DON R. WILLETT, Circuit Judge:
Three Africa-based cybercriminals-Oladimeji Ayelotan, Femi Mewase, and Rasaq Raheem-masterminded a sprawling international romance scam that stole hearts, and money. Posing as bachelors (and bachelorettes) online, these Nigerian nationals kindled digital romances with scores of lovelorn Americans. The fraudsters sat at overseas computers, prowling the Internet and spinning false promises of love and romance, ultimately duping their unsuspecting victims into sending money to Nigeria and South Africa.
Many fauxmance swindlers escape scot-free, their victims, broke and brokenhearted, too embarrassed to come forward. Not this time. A wary target reported her suspicions, and the scammers didn't fare as well in court as they had online. After a 16-day trial, the jury convicted them, and the district court imposed lengthy prison sentences. This appeal alleges several errors-the district court's imposition of leg restraints during trial; the admission of emails and a nonoriginal passport; the dismissal of a juror during trial; and the sentences handed down.
Each argument is meritless, and we AFFIRM.
I. BACKGROUND
The transnational romance scam worked like this: Ayelotan, Mewase, and Raheem-along with their coconspirators-stole personal information such as names, Social Security numbers, credit card numbers, and bank account numbers. They then impersonated their victims-getting cash advances and transferring funds out of the victims' accounts.
But they needed a safe way to transfer the fruits of their crimes. Thus their digital dalliances. Using dating websites like "seniorpeoplemeet.com," well-honed conversation scripts, and step-by-step guides, the conspirators cultivated online relationships, then sweet-talked their "paramours" into laundering their money. Next, the conspirators would cajole their enamored victims into becoming money mules, conduits for stolen funds, even providing prepaid shipping labels for the swindled cash and goods.
Everything was going according to plan until one prospective money mule grew suspicious. She reported her experience to the police, who ran it by Homeland Security Investigations (the Department of Homeland Security's investigation arm). Agent Todd Williams, posing over email as
*400
the target victim, helped unravel the whole scheme.
The district court held a 16-day trial, during which it put the three defendants in leg restraints. The court also removed and replaced one of the jurors.
The jury convicted Ayelotan and Raheem on several counts of conspiracy to commit mail fraud, wire fraud, and bank fraud; conspiracy to commit identity theft, use of unauthorized access devices, and theft of government property; mail fraud; and conspiracy to commit money laundering.
The jury acquitted Mewase of conspiracy to commit money laundering. But it convicted him of conspiracy to commit mail fraud, wire fraud, and bank fraud; and conspiracy to commit identity theft, use of unauthorized access devices, and theft of government property.
The defendants filed post-trial motions for relief. But the district court sentenced each defendant to the statutory maximum for each conviction, running consecutively.
II. STANDARD OF REVIEW
The arguments raised on appeal involve varying standards of review. We review decisions to shackle criminal defendants, admit evidence, and remove jurors for abuse of discretion.
And it is an abuse of discretion to apply an erroneous view of the law or to clearly err in assessing evidence.
We review alleged Confrontation Clause violations de novo.
As for sentencing, we review fact findings for clear error and application of the Sentencing Guidelines de novo.
And as the Supreme Court requires, we review the substantive reasonableness of sentences for abuse of discretion.
Finally, the Supreme Court directs us to review arguments raised for the first time on appeal for plain error.
In
Puckett
, the Court elaborated that under this standard of review, defendants must first establish an error.
Next, they must show that it's clear or obvious.
Then, they must prove that the error affected their substantial rights.
And if they satisfy these three prongs, we may correct the error-
if
it "seriously affects the fairness, integrity, or public reputation of judicial proceedings."
III. DISCUSSION
Ayelotan, Raheem, and Mewase bring assorted challenges-none availing-to their convictions and sentences:
•
Ayelotan and Raheem
-the shackling of their legs during trial
•
Raheem and Mewase
-the admission of various emails and Mewase's nonoriginal passport
*401
•
Ayelotan and Mewase
-the dismissal of a juror
•
All three defendants
-their sentences
A. Shackling the defendants at trial was not an abuse of discretion.
Ayelotan and Raheem claim that shackling their legs violated their due process rights. As the Supreme Court explained 14 years ago in
Deck
, the Fifth and Fourteenth Amendments' due process clauses require courts to have a "particular reason" for shackling; "only in the presence of a special need."
And the Court forbids visible restraints altogether unless "justified by an essential state interest" specific to that trial.
