MORRIS SHEPPARD ARNOLD, Circuit Judge.

Pamela A. Habhab appeals her convictions on seven counts of mail fraud, one count of wire fraud, and one count of money laundering. We affirm the judgment of the trial court 1 in all respects.

Ms. Habhab formed a company to prepare magazines consisting solely of advertising. To sell advertising in the various Habhab publications, and thus to generate revenues for Ms. Habhab’s company, Ms. Habhab’s employees solicited the business of people who had advertised in competing publications. During a solicitation call, Ms. Ha-bhab’s employees informed the potential advertiser of a certain readership or circulation figure for a relevant Habhab magazine.

Following a complaint from a rival publisher, the United States Postal Inspection Service began an investigation to determine whether Ms. Habhab’s company misrepresented its readership or circulation figures to potential advertisers. After a grand jury indicted Ms. Habhab on charges including mail fraud, wire fraud, and money laundering, a bench trial followed, and the court found Ms. Habhab guilty of the charges at issue here.

Ms. Habhab argues that there is insufficient evidence in the record to support the trial court’s findings that the circulation or readership figures claimed for the relevant publications were false, that she authorized fraudulent representations, or that she committed unlawful activity sufficient to support a conviction for money laundering. Because this was a bench trial, we review the court’s findings of fact for clear error. Estate of Davis v. Delo, 115 F.3d 1388, 1393 (8th Cir.1997). A finding is clearly erroneous if it lacks substantial evidence to support it or if, even when there is evidence to support the finding, the reviewing court on the entire evidence is left with the definite and firm conviction that a mistake has been made. Day v. Johnson, 119 F.3d 650, 654 (8th Cir.1997). For the reasons that will shortly appear, we find that the trial court did not clearly err in reaching its factual conclusions.

II.

We believe that there is considerable evidence in the record to support a finding that the circulation or readership figures claimed by Ms. Habhab’s employees were false. Three of the counts of mail fraud on which Ms. Habhab was convicted arose from representations made about the circulation or readership of a magazine called Farm Market Directory. There was evidence that Ms. Habhab’s employees told three customers that circulation for that publication was 30,-000. The printer of this magazine, however, testified that he printed only 4,000 to 5,000 copies of it for each issue. A magazine for which only 5,000 copies are printed manifestly cannot have a circulation of more than 5,000 copies, let alone six times that.

Another count also arose from representations made about Farm Market Directory. In this instance, the relevant customer was informed as to the readership, rather than the circulation, of the publication. Readership may exceed circulation since one magazine can be read by several different people. There was testimony that Ms. Habhab’s employees told the customer that the magazine would be read by about 70,000 farmers. If, as the trial court reasonably found, only *413 5,000 copies were printed, each magazine would have to be read by fourteen people for the readership figures claimed by Ms. Ha-bhab’s employees to be accurate. The government offered evidence that the readership per issue for another magazine, Successful Farming, was 2.8 people and that readership-to-cireulation ratios ranging from 2.79 to 4.6 existed for other magazines. The defendant’s own readership expert offered an estimate that seven to ten people would read each copy of Ms. Habhab’s relevant magazines, still well short of the multiple of fourteen that Ms. Habhab’s employees claimed. The trial court could find falsity solely on the basis of the government’s evidence, but even if the court credited the defendant’s expert, it could still have reasonably concluded that the readership claim serving as the basis for the count was false.

Two of the counts of mail fraud of which Ms. Habhab was convicted were based upon representations made by Ms. Habhab’s employees who were selling advertising for a magazine called Computer Bargain Line. There was evidence that one advertising customer was told that copies of the publication would be mailed to 40,000 names on a mailing list, and there was testimony that another advertising customer received a,letter claiming that the publication would be sent out to a mailing list consisting of 52,000 names, with an additional 10,000 copies going to a trade show. Yet, the trial court reasonably found, no more than 5,000 copies of Computer Bargain Line were ever printed.

The remaining count of mail fraud arose from representations made- in connection with advertising sales for Ms. Habhab’s Travel America magazine. There was evidence that one of Ms. Habhab’s employees told a customer that the circulation for Travel America would be 50,000 per issue and that that customer later received a letter representing that Travel America had a readership of approximately 50,000 per issue. Whether Ms. Habhab’s employees represented circulation or readership here is of no moment, because, in either event, the claim would be false. Even assuming, as the trial court reasonably did, that seven people read each issue of the magazine, and that 5,000 issues were printed, the evidence would yield a readership of only 35,000.

