The Limits of Literal Truth as Defense to Perjury (or Even 18 USC 1001) (7/24/10)

Bronston v. United States, 409 U.S. 352 (1973), here, is frequently cited for the proposition that literal truth is a defense to a charge of perjury. Actually, Bronston involved an unresponsive literal truth to a question not asked with the answer intended to mislead (or at least avoid the question that was not asked). Usually, the battle ground is not the unresponsive answer where the defendant raises Bronston as a defense to perjury but the responsive answer that the defendant argues is literally true. Consider this example based on United States v. Thomas, 612 F.3d 1107 (9th Cir. 2010), here, involving a charge of perjury from a Q&A in a grand jury room.

Q. Did the drug dealer give you the drugs?

A. No.

The context leading up to this crisp Q&A is clear that Government attorney used the word give in a general sense – just a transfer and not just a transfer without any consideration (drug dealers do not generally give away drugs). The grand jury witness is charged with perjury for that answer. The Government proves the drug purchase and the delivery of the drugs. Is truth a defense? No. Even though the literal question without context may be capable of more than one interpretation and it is not impossible that the grand jury witness interpreted the question in the more technical gift sense, the jury is entitled to consider context in answering the ultimate question which is not whether the witness could have had some other understanding that would make the answer literally in his or her mind, but whether the witness actually had that understanding of what the prosecutor was asking. As stated by the Ninth Circuit, the issue is “whether the jury could conclude beyond a reasonable doubt that the defendant understood the question as did the government and that, so understood, the defendant's answer was false.”

The question then arises is the relevance of these notions to 18 USC 1001, false statements, which is more often seen in tax cases where allegedly false and / or misleading statements are made to an IRS agent or, sometimes, a prosecutor in a proffer session.  Some of the earlier versions of the DOJ Tax CTM actually seemed to read Section 1001 more expansively than perjury. Consider the following from the 2001 version of the DOJ CTM 24.04:

A statement is false for purposes of this statute even if it is a technically true statement, but it is knowingly put to a false use. In Peterson v. United States, 344 F.2d 419 (5th Cir. 1965) [here], in response to the question whether a payment was for past earned fees or fees to be earned, the defendant submitted a letter stating that his records showed the payment was an accrued fee, and accordingly, the payment was a deductible expense for a particular year. The court held that even if the literal language of the letter was true as to what the records reflected, it was clearly open to the jury to find that the statement in the letter as to the payment being an accrued fee was false. Peterson, 344 F.2d at 427. See also United States v. Brack, 747 F.2d 1142, 1150 (7th Cir. 1984), cert. denied, 469 U.S. 1216 (1985) ("even though the statements were accurate as to the total amount of the contract, they constituted false statements within the meaning of section 1001 by concealing the fraudulent nature of the contract").  Cf. Bronston v. United States, 409 U.S. 352, 358 n.4 (1973) (fraudulent statements include "intentional creation of false impressions by a selection of literally true representations") (citations omitted).

The 2008 version has this discussion of Peterson (CTM 24.05):

There, in response to a question whether a payment was for past earned fees or fees to be earned, a defendant submitted a letter stating that his records reflected that the payment was for accrued fees and that the fees were accordingly a deductible expense for the codefendant for a particular year. 344 F.2d at 427. On appeal, the Fifth Circuit held that whether the letter was true was a question for the jury and that even if the literal language of the letter was true as to what the records reflected, it was clearly open to the jury to find that the statement in the letter as to the payment’s being for an accrued fee was false. Id.; see also United States v. Brack, 747 F.2d 1142, 1150 (7th Cir. 1984) ("even though the statements were accurate as to the total amount of the contract they constituted false statements within the meaning of § 1001 by concealing the fraudulent nature of the contract").

