FOR PUBLICATION
 UNITED STATES COURT OF APPEALS
      FOR THE NINTH CIRCUIT

UNITED STATES OF AMERICA,             
                Plaintiff-Appellee,         No. 05-30076
               v.                            D.C. No.
PATRICIA ANN LARSON,                      CR-04-00110-SEH
             Defendant-Appellant.
                                      

UNITED STATES OF AMERICA,                  No. 05-30077
                Plaintiff-Appellee,
               v.                            D.C. No.
                                          CR-04-00110-SEH
LEON NELS LAVERDURE,
                                             OPINION
             Defendant-Appellant.
                                      
       Appeals from the United States District Court
                for the District of Montana
        Sam E. Haddon, District Judge, Presiding

             Argued and Submitted En Banc
        March 20, 2007—San Francisco, California

                   Filed August 1, 2007

     Before: Mary M. Schroeder, Chief Circuit Judge,
 Harry Pregerson, Stephen Reinhardt, Pamela Ann Rymer,
         Michael Daly Hawkins, Susan P. Graber,
        Kim McLane Wardlaw, Raymond C. Fisher,
  Ronald M. Gould, Richard A. Paez, Richard C. Tallman,
 Richard R. Clifton, Carlos T. Bea, Milan D. Smith, Jr., and
              Sandra S. Ikuta, Circuit Judges.

                            9139
9140      UNITED STATES v. LARSON
         Opinion by Judge Paez;
       Concurrence by Judge Graber;
        Dissent by Judge Hawkins
                  UNITED STATES v. LARSON              9143


                        COUNSEL

David F. Ness, Assistant Federal Defender, Federal Defenders
of Montana, Great Falls, Montana, and James B. Obie,
Helena, Montana, for the defendants-appellants.
9144                  UNITED STATES v. LARSON
Joseph E. Thaggard, Assistant United States Attorney, United
States Attorney’s Office, Great Falls, Montana, for the
plaintiff-appellee.


                             OPINION

PAEZ, Circuit Judge:

   These appeals present the question whether the Sixth
Amendment Confrontation Clause rights of Defendants Patri-
cia Ann Larson and Leon Nels Laverdure were violated when
they were barred from cross-examining two witnesses about
the mandatory minimum prison sentences that they would
have faced but for their cooperation with the Government. A
three-judge panel of our court held that there was no constitu-
tional violation and affirmed Defendants’ convictions. United
States v. Larson, 

(9th Cir. 2006). We subse-
quently granted rehearing en banc. Before addressing the mer-
its of Defendants’ constitutional arguments, we clarify the
standard of review that we apply to Confrontation Clause
challenges. Under the circumstances here, we review for
abuse of discretion, and we hold that Defendants’ Confronta-
tion Clause rights were violated. The error was harmless,
however, and we therefore affirm their convictions.1

                                  I.

                            Background

   In 2003, the Great Falls Police Department began investi-
gating a number of drug dealers in the Great Falls, Montana
area. In April 2003, the Department paid informant Connie
Riggs to make a controlled purchase of 1.8 grams of metham-
phetamine from Larson. The police unsuccessfully attempted
  1
   Defendants do not challenge their sentences.
                       UNITED STATES v. LARSON                         9145
a second controlled buy from Larson through Riggs. In Octo-
ber 2003, informant Jason Gilstrap made two controlled pur-
chases of methamphetamine from Laverdure, one in the
amount of 1.46 grams and the other in the amount of 1.79
grams. In December 2003, a third informant purchased 3.2
grams of methamphetamine from Joy Lynn Poitra and her
cousin, Rick Lee Lamere. Two weeks later, the same infor-
mant purchased another twenty-one grams of methamphet-
amine from Poitra.

   On July 23, 2004, a federal grand jury indicted Larson,
Laverdure, Poitra, and Lamere in a single indictment, charg-
ing each with one count of conspiracy to possess metham-
phetamine with intent to distribute.2 The Government
informed Lamere that because he had at least two prior felony
drug convictions, he faced a statutory mandatory minimum
sentence of life imprisonment without the possibility of
release. See 21 U.S.C. § 841(b)(1)(A).