The sole count of wire fraud on which Ms. Habhab was convicted arose from representations made in connection with advertising sales for Travel America. According to notes of her conversation with some of Ms. Habhab’s employees, an advertising customer was promised a mailing, or circulation, of 120,000 over four issues. While a defense witness offered testimony that she made claims concerning only readership, and not circulation, the trial court did not err in concluding that the claim was for circulation. A district court judge, of course, does not commit clear error merely because he or she does not credit a defendant’s testimony. On the basis of 5,000 copies of the magazine printed per issue, Travel America would have a four issue circulation of 20,000, not 120,000.

III.

Ms. Habhab maintains that the trial court erred in concluding that she authorized any of the fraudulent representations serving as the basis for her fraud convictions. Authorization was a key element in the government’s case for mail fraud and wire fraud because the government did not contend that Ms. Habhab herself made any fraudulent representations to the advertising customers. It is well settled in this circuit that, to be subject to criminal liability for false representations, the defendant must either make or authorize the representations at issue. See Pritchard v. United States, 386 F.2d 760, 766 (8th Cir.1967), cert. denied, 390 U.S. 1004, 88 S.Ct. 1247, 20 L.Ed.2d 104 (1968). The authorization may be either express or implied. As the Pritchard court recognized, material representations made in the presence of the defendant or by a person authorized by the defendant to make them are sufficient to support a finding of guilt. See id.

Despite Ms. Habhab’s contentions, there is substantial evidence of her implied authorization of fraudulent representations. In the first place, there is a great deal of testimony tending to show that her employ *414 ees made such representations, that, at times, Ms. Habhab used a telephone from which she could listen in on any of the telephone lines used by her employees, that she hired people to observe and listen to the employees on the telephone, and that these people would then report back to her on what they overheard. It is undisputed, moreover, that Ms. Habhab’s office was in close proximity to her employees.

While, as we have said, implied authorization can serve as a sufficient basis for criminal liability, the government’s case was buttressed here by the presence in the record of evidence of Ms. Habhab’s express authorization of fraudulent activity. According to one witness, Ms. Habhab approved every typeset document, including cover letters sent to customers. One cover letter included a fraudulent claim regarding the readership of Travel America. Other typeset documents madé additional representations regarding the circulation or readership of Travel America and Farm Market Directory. Yet another document, on which Ms. Habhab wrote “OK to type PH,” claimed circulation of 35,000 for Farm Market Directory.

We need not decide whether Ms. Ha-bhab specifically authorized each and every representation made by her employees, because there is sufficient evidence here from which the trial court could have reasonably concluded that Ms. Habhab authorized a common scheme of fraudulent claims regarding the circulation or readership of her magazines. Evidence of authorization of a common scheme of deception is sufficient to support a conviction for specific acts that were committed pursuant to that common scheme. See United States v. Cohen, 516 F.2d 1358, 1364 (8th Cir.1975).

IV.

Ms. Habhab also asserts that there was insufficient evidence to support her conviction for money laundering under 18 U.S.C. § 1956. To violate this statute, as relevant here, one must conduct or attempt to conduct a financial transaction that involves the proceeds of specified unlawful activity. Ms. Ha-bhab does not contest the trial court’s finding that a financial transaction took place on June 26, 1989. Instead, she challenges the trial court’s determination that specified unlawful activity took place prior to that date. Ms. Habhab further contends that even if such prior specified unlawful activity took place, the evidence was insufficient to support a conclusion that the June 26, 1989, financial transaction involved the proceeds of that prior activity. We disagree with both contentions.

While the evidence of prior specified unlawful activity is not overwhelming, it is sufficient to support the trial court’s conclusion, and our review of the record does not leave us with a definite and firm conviction that a mistake has been made. Employee Mike Moore testified that he started working for Ms. Habhab’s company in 1986 or 1987. His work for the company included communicating asserted readership figures to prospective advertisers. In 1987, he said, he claimed readership of 200,000 over four issues for the publication National Gun Sportsman. Another Habhab employee, Randy Sells, testified that he claimed readership of 37,500 per issue for a publication entitled Industrial Machine Connection. Such testimony, combined with the evidence of Ms. Habhab’s later activity (that is, the activity supporting Ms. Habhab’s convictions for mail and wire fraud), provides sufficient evidence for the trial court to have concluded that Ms. Habhab engaged in a scheme to defraud potential advertisers that started before June 26, 1989. Activity in furtherance of a scheme to defraud is, of course, specified unlawful conduct that can satisfy the requirements of § 1956. Cf. United States v. Massey, 48 F.3d 1560, 1566 (10th Cir.), cert. denied, 515 U.S. 1167, 115 S.Ct. 2628, 132 L.Ed.2d 868 (1995).