What reasons are enough? Some 25 years ago, in
Wilkerson
, we held that courts may shackle defendants when there's a danger of harm or escape.
As we said then, "[w]e do not underestimate the need to restrain dangerous defendants to prevent courtroom attacks, or the need to give trial courts latitude in making individualized security determinations."
And a few years before that, in
Ellender
, we stated that district courts "may rely heavily on the U.S. Marshal's advice" in considering restraints.
Here, the court had valid reasons: Ayelotan and Raheem posed a danger. And the facts show that. For example, at Ayelotan's extradition hearing, he and other defendants caused such a ruckus that SWAT had to be called in. The district court stressed this concern. And given the extradition escapades, the U.S. Marshals expressed unease too-recommending that the court restrain the defendants at trial. The court had a valid, particularized reason for shackling the defendants.
Plus, the defendants' restraints weren't visible. Neither Ayelotan nor Raheem even
claim
that the restraints were visible. Neither defendant raises any evidence suggesting that the jury saw the restraints. And Raheem asserts merely that it's impossible to be sure that a juror never saw him in shackles.
In sum, the district court had compelling reasons for shackling the defendants, and there's no evidence the jury even saw the restraints. And so the district court didn't violate the defendants' due process rights.
B. The emails and copy of Mewase's passport were admissible.
At trial, the Government admitted oodles of emails that the defendants sent to their romantic targets. These emails all came from Google and Yahoo! accounts. And they revealed the defendants' fraudulent activities. They also included instructions for money mules to send cash and progress updates on their various schemes.
The defendants filed pretrial objections to admission of the emails. But the district court held that the emails and transmittal records, accompanied by Google and Yahoo! records-custodian certificates, were admissible self-authenticating business records.
On appeal, Raheem and Mewase challenge the admission of these emails and
*402
records under the Federal Rules of Evidence and the Confrontation Clause. Mewase also contends that admission of a duplicate copy of his passport identification page violated the Best Evidence Rule. Both challenges fail.
The Federal Rules of Evidence ban hearsay-out-of-court statements made to prove the truth of what's asserted.
But there are exceptions. Here, each email represented two "statements" for purposes of the Federal Rules of Evidence. First was the transmittal certificate-effectively, the email provider's statement that one user wrote and sent a message to another user at the recorded time. Second, the content of each email is also a statement. We take them in that order.
The Federal Rules of Evidence except business records from hearsay.
To qualify, a records custodian with knowledge must testify unless they're "self-authenticating."
Records are self-authenticating if they include a custodian certification that the records "meet[ ] the requirements of Rule 803(6)(A)-(C)."
All the email records from this conspiracy included certificates. The certificates stated that Google or Yahoo! recorded the transmittal data automatically when users send emails, as part of the regular practice of a regularly conducted business activity. This satisfies Rule 803(6) 's requirements for admission.
As we explained in our 1991
Wilson
decision, certificates from a records custodian that "track the language of Rule 803(6) nearly word for word" render the records self-authenticating.
Thus, the district court didn't abuse its discretion by finding a valid hearsay exception for the certificates.
The second possible hearsay is the emails' substantive content: the messages between the defendants and their coconspirators. But these were admissible too. The Government didn't offer these statements to prove the content of them.
Same with the e-mule messages.
Rather, these statements were "the operative words of [the] criminal action"-what we called "paradigmatic nonhearsay" in our 1981 case
Jones
.
The remaining content in the emails-updates between the coconspirators about their criminal
*403
scheme-was admissible as opposing party and coconspirator statements under Rule 801(d).
2. Confrontation Clause
Our Confrontation Clause analysis resembles our Federal Rules of Evidence analysis. And likewise, the district court didn't misstep in admitting the emails and records.
In 2004, the Supreme Court held in
Crawford
that the Confrontation Clause prohibits admitting out-of-court statements as evidence against defendants in a criminal case unless they can cross-examine the declarant.
But that prohibition applies only if the statements are "testimonial."
And two years later in
Davis
, the Court explained that statements are "testimonial" if their "primary purpose ... is to establish or prove past events potentially relevant to later criminal prosecution."
Or as the Court explained in
Melendez-Diaz
three years after that, business records must be "for the administration of the [business's] affairs and not for the purpose of establishing or proving some fact at trial."
Under this framework, Google and Yahoo!'s transmittal-data records aren't testimonial. Consider our
Towns
decision in 2013.
There, we considered a pharmacy's prescription-purchase logs.
We held that those weren't testimonial.