   Both Poitra and Lamere entered into plea agreements
whereby they agreed to cooperate and to testify against Lar-
son and Laverdure. In return, the Government represented that
if Poitra and Lamere testified to its satisfaction, it would file
a motion for a reduced sentence, recommending to the court
the extent of departure warranted by the cooperation. See
U.S.S.G. § 5K1.1 (2004). Lamere pleaded guilty to the con-
spiracy charge and admitted to having distributed five kilo-
grams of a substance containing methamphetamine. Only by
  2
    The indictment alleged that between January 1, 1999, and February 29,
2004, the four (and others) conspired “to possess with the intent to distrib-
ute controlled substances, including but not limited to 500 grams or more
of a substance containing a detectable amount of methamphetamine,” in
violation of 21 U.S.C. §§ 841(a)(1) and 846. The indictment attributed
five to fifteen kilograms of a substance containing methamphetamine and
two to 3.5 kilograms of cocaine to Larson, five to fifteen kilograms of a
substance containing methamphetamine to Laverdure, and five to fifteen
kilograms of a substance containing methamphetamine to Lamere. (No
specific amount of methamphetamine was attributed to Poitra.)
9146                 UNITED STATES v. LARSON
the Government’s motion for substantial assistance could
Lamere receive a sentence below the statutory mandatory
minimum of life imprisonment. See 18 U.S.C. § 3553(3);
U.S.S.G. § 5K1.1 cmt. n.1 (2004). Poitra pleaded guilty under
a superseding indictment that charged her with conspiring to
possess with intent to distribute at least fifty grams of a sub-
stance containing a detectable amount of methamphetamine.
As a result of her plea, she faced a statutory minimum sen-
tence of five years imprisonment in the absence of a § 5K1.1
motion by the Government, with a potential maximum sen-
tence of forty years in prison. 21 U.S.C. § 841(b)(1)(B).

   In October 2004, Larson and Laverdure proceeded to trial
before a jury. Poitra and Lamere testified as witnesses for the
Government.3 Poitra testified that in 2003, Laverdure, her
third cousin, offered to help her purchase methamphetamine.
Poitra agreed and later accompanied Laverdure when he went
to purchase the methamphetamine. She waited in an alley
while Laverdure walked toward Larson’s house. Laverdure
returned with .25 ounces of methamphetamine, which Poitra
later sold. Poitra continued to purchase methamphetamine
regularly from Laverdure, totaling about 3.5 ounces. She testi-
fied that she overheard Laverdure calling “Patty” Larson to
obtain methamphetamine on one or two occasions. Poitra fur-
ther testified that she obtained from Lamere both the 3.2 and
twenty-one grams of methamphetamine that she sold to the
confidential informant, and that Lamere obtained the latter
quantity for her from Larson.

   Poitra acknowledged before the jury that she had also been
charged with conspiring to possess methamphetamine with
intent to distribute and that she was cooperating with the
expectation that the Government would move to reduce her
sentence. She admitted that she was a drug addict and a
dealer. On cross-examination, Poitra also admitted that she
had engaged the police in a high-speed car chase and that,
  3
   Neither Poitra nor Lamere had been sentenced at the time.
                   UNITED STATES v. LARSON                 9147
when caught, she had been found with methamphetamine,
marijuana, a digital scale to weigh drugs, and ammunition for
a 9mm handgun. She acknowledged that she had distributed
280 grams, roughly ten ounces, of methamphetamine. Lar-
son’s counsel questioned Poitra about inconsistencies in her
testimony, including her previous statement to police that a
person other than Larson or Laverdure had been Lamere’s
only source of methamphetamine.

   Larson’s counsel asked Poitra whether she faced a manda-
tory minimum sentence of five-years imprisonment in the
absence of a motion by the Government to reduce her sen-
tence; she answered that she d

judge interjected
that this line of questioning was improper and informed the
jury that “all matters related to sentencing are the decision of
the court and the court only.” The court, however, permitted
counsel to ask Poitra whether she understood that only the
Assistant United States Attorney could move to reduce her
sentence. Poitra acknowledged that she did, and further testi-
fied that she had a two-year-old child and did not want a
prison sentence.