Nor did the trial court err in concluding that the June 26, 1989, financial transaction involved the proceeds of the unlawful scheme to defraud advertisers. The government need. not trace the laundered funds to a particular instance of fraud. See United States v. Blackman, 904 F.2d 1250, 1257 (8th Cir.1990). Furthermore, only part of the money involved in the transaction need be derived from the specified unlawful activi *415 ty. See United States v. Marbella, 73 F.3d 1508, 1514 (9th Cir.), cert. denied, — U.S. -, 116 S.Ct. 2555, 135 L.Ed.2d 1073 (1996). The trial court found that Ms. Ha-bhab’s advertising proceeds were placed in the bank account from which the check for the June 26, 1989, transaction was written. Exhibits admitted into evidence support that finding by detailing the income and expenditures for Ms. Habhab’s company during the period in question.

Ms. Habhab complains on appeal of three legal errors committed by the trial court. After careful consideration of her arguments, we find no error. First, Ms. Habhab contends that the trial court erred by reserving judgment on her motion, made immediately after the close of the government’s ease, for judgment of acquittal. To support this contention, Ms. Habhab relies on Fed. R. Crim P. 29(a) as it existed at the time of trial (the rule has since been amended to avoid the very difficulty at issue here). At the time of trial, Fed.R.Crim.P. 29(a) specifically required a court to determine the sufficiency of the evidence presented by the government prior to the presentation of the defense. The rule was apparently designed to prevent a defendant from being forced to present a case that might influence the court in its ruling on the motion. Whatever its rationale, we have held under the former rule that there is no prejudicial error in reserving a ruling on a motion under Fed.R.Crim.P. 29(a) “if at the time the motion was made the Government had produced sufficient evidence to justify submission of the case” to the finder of fact. United States v. House, 551 F.2d 756, 758 (8th Cir.), cert. denied, 434 U.S. 850, 98 S.Ct. 161, 54 L.Ed.2d 119 (1977), quoting Moore v. United States, 375 F.2d 877, 879 (8th Cir.), cert. denied, 389 U.S. 844, 88 S.Ct. 92, 19 L.Ed.2d 110 (1967) (emphasis supplied in House). Here, the trial court took the additional precautionary step of noting that it would not consider evidence presented by the defendant when it later ruled on the motion under Fed.R.Crim.P. 29(a) (as it existed at the time of trial).

Ms. Habhab argues that the government’s evidence as to readership was in fact insufficient at the close of the government’s case. We disagree. The government, during its case-in-chief, presented testimony on the readership issue. Former Habhab employee Mona Dillard testified that the average number of people reading a given copy of a magazine was four. An expert witness for the government testified that research showed a range of circulation-to-readership ratios of between 2.79 and 4.6 per magazine. The trial court could have reasonably determined a reliable readership figure on the basis of this testimony. It is. true that the court referred to the defense expert’s readership figures in its findings, but only to note that the defendant would be guilty of most of the various charges of fraud even if it credited figures that her own witnesses proposed.

Ms. Habhab also challenges her convictions on the basis that government investigators intimidated two potential witnesses, Carla Jones and Nancy Beisser, thus denying Ms. Habhab due process of law in her trial. We recognize that government conduct designed to intimidate potential defense witnesses is improper. See Harrington v. Nix, 983 F.2d 872, 874-75 (8th Cir.1993) (per curiam). We also understand that intimidation may occur when a potential witness is informed, say, on the eve of trial, that he or she is under criminal investigation. See United States v. Thirion, 813 F.2d 146, 156-57 (8th Cir.1987). Here, however, there is no evidence that the government’s actions were designed to intimidate witnesses or that Ms. Habhab suffered any actual prejudice as a result.