Rather, the pharmacy recorded those logs "
ex ante
to comply with state regulatory measures, not in response to active prosecution."
It's the same with Google and Yahoo! They didn't create the records to prove a particular fact at a particular trial-let alone this trial. The records are admissible as far as the Confrontation Clause goes.
So are the coconspirator statements. Take our 2011 decision in
Jackson
.
In that case, we explained that-as a general matter-"coconspirator statements made during the course and in furtherance of a conspiracy" aren't testimonial.
So the district court didn't violate the defendants' Confrontation Clause rights by admitting this evidence.
*404
3. Best Evidence Rule
Mewase also contends that the passport copy was inadmissible under the Best Evidence Rule. True, the Federal Rules of Evidence require using "[a]n original writing, recording, or photograph" as evidence of that item's contents at trial.
But the Rules clarify that "[a] duplicate is admissible to the same extent as the original unless a genuine question is raised about the original's authenticity or the circumstances make it unfair to admit a duplicate."
Mewase's original passport went missing during his extradition. And Mewase acknowledges that it wasn't the Government's fault. Even so, he still claims that relying on the duplicate isn't fair. And yet he doesn't explain why. So the district court didn't abuse its discretion in admitting the duplicate.
C. The district court properly removed and replaced Juror 20.
Ayelotan and Mewase contend that the district court's decision to remove Juror 20 violated their due process rights. We review the removal and replacement of a juror for abuse of discretion.
And as we explained in
Huntress
, a court may not dismiss a juror "without factual support or for a legally irrelevant reason."
The Federal Rules of Criminal Procedure allow a judge to remove jurors who cannot perform their duties.
We explained what that means long ago. Forty-plus years ago in
Smith
, we held that "a juror who cannot remain awake during much of the trial is unable to perform his duty."
And in 2002, we explained in
Edwards
that other good reasons for removal include a "lack of candor" and an "inability or unwillingness to follow instructions."
Here, the district court identified legally relevant reasons for removing Juror 20. The juror slept through witness testimony; misrepresented this fact to the district court when asked; didn't understand, or else didn't follow, the jury instructions; and didn't deliberate.
Ample evidence supported these legitimate reasons. Near the end of the first day of jury deliberations, the foreperson delivered two notes to the court:
We have one person that is undecided, and we [cannot] get an answer from him.
One, we have a juror that has admitted that he slept through some of the testimony; two, he doesn't believe you can't go to Western Union and pick up transaction, his opinion/own notation; three, he refuses to follow the judge's instruction.
After reviewing the notes, the court called in the foreperson, Juror 20, and every other juror one at a time.
When the district court asked Juror 20 about sleeping through the trial, he admitted that he may have "nodded off." But he said that if he did, "it was very minimal." Yet his fellow jurors testified otherwise. Some said that Juror 20 admitted that he "was asleep some of the time, so [he] may have missed things." And others said he
*405
confessed that "he may not have heard everything he needed to hear" since he fell asleep.
Many jurors even said they saw him sleeping more than just a "minimal" amount, including his neighboring juror:
I sat next to [Juror 20], I noticed that there was quite a few times, a lot of times, that I would look over and notice that he was asleep. ... It was more than once a day .... I do believe that it was enough to miss key information.
And several other jurors believed that Juror 20 simply couldn't follow the law.
Juror 19, for example, testified that the jury pointed out specific instructions to Juror 20, which "he refused to follow"-e.g., "considering only the evidence presented" and following "the law that's in the instructions." Another juror questioned Juror 20's comprehension: "He spent three hours reading [the Court's] instructions yesterday, and he got to Page 5. We've tried to talk to him, and he just refuses."
Based on this evidence, the district court removed Juror 20. But Ayelotan and Mewase argue that the court really removed Juror 20 because he was a "hold-out"; because he "clearly questioned Mr. Ayelotan's guilt." True, Juror 20 resisted convicting. Yet as we noted in
Edwards
, even "hold-out jurors are not immune from dismissal based upon just cause."
The district court had legitimate reasons to dismiss Juror 20. And there was strong factual support for those legitimate reasons. Removing Juror 20 was not an abuse of discretion.
D. Sentencing Challenge
The district court sentenced Ayelotan, Raheem, and Mewase to practically interminable prison terms. The court calculated the total intended loss of their scheme: over $25 million. At a joint sentencing hearing, the court heard testimony from a postal inspector and two case agents from Homeland Security Investigations. They testified on the loss amount and the defendants' roles.