   Lamere testified that he obtained about ten ounces of
methamphetamine from Laverdure. On four occasions,
Lamere was told by Laverdure that Larson was his metham-
phetamine source, identifying her as the source of 1.5 ounces
total. Lamere also testified that he purchased six ounces of
methamphetamine from a man named Fasto Komeotis.
Komeotis told Lamere that the drugs came from Larson.
Lamere testified that he would drop Komeotis off about a
block away from Larson’s house and that Komeotis would
return with the drugs.

   Lamere testified that he pleaded guilty to conspiring to dis-
tribute more than 500 grams of methamphetamine and that he
was cooperating with the Government in the hope that he
would receive a reduced sentence. He admitted that he was
both a drug addict and a drug dealer. He testified that he had
9148               UNITED STATES v. LARSON
sold at least ten pounds of methamphetamine and that he had
been averaging $3,600 a day in drug sales. He also admitted
that he had been convicted of seven felonies, including drug-
related felonies, and that he had served 2-1/2 years in prison.
On cross-examination, Lamere acknowledged that on direct
examination he testified that he made about fifty purchases
from Laverdure, but that at his change of plea hearing he had
testified that he only dealt with Laverdure three or four times.

   Other witnesses also testified on behalf of the Government.
Informant Riggs testified regarding her controlled purchase of
approximately 1/16 ounce of methamphetamine from Larson.
Riggs admitted to having had a drug and alcohol problem for
twenty-eight years, and to improperly using the money the
Government gave her for making the controlled purchase to
buy drugs. Riggs acknowledged that drug users often help
each other buy drugs and that, when Larson obtained the
methamphetamine for her, Larson had originally asked for
some of the drug in exchange for her assistance. Riggs testi-
fied that she had attempted a second, unsuccessful controlled
buy from Larson.

   Informant Gilstrap testified that he made two controlled
purchases from Laverdure. He testified that he had bought
some methamphetamine from Laverdure “a few times”
before, in “very little” amounts of approximately 1/4 gram.
Gilstrap admitted that he was a drug addict who had been
arrested for drug possession and had gone to prison three
times for drug-related felonies. Gilstrap acknowledged that, as
a benefit of his involvement in the controlled purchases, he
was permitted to live with his family while still an inmate.

   After the Government rested its case, Larson and Laverdure
moved for the district court to reconsider its ruling preventing
them from inquiring into the statutory mandatory minimum
sentences that Poitra and Lamere faced. The court denied the
motion. The jury found Larson and Laverdure guilty. The
court later sentenced Larson to 97-months imprisonment, fol-
                      UNITED STATES v. LARSON                      9149
lowed by four years of supervised release. The court sen-
tenced Laverdure to 188-months imprisonment, followed by
four years of supervised release.

   On appeal, Larson and Laverdure argue, among other
things, that the district court violated their Confrontation
Clause rights when it refused to allow counsel to cross-
examine Poitra and Lamere about the mandatory minimum
sentences that they faced in the absence of cooperation with
the Government. A three-judge panel of our court determined
that their Confrontation Clause rights were not violated
because Poitra’s five-year minimum sentence was of slight
probative value, examination regarding Lamere’s potential
life sentence created a significant risk of undue prejudice, and
“the jury otherwise received sufficient information from
which to evaluate the cooperating witness’s biases and moti-
vations.” 

. The three-judge panel
also rejected Defendants’ challenge to the district court’s evi-
dentiary ruling allowing Lamere to testify that Laverdure and
Komeotis told him that Larson was the source of their
methamphetamine, and to the court’s denial of their request
to sit at the counsel table, rather than behind counsel. See 

   In this en banc proceeding, we focus our inquiry on Defen-
dants’ Confrontation Clause challenge. We hold that the dis-
trict court did not abuse its discretion by limiting Defendants’
cross-examination of Poitra, but that Defendants’ Confronta-
tion Clause rights were violated when the district court barred
Defendants from questioning Lamere about the mandatory
life sentence that he faced in the absence of a motion by the
Government to reduce his sentence. We conclude, however,
that the error was harmless, and we affirm Defendants’ convic-
tions.4
  4
   Because we agree with the three-judge panel’s disposition of the
remaining issues that Defendants raise on appeal, we adopt those portions
of the panel opinion with the exception of the final three paragraphs of
9150                   UNITED STATES v. LARSON
                                    II.