The 2008 version, it seems to me, is less sweeping in its claims than the 2001 version.  And, I am not sure the courts interpreting Section 1001 would agree that it is not limited by the Bronston analysis.  Thus, for example, consider the following from United States v. Good, 326 F.2d 589 (4th Cir. 2003), here:

The principle articulated in Bronston holds true for convictions under Section 1001 and in this case today. See Mandanici, 729 F.2d at 921 (2d Cir. 1984); United States v. Moses, 94 F.3d 182, 188-89 (5th Cir. 1996) (“We cannot uphold a conviction . . . where the alleged statement forming the basis of a violation of section 1001 is true on its face.”); United States v. Vesaas, 586 F.2d 101, 104 (8th Cir. 1978) (“[A] prosecution for a false statement under § 1001 or under the perjury statutes cannot be based on an ambiguous question where the response may be literally and factually correct.”) (citing Bronston, 409 U.S. at 366). The defendant's response to the question was literally true, and thus, her indictment charging a violation of Section 1001 was properly dismissed.

Still, this is an area fraught with danger for the practitioner. It is not just the formal Q&A that gives rise to the danger, but can happen in the context of any IRS request for information. Consider an early stage of an audit. Since the agent may have not yet focused upon the particular facts, his or her questions may be unfocused, and the taxpayer and the practitioner (who do know the real facts) need to be careful how they respond. A gambit sometimes employed is to respond to the question that you would have preferred the agent to have asked – in Peterson, obviously, the taxpayer would have preferred that the question was “What do your books show” rather than, in effect, “What was the payment for.” Many practitioners will respond to this type of risk by rephrasing the question explicitly and then answering the question as rephrased. In other words, in Peterson, if the taxpayer had expressly rephrased the question to be “What do your books show” and responded, there should not have been a problem, because the agent (or other inquirer) could then, realizing that his question was not answered, repeat the question. I hope you will have picked up that this has the same essential characteristics of trials or hearings as in Bronston, where a person answers the question he or she would have preferred the examiner to have asked rather than the one asked. In an adversary hearing context, Bronston establishes that there is no problem where the answer is literally true even if misleading and, certainly, there would be no perjury problem in litigation if the answerer expressly makes clear which question he is answering. So too, it would seem that there should be no problem in the Section 1001, so long as the answerer makes clear that he is answering a question other than asked by the examiner. The key open issue under Section 1001 is whether the answer can violate Section 1001 in situations where he would not be held for perjury under 18 U.S. C. § 1621.

As noted, I would argue that in an audit context, which is inherently adversarial, the examiner’s responsibility is to pick up on shifts in focus of answers and follow through if he is not sure that he got a responsive answer. Nevertheless, in eggshell civil audits, taxpayers and their practitioners must be very careful as to what they say.

Addendum 7/28/10:
In a comment below, Anonymous cites the following article which I provide here for one click access:  Solomon L. Weisenberg, How to Avoid Going to Jail under 18 USC Section 1001 for Lying to Government Agents, Findlaw (2004) (although an older article, the points are still very good).

Addendum: 8/3/10
For a case similar in outcome to Thomas where the fight was over what the word "have" meant and holding that a jury of the defendant's peers can resolve that issue in context, see United States v. Gorman, 613 F.3d 711 (7th Cir. 2010), here.  The guts of the court's holding is (case citations omitted):

And in any event, our own determination of the meaning Jamarkus ascribed to "have" means little in this context, when Jamarkus was convicted by a jury of his peers. We agree initially with Jamarkus that "to have" has more than one meaning. See Webster's Third New International Dictionary 1039 (Philip Babcock Gove ed., 3d ed. 1986). When a word has more than one potential meaning, it must be examined in context to determine the meaning the defendant ascribed to it. As Jamarkus correctly points out, the fault for unclear, ambiguous, or vague answers rests with the questioner. But what Jamarkus ignores is that our precedent dictates that even when a question or answer is ambiguous, a conviction may still be upheld if a jury has been called upon "to determine that the question as the defendant understood it was falsely answered . . . ." 

What meaning Jamarkus ascribed to the question and whether his denial was knowingly false were questions within the jury's province. Id. The jury necessarily determined that Jamarkus understood "have" as signifying possession, and yet he readily and knowingly made a false denial. Under these circumstances, we conclude that any rational trier of fact could have found the elements of perjury beyond a reasonable doubt.