATF illegally registered firearms after the 1968 amnesty expired. The OIG determined that ATF registered nearly 2,500 firearms to individuals after the 1968 amnesty expired, and did not publish a notice in the Federal Register as required by law, which means the firearms were registered in violation of the NFA. The OIG further determined that 36 registration applications of a sample of 65 such applications (more than 55 percent) "did not have sufficient documentation to clearly show if the registrant met the criteria for amnesty." The OIG thus confirmed my allegation. 3. I alleged that two ATF employees committed felony perjury in letters to me regarding an NFA firearm classification issue. The OIG determined that these employees would not have committed perjury if a career criminal had been in "constructive possession" of an NFA firearm. In fact, the firearm in question was in the possession of another career criminal who claimed he was holding it as a favor, and nobody was ultimately charged with its illegal possession. The "if" nature of the OIG finding does not state that the career criminal was in constructive possession. The OIG thus confirmed my allegation by failing to determine that the career criminal ATF identified was in constructive possession of the firearm in question. 4. I alleged that some registration ATF classifies as "OTHER" may include firearm registrations that were contrary to law. The OIG determined that the foregoing registrations in some cases did not conform to authorized procedures, but that there was no evidence of illegal registrations. The OIG did not sustain my allegation, and I accept the OIG's findings. 5. I alleged that a large number of NFA firearms are currently registered to persons who are deceased. The OIG determined that more than 100,000 NFA firearms are likely now registered to persons who are deceased. The OIG thus confirmed my allegation. Your November 19, 1999, letter also states that you "disagree" with me that "the registration of NFA firearms to persons who may be deceased is a problem that seriously impacts the accuracy of the NFRTR." In its December 1998 report, the OIG seriously disagrees with you, stating: "Although the accuracy of the certification [of ownership] may not be affected, we believe that unless ATF acts to address this issue the number of deceased registrants will become greater as time passes and could adversely impact Congress' intention to control these weapons. ATF could be subject to further public criticism which would reflect negatively on ATF's efforts to administer the [NFRTR]." Further, your letter contains no specific statements identifying what ATF plans to do to address the issue of deceased registrants. Finally, you state that "the NFRTR continues to be an accurate and reliable database of firearms transactions." This statement is legally incorrect, because there is no basis for itin fact, the best current evidence indicates there are serious errors in the NFRTR, and that the NFRTR is not an accurate and reliable data base. Specifically, in its December 1998 report, the OIG describes the results of analyzing "discovery samples" (these are random, but nonrepresentative samples) for three NFA document categories page 3 (Form 4467, Letter, and Other), which constitute less than 4 percent of all documents. Under the standards for analyzing discovery samples, which the OIG implied that it followed, detecting more than 1 error in a discovery sample requires that a representative sample of the document base be drawn and the true error rate calculated. In fact, calculating the true error rate and its effect upon the audit is a requirement of Generally Accepted Government Auditing Standards (GAGAS), which the OIG stated it followed. Yet, after determining that each "discovery sample" contained more than 1 error (Form 4467, discovery sample size of 141, had 6 errors; Letter, discovery sample size of 179, had 15 errors; and Other, discovery sample size 202, had 16 errors), the OIG did not follow GAGAS by drawing statistically representative samples, calculating the true error rates, and determining the effects of these errors upon its audit. The OIG further states that its audit work . . . did not include a review of the accuracy of ATF's certifications in criminal prosecutions that no record of registration of a particular weapon could be found in the registry. We also did not evaluate the procedures that ATF personnel use to search the registry to enable them to provide an assurance to the court that no such registration exists in specific cases. Accordingly, this [audit] does not provide an opinion as to the accuracy of the registry searches conducted by ATF. Thus, the OIG does not give the NFRTR data base, or ATF's searches of it in preparing federal court certifications, a clean bill of health. Put another way, the OIG's statements cannot reasonably be construed to state that the NFRTR data are accurate and reliable. It is impossible to know what the true error rates are for Form 4467, Letter and Other document categories, because the OIG did not draw statistically representative samples and calculate the true error rates. The relevant finding from the OIG's analysis of its discovery samples, is that there is reasonable doubt the NFRTR data are accurate and reliable. Put it another way, the OIG's audit work indicates it is likely that the NFRTR may be inaccurate and unreliable. As demonstrated in my testimonies submitted to the House Subcommittee on Treasury, Postal Service, and General Government Appropriations, my analyses of NFRTR data that ATF has publicly released indicates it is extremely likely there are significant errors in Form 4467, Letter and Other document categories. Further, my analyses of most of the most common forms in the NFRTR (Forms 1 through 5, which constituted 2,184,474 firearm registration transactions, or 85 percent of the 2,571,766 reported by ATF for the years 1934 to July 31, 1998), determines it is likely there may be significant errors in each of these categories as well. Again, there is no comfort from the OIG's determination that it "cannot make any probability statement or draw any other conclusions concerning the more commonly used forms and corresponding data" in the NFRTR. page 4 Consequently, Director Magaw, there is no basis other than your personal opinion for your statement that "the NFRTR continues to be an accurate and reliable database of firearms transactions." The reason is that all of the most recent best evidence, which I have specifically identified, indicates that the NFRTR may not be accurate and reliable. I understand that ATF can't afford to admit that the NFRTR may not be accurate and reliable. Anyone who has ever been convicted of illegal possession of an NFA firearm was convicted for having something that didn't show as being registered to that person. What happens when it's proven that the NFRTR isn't accurate and reliable? Reasonable doubt happensthat's what. At this very late date, I again respectfully urge you to exercise leadership and direct ATF's employees to correct the errors in the NFRTR; and that you do so as soon as possible. Sincerely, (signed---Eric M. Larson) cc: The Honorable Orrin G. Hatch, Chairman, Senate Committee on the Judiciary The Honorable Henry Hyde, Chairman, House Committee on the Judiciary The Honorable Jim Kolbe, Chairman, Subcommittee on Treasury, Postal Service and General Government Appropriations The Honorable Dan Burton, Chairman, House Committee on Government Reform The Honorable William Jefferson Clinton