                        Standard of Review

  [1] We begin by acknowledging an intra-circuit conflict
regarding the standard of review for Confrontation Clause
challenges to a trial court’s limitations on cross-examination.5
See United States v. Rodriguez-Rodriguez, 

, 856

Section III.B. See 

-17. These final paragraphs
addressed an issue that was unresolved at the time: whether the test articu-
lated in Ohio v. Roberts, 

(1980), to determine the admissibil-
ity of out-of-court nontestimonial statements survived Crawford v.
Washington, 

(2004). 

. The Supreme
Court has since clarified, however, that Crawford “eliminat[es] Confronta-
tion clause protection against the admission of unreliable out-of-court non-
testimonial statements” and that “the Confrontation Clause has no
application to such statements and therefore permits their admission even
if they lack indicia of reliability.” Whorton v. Bockting, 

,
1183 (2007). Adopting the portions of the three-judge panel opinion that
concluded that the out-of-court statements by Komeotis and Laverdure
were made in furtherance of the conspiracy and were 

, we hold that under Crawford, Defendants’ Confrontation
Clause rights were not violated by the admission of these statements.
   5
     There is also disagreement among the circuits mirroring our intra-
circuit conflict. Compare United States v. Vitale, 

, 195 (2d
Cir. 2006) (reviewing Confrontation Clause claims de novo) with United
States v. Orisnord, 

, 1178 (11th Cir. 2007) (publication
pending) (reviewing for abuse of discretion a Confrontation Clause claim
based on a limitation on cross-examination); United States v. Smith, 

, 220 (4th Cir. 2006) (reviewing restrictions on cross-examination
claims for abuse of discretion); United States v. Tykarsky, 

,
475 (3d Cir. 2006) (same) and United States v. Kenyon, 

,
1063 (8th Cir. 2007) (“We review evidentiary rulings regarding the scope
of a cross examination for abuse of discretion, but where the Confronta-
tion Clause is implicated, we consider the matter de novo.”) (citations
omitted); United States v. Townley, 

, 1271 (10th Cir. 2007)
(“Although a district court’s evidentiary rulings are reviewed for abuse of
discretion, whether admission of such evidence violates the Confrontation
Clause is reviewed de novo.”), petition for cert. filed, ___ U.S.L.W. ___
(U.S. Mar. 12, 2007) (No. 06-10032).
                    UNITED STATES v. LARSON                 9151
(9th Cir.) (recognizing the conflict), cert. denied, 

(2005). One line of Ninth Circuit cases has applied de
novo review. See, e.g., United States v. Holler, 

, 1066 (9th Cir.) (“We review de novo whether a limita-
tion on cross examination violated the defendant’s right of
confrontation.”), cert. denied, 

(2005); United
States v. Wilmore, 

, 871 (9th Cir. 2004)
(“Whether a limitation on cross-examination is so restrictive
that it constitutes a violation of a defendant’s Sixth Amend-
ment right is reviewed de novo.”); United States v. Adamson,

, 612 (9th Cir. 2002) (“We review de novo
whether limitations on cross-examination are so severe as to
violate the Confrontation Clause.”); United States v. Ortega,

, 682 (9th Cir. 2000) (“Whether limitations on
cross-examination are so severe as to violate the Confronta-
tion Clause is a question of law reviewed de novo.”).