Ms. Habhab’s intimidation claim rests mainly on a letter that Assistant United States Attorney Janet Papenthien sent to Ms. Jones and Ms. Beisser, indicating that a plea agreement might be forthcoming if they were to provide pertinent information to the government. The letter is not, on its own, a threat. Indeed, returning the letter and agreeing to proffer information is completely voluntary; nothing in the letter indicates a penalty for failing to cooperate.

We do not hold that such a letter may never be a source of intimidation. For exam- *416 pie, a letter that also serves as a potential witness’s first notification that he or she is under investigation, especially if delivered immediately before the witness is to testify, would be highly suspect. See Thirion, 813 F.2d at 157. Here, however, the potential witnesses knew or should have known of the investigation pending against them. Both of them assisted Ms. Habhab’s counsel in reviewing discovery materials, which included documents previously naming them as targets of the investigation. Furthermore, they hardly received their letters on the eve of trial; instead, the letters arrived two months before the trial began. The letters are simply evidence of the United States Attorney’s Office offering to negotiate with two people under investigation, two months before they might testify for a third person. One defendant’s trial should not necessarily sever all lines of communication, for some extended period of time such as two months, between the United States Attorney’s Office and other people under investigation.

Once there is a determination that there was no intimidation in this instance, Ms. Habhab’s challenge resolves into a conflict between Ms. Habhab’s right to call witnesses in her defense and the rights of those witnesses to avoid self-incrimination. It is well settled that an accused’s right to compulsory process must yield to a witness’s Fifth Amendment privilege not to give testimony that would tend to incriminate him or her. See Culkin v. Purkett, 45 F.3d 1229, 1233 (8th Cir.), cert. denied, — U.S. -, 116 S.Ct. 127, 133 L.Ed.2d 76 (1995).

Finally, Ms. Habhab requests a new trial on the grounds that the nineteen-month delay between the conclusion of the bench trial and the court’s entry of judgment violated her right to a speedy trial. While we have not previously addressed the speedy trial right in this exact context, we think that our cases dealing with pre-trial, post-indictment delays provide some guidance on the matter. We typically resort to balancing to evaluate pre-trial, post-indictment delays, weighing considerations such as the length of the delay, the reason for the delay, whether the defendant asserted his or her right to a speedy trial, and the actual prejudice to the defendant. See Matthews v. Lockhart, 726 F.2d 394, 396 (8th Cir.1984).

The first of these considerations, the length of the delay, is a threshold issue: if the delay is not “presumptively prejudicial,” id., further inquiry becomes unnecessary. We have found a seventeen-month delay, in the context of a pre-trial, post-indictment delay, to be presumptively prejudicial. While it is true that prejudice of the same character or degree may not necessarily arise from post-trial delay, we believe that a nineteen-month delay between the conclusion of testimony and the finding of facts in a bench trial is so extraordinarily long that it satisfies the threshold requirement of presumed prejudice and leads us to make a more detailed inquiry.

Each of the considerations described in Matthews other than the length of the delay, however, weighs against finding a violation of the speedy trial right here. The reason for the delay, according to the trial court, was to benefit the defendant. The court stated that it waited to sentence Ms. Habhab until Congress took action on pending recommendations to change the sentencing guidelines relevant to money laundering. The recommended sentencing guidelines lowered the base offense level for money laundering, thus shortening Ms. Habhab’s period of incarceration. Shortly after Congress acted on the proposed guideline changes, the trial court decided the case.

Ms. Habhab did not assert her right to a speedy trial until nearly fourteen months after trial. While we understand her argument that a request for a decision might alienate the trial court and precipitate a rash or unfavorable decision, we will not presume that to be the case. Indeed, we have every confidence that it is not. If Ms. Habhab wanted a decision shortly after trial, she could have requested one. We will not allow Ms. Habhab to gamble on the potential advantage of the delay and then seek reversal on the grounds that the delay was excessive.

The actual prejudice suffered by Ms. Ha-bhab was relatively small. Ms. Habhab was not incarcerated during the time in question. While she alleges limitations on her ability to *417 travel and to convey property, she has made no showing that a request to travel was denied or that she wanted to sell property and was unable to do so. The delay did not affect Ms. Habhab’s ability to prepare an effective defense. While a pre-trial delay may allow necessary witnesses to die, become unavailable, or simply forget important facts, a post-trial delay cannot lead to the same complications.

VI.

For the foregoing reasons, we affirm each of the convictions.

Notes

. The Honorable Donald E. O'Brien, United States District Judge for the Northern District of Iowa.