For Ayelotan, the court imposed a 22-level increase for the intended loss. And it imposed a 4-level enhancement for his leadership role. This came to a total Guidelines offense level of 43. Given Ayelotan's criminal history, his Guidelines sentence was life imprisonment. The court reduced that to the within-Guidelines statutory maximum of 95 years.
And it sentenced Ayelotan to that.
For Raheem, the court also assigned a total offense level of 43 for the intended loss and his leadership role. His initial sentencing range was life imprisonment too. But the court again decreased it. His within-Guidelines statutory maximum was 115 years.
And the court sentenced him to that.
Finally, for Mewase, the court also applied the same 22-level increase for the intended loss. The resulting Guidelines range was 262 to 327 months. The court reduced that to 262 to 300, the statutory maximum. The court then sentenced him to 25 years.
All three defendants now challenge their sentences on various grounds.
1. Leadership Enhancement
On appeal, Ayelotan and Raheem
*406
object to their leadership enhancements.
And Ayelotan says that the court should've given him a minor-role reduction.
As we explained in
Hebert
, the preponderance of the evidence must show that the defendant was "the organizer or leader of a criminal activity that involved five or more participants or was otherwise extensive."
We will reverse the enhancement only if the district court clearly erred.
To begin with, the district court found-and the trial evidence supported-that this was a major criminal undertaking. It involved far more than five participants. As for leadership, the commentary to § 3B1.1 lists factors for assessing a defendant's role:
• the exercise of decision-making authority;
• the nature of participation in the commission of the offense;
• the recruitment of accomplices;
• the claimed right to a larger share of the fruits of the crime;
• the degree of participation in planning or organizing the offense;
• the nature and scope of the illegal activity;
• and the degree of control and authority exercised over others.
The commentary also clarifies that more than one person can qualify as a leader of the same criminal group.
Ayelotan argues that the second superseding indictment proves that his role was "peripheral in substance" and "peripheral in duration." He claims that the only accusation against him was that he emailed shipping-label information.
But the district court specifically relied on several § 3B1.1 factors in determining leadership roles. At sentencing, the court heard testimony from case agent Williams. Agent Williams testified that Ayelotan was among the "top people within the group of indicted defendants." Williams asserted that Ayelotan was a "central hub to a number of other co-conspirators on a daily basis." Agent Williams also identified Raheem as one of four coconspirators "at the top of th[e] hierarchy."
Williams based this on their time spent on the conspiracy, their day-to-day involvement, and the number of coconspirators who reported to them. The district court relied on this testimony. And it determined that Ayelotan "was involved in almost every scam involved in the offense"; that he "participated in planning or organizing it"; and that he directly controlled other conspirators. Similarly, the court credited Williams's testimony on Raheem's leadership in the organization.
The district court did not clearly err considering the evidence before it-not with Ayelotan and not with Raheem.
2. Eighth Amendment
Now for the first time on appeal, Ayelotan and Raheem raise an Eighth Amendment argument. They assert that effective life sentences for nonviolent crimes are cruel and unusual punishment. We review arguments raised for the first time on appeal for plain error.
Although Ayelotan
*407
acknowledges the plain-error standard, he never briefed anything beyond the first step of the plain-error test: whether an error occurred.
Yet even if Ayelotan had sufficiently briefed the other steps of the plain-error analysis, he still wouldn't be entitled to relief. Yes, in extraordinary cases, term-of-years sentences can be cruel and unusual. In 1992, we held in
McGruder
that sentences violate the Eighth Amendment when they're "grossly disproportionate" to the convicted conduct.
But as we've explained before, this review is narrow. In our 2016 case
Mills
, we underscored that a reviewing court may not "substitute its judgment for that of the legislature nor the sentencing court."
And four years before that, in
Looney
, we acknowledged that under Supreme Court precedent, "successful Eighth Amendment challenges to prison-term lengths will be rare."
Ayelotan maintains that an effective life sentence for a nonviolent, financial crime is-by definition-grossly disproportionate. Yet he cites no cases finding that lengthy sentences for financial crimes violate the Eighth Amendment. He cites only to the Supreme Court's 2010 decision in
Graham
and its 1992 decision in
Robinson
.
But in
Graham
, the Supreme Court limited its decision to life sentences for minors.
And in
Robinson
, the Court held that a law criminalizing narcotics addiction was unconstitutional.
The Court struck down the law because its purpose was to stigmatize addicts; not because the sentence was too long.
Neither decision applies here.
Yet our decision in
Mills
does apply. There, we held that "the Guidelines are a convincing objective indicator of proportionality."