   A second line of cases has applied an abuse of discretion
standard. See, e.g., United States v. Lo, 

, 482 (9th
Cir. 2000) (“[W]e review a trial court’s decision to limit the
scope of cross-examination for abuse of discretion, and will
find a Confrontation Clause violation only if the trial court’s
ruling limits relevant testimony[,] . . . prejudices the defen-
dant . . . and denies the jury sufficient information to appraise
the biases and motivations of the witness.”) (internal quota-
tion marks omitted) (alterations in original); United States v.
Shabani, 

, 403 (9th Cir. 1995) (reviewing limits
on cross-examination for abuse of discretion); Wood v.
Alaska, 

, 1550 (9th Cir. 1992) (“Because trial
judges have broad discretion both to determine relevance and
to determine whether prejudicial effect or other concerns out-
weigh the probative value of the evidence, we will find a
Sixth Amendment violation only if we conclude that the trial
court abused its discretion.”).

   A third line of cases combines these two approaches. See,
e.g., United States v. Shryock, 

, 979-80 (9th Cir.
2003) (“Whether limitations on cross-examination are so
9152               UNITED STATES v. LARSON
severe as to violate the Confrontation Clause is a question of
law we review de novo . . . . We review for an abuse of dis-
cretion the district court’s limitation of cross-examination.”);
United States v. Bensimon, 

, 1128 (9th Cir.
1999) (“This court reviews de novo whether the limitation on
cross-examination violated [the defendant’s] right of confron-
tation. The district court, however, has considerable discretion
in restricting cross-examination, and this court will find error
only when that discretion has been abused.”) (citations omit-
ted); United States v. James, 

, 713 (9th Cir.
1998) (“The district court may impose reasonable limits on
cross-examination, which we will review for an abuse of dis-
cretion. Whether limitations on cross-examination are so
severe as to amount to a violation of the confrontation clause
is a question of law we review de novo.”) (citation and inter-
nal quotation marks omitted); United States v. Vargas, 

, 704 (9th Cir. 1991) (“The trial judge has discretion
to impose reasonable limits on cross-examination, and this
court finds error only when that discretion has been abused.
. . . Whether limitations on cross-examination are so severe as
to amount to a violation of the confrontation clause is a ques-
tion of law reviewed de novo.”).

   [2] We resolve this conflict by concluding that the third
approach is most appropriate. We hold that the following
approach should be used to review whether a trial court
improperly restricted a defendant from cross-examining a
prosecution witness: If the defendant raises a Confrontation
Clause challenge based on the exclusion of an area of inquiry,
we review de novo. In reviewing a limitation on the scope of
questioning within a given area, we recognize that “trial
judges retain wide latitude insofar as the Confrontation Clause
is concerned to impose reasonable limits on such cross-
examination based on concerns about, among other things,
harassment, prejudice, confusion of the issues, the witness’
safety, or interrogation that is repetitive or only marginally
relevant.” Delaware v. Van Arsdall, 

, 679
(1986). A challenge to a trial court’s restrictions on the man-
                        UNITED STATES v. LARSON                         9153
ner or scope of cross-examination on nonconstitutional
grounds is thus reviewed for abuse of discretion.

   This approach is in harmony with the Supreme Court’s
decision in Van Arsdall.6 In Van Arsdall, the Court recognized
that “trial judges retain wide latitude insofar as the Confronta-
tion Clause is concerned to impose reasonable limits” on
cross-examination. 

held, however, that where
“the trial court prohibited all inquiry into the possibility” of
a witness’ bias, the defendant’s Confrontational Clause rights
were violated. 

also recognizes that whether
there has been a Confrontation Clause violation is ultimately
a question of law that must be reviewed de novo. See 

; cf. Ornelas v. United States, 

,
  6
    It also brings us in line with a number of our sister circuits. See, e.g.,
United States v. Holt, ___ F.3d ___, 

, *3 (7th Cir. 2007)
(publication pending) (“We review the district court’s decision to limit the
scope of cross-examination for an abuse of discretion. If the Sixth Amend-
ment right to confrontation is directly implicated, however, our review is
de novo.”) (citations omitted); United States v. Kenyon, 

,1063 (8th Cir. 2007) (“We review evidentiary rulings regarding the
scope of a cross examination for abuse of discretion, but where the Con-
frontation Clause is implicated, we consider the matter de novo.”) (cita-
tions omitted); United States v. Martinez-Vives, 

, 53 (1st Cir.
2007) (“On a challenge to a district court’s limitation of cross-
examination, we first perform a de novo review to determine whether a
defendant was afforded a reasonable opportunity to impeach adverse wit-
nesses consistent with the Confrontation Clause. Provided that threshold
is reached, we then review the particular limitations only for abuse of dis-
cretion.”) (citations and internal quotation marks omitted); United States
v. Hitt, 

, 155-56 (5th Cir. 2006) (“A district court’s limitation
of cross-examination of a witness is reviewed for abuse of discretion.
Abuse-of-discretion review is only invoked if the limitation did not curtail
the defendant’s Sixth Amendment right to confront witnesses. Whether a
defendant’s Sixth Amendment rights were violated is reviewed de novo.”)
(citations omitted), cert. denied, 

(2007); United States v.
Townley, 

, 1271 (10th Cir. 2007) (“Although a district
court’s evidentiary rulings are reviewed for abuse of discretion, whether
admission of such evidence violates the Confrontation Clause is reviewed
de novo.”).
9154                UNITED STATES v. LARSON
697-99 (1996) (holding that courts of appeal are to review de
novo district court determinations regarding reasonable suspi-
cion and probable cause, reasoning that de novo review pre-
vents divergent interpretations of similar facts, enables
appellate courts to maintain control of legal rules, and unifies
precedent). At the same time, this standard allows the trial
court discretion to manage the presentation of evidence and
restrict examination based on relevant considerations. See Van

.

   [3] Applying this approach here, Defendants’ Confronta-
tion Clause claims challenge the district court’s limitation on
the scope of cross-examination within an area of inquiry: the
biases and motivations to lie of the Government’s cooperating
witnesses, Poitra and Lamere. We therefore review Defen-
dants’ claims for abuse of discretion.

                              III.

                          Discussion

A.     Constitutional Error

   [4] The Confrontation Clause of the Sixth Amendment,
which “guarantees the right of an accused in a criminal prose-
cution ‘to be confronted with the witnesses against him,’ ”
Van 

, includes “the right of effective
cross-examination,” Davis v. Alaska, 

, 318
(1974). Effective cross-examination is critical to a fair trial
because “[c]ross-examination is the principal means by which
the believability of a witness and the truth of his testimony are
tested.” 

. We, like the Supreme Court,
have “emphasized the policy favoring expansive witness
cross-examination in criminal trials.” 

; see
also Van 

-79; 

.

  [5] The Supreme Court has recognized that “the exposure
of a witness’ motivation in testifying is a proper and impor-
                   UNITED STATES v. LARSON                  9155
tant function of the constitutionally protected right of cross-
examination.” 

-17. Thus, “jurors [are]
entitled to have the benefit of the defense theory before them
so that they [can] make an informed judgment as to the
weight to place on [the Government witness’] testimony.” 

As we explained in United States v. Schoneberg:

    The constitutional right to cross-examine is
    “[s]ubject always to the broad discretion of a trial
    judge to preclude repetitive and unduly harassing
    interrogation,” but that limitation cannot preclude a
    defendant from asking, not only “whether [the wit-
    ness] was biased” but also “to make a record from
    which to argue why [the witness] might have been
    biased.”

, 1042 (9th Cir. 2005) (quoting 

) (footnotes omitted) (alterations in original).

   [6] The issue here is whether the district court erred when
it prevented defense counsel from exploring the mandatory
minimum sentences that Poitra and Lamere faced in the
absence of a U.S.S.G. § 5K1.1 motion by the Government.
We have identified three factors courts should consider in
determining whether a defendant’s Confrontation Clause right
to cross-examination was violated:

    (1) [whether] the excluded evidence was relevant;
    (2) [whether] there were other legitimate interests
    outweighing the defendant’s interest in presenting
    the evidence; and (3) [whether] the exclusion of evi-
    dence left the jury with sufficient information to
    assess the credibility of the witness.