And in
Looney
, we upheld effective life sentences for the nonviolent offenses of possession with intent to distribute methamphetamine and possession of firearms.
Ayelotan's and Raheem's within-Guidelines sentences didn't violate the Eighth Amendment. The district court did not err, plainly or otherwise.
3. Intended Loss Calculation
Raheem and Mewase challenge the intended loss calculation of over $25 million. In 2012, we laid out the rule for calculating intended losses in
Hebron
: "The applicable loss is generally the greater of actual loss-which includes only reasonably foreseeable harm resulting from the fraud-and intended loss-which includes the harm intended to result from the offense."
And district courts need
*408
only make "a reasonable estimate of the loss" based on the evidence.
We review these calculations for clear error.
In
Hull
, we explained that we review the foreseeability of the loss for clear error too.
Then in
Sanders
, we established that we review the loss calculation itself for clear error.
We won't overturn these factual findings "unless they are implausible in light of the record as a whole."
Raheem retorts that the record doesn't show that they "intended to inflict a loss" of $25 million or more. First, that misstates the standard. The Sentencing Guidelines don't require the defendant to have intended the
specific
loss amount.
Instead, the district court simply has to conclude that the defendant knew or reasonably should have known that the scheme would cause the harm.
And the district court relied on testimony that showed this: "Mr. Raheem and Mr. Ayelotan were directly involved in most, if not all of [the activity used to calculate the loss amount], and to the extent they were not directly involved, it was certainly reasonably foreseeable to them as part of the jointly undertaken criminal activity which was part and parcel to this conspiracy."
Second, the record does show that the intended loss was over $25 million. The district court considered the PSR's calculations. And it heard testimony from the postal inspector and two case agents. Based on that, the court calculated the loss at $52 million.
Finally, Raheem disputes the district court's determination that there were at least 37,817 credit cards involved in the conspiracy. Rather, Raheem urges that "there is no information all of the alleged credit cards were real." Yet the district court heard specific evidence suggesting that the credit card numbers corresponded to real accounts. For example, Williams testified that the first six numbers on the cards-generally identifying the financial institutions-matched with real institutions; and that the stolen information included three-digit CVV codes on the back of each card.
In any event, Raheem and his coconspirators' intended to get their hands on valid credit card numbers. Otherwise, their scheme could bear no fruit. Phony credit cards won't even buy you Monopoly Money-let alone cash advances. Even if some card numbers were fake, that wouldn't matter. Again, their
intent
dictates the
intended
loss.
4. Substantive Reasonableness of Sentence
Finally, Raheem and Mewase contend that their sentences are substantively unreasonable. But in 2009, we reiterated in
Cooks
that we presume a within-Guidelines-range sentence is reasonable.
And we tend to defer to the sentencing court. As we noted in our 2011 decision in
Scott
, "the sentencing court is in a better position
*409
to find facts and judge their import under the § 3553(a) factors."
To rebut the presumption, a defendant must show one of three things: (1) the court failed to consider a factor that it should've given significant weight; (2) the court gave significant weight to an irrelevant or improper factor; or (3) the court clearly erred in balancing sentencing factors.
Raheem argues that his sentence exceeds the aim of the § 3553(a) sentencing factors. But his sentence was within-Guidelines. So it's presumed reasonable. And here, Raheem hasn't rebutted that presumption.
The district court detailed its justifications and discussed the § 3553(a) factors. The court noted that Raheem was "a leader or organizer" in "a broad and wide-reaching and extensive conspiracy involving criminal conduct that went on for a number of years"; one that "involved the manipulation and taking advantage of, whether financially or otherwise, a large number of victims."
Mewase argues that his sentence is unreasonable because of "unwarranted sentencing disparities" between him and some coconspirators who pleaded guilty. But this doesn't rebut the presumption of reasonableness either. Mewase offers no proof that he was so similarly situated to coconspirators who received lesser sentences that the disparities are "unwarranted." Nor does the PSR suggest that. Rather, many other coconspirators played much smaller roles, accepted responsibility, or cooperated with the Government.
In sum, we have consistently declined to merely reweigh the sentencing factors. Yes, Raheem's within-Guidelines sentence is severe. But the district court weighed the appropriate factors. And its sentence isn't substantively unreasonable. The district court also grounded Mewase's sentence in an in-depth consideration of the § 3553 factors. So neither sentence is substantively unreasonable.
IV. CONCLUSION
The district court committed no error, and we AFFIRM in all respects.