United States v. Beardslee, 

, 383 (9th Cir. 1999)
(citing United States v. James, 

, 713 (9th Cir.
1998)); see also 

(“A limitation on
cross examination does not violate the Confrontation Clause
9156               UNITED STATES v. LARSON
unless it limits relevant testimony and prejudices the defen-
dant, and denies the jury sufficient information to appraise the
biases and motivations of the witness.”) (internal quotation
marks omitted).

   We review the limitation on the cross-examination of each
witness separately. Cf. Van 

(stating
that the inquiry into the prejudice necessary to establish a
Confrontation Clause violation “must be on the particular wit-
ness”). At trial, Larson’s counsel attempted to cross-examine
Poitra about the minimum sentence she faced in the absence
of cooperation with the Government. However, he was inter-
rupted by the district judge:

    Q.   In fact, you’re going to prison for a minimum
         of five years, right?

    A.   Yes.

    THE COURT: Well, just a minute, counsel. You
       know that the sentencing of defendants in this
       court is the responsibility of the court. And I
       will make the decision about the appropriate
       sentence at the appropriate time. That’s not a
       proper subject of cross-examination.

    [DEFENSE COUNSEL]: Can I ask her about the
       penalty she’s facing?

    THE COURT:        I don’t give advisory opinions.

    [DEFENSE COUNSEL to Poitra]: What’s your
       understanding of the penalties that you’re fac-
       ing?

    [GOVERNMENT]:          Judge, I’m going to object to
       this.
                        UNITED STATES v. LARSON                         9157
      THE COURT:           Sustained.

              And ladies and gentleman, that is
           because in this court, all matters related to
           sentencing are the decision of the court and
           the court only. And that’s the way the law
           is set up and that’s the way it works in this
           court.

   [7] We conclude that while the district court restricted
defense counsel’s cross-examination of Poitra, defense coun-
sel was able to adequately explore Poitra’s motivation to lie
such that the court’s restriction was not an abuse of discretion
and did not violate Defendants’ Confrontation Clause rights.
Poitra testified that she had pleaded guilty and agreed to coop-
erate with the Government in exchange for a motion to reduce
her sentence. She acknowledged that only the Government
had the authority to make such a motion, and that she had a
young child and did not want to go to prison. Poitra was asked
about the mandatory minimum sentence she faced; she
acknowledged that she faced at least five years in prison.
Although the district court firmly ruled that no further exami-
nation would be permitted regarding the actual sentences the
cooperating witnesses faced, it did not strike Poitra’s response.7
Because the jury was sufficiently apprised of Poitra’s incen-
tive to testify to the Government’s satisfaction — including
the mandatory minimum sentence that Poitra faced in the
absence of cooperation with the Government — there was no
Sixth Amendment error.
  7
    The dissent treats the district judge’s remarks as tantamount to striking
Poitra’s answer. The judge, however, never admonished the jury to disre-
gard Poitra’s testimony, nor did any of his comments negate Poitra’s testi-
mony that she faced a five-year minimum sentence in the absence of
cooperation. The court, in its final instructions, instructed the jury to con-
sider all the evidence. The jury was therefore free to take into account that
Poitra faced a five-year minimum sentence in the absence of cooperation
and to give such testimony as much weight as it deemed appropriate. On
this record, we can only assume that the jury followed the court’s instruc-
tions.
9158                      UNITED STATES v. LARSON
   With respect to Lamere, the district court’s prior ruling
made it abundantly clear that it would not permit defense
counsel to cross-examine Lamere regarding the mandatory
minimum life sentence that he faced in the absence of cooper-
ation. Defense counsel heeded the court’s admonition and did
not affirmatively explore the issue with Lamere.8

   [8] The fact that Lamere faced a mandatory minimum sen-
tence of life imprisonment in the absence of cooperation with
the Government was clearly relevant to assessing his testi-
  8
   Defense counsel for Larson examined Lamere regarding his incentive
to cooperate with the Government as follows:
      Q.   And you pled guilty to conspiracy in this case; correct?
      A.   Yes, I have.
      Q.   And you are facing a prison sentence as a result of this con-
           viction as well; right?
      A.   Yes.
      Q.   And by testifying here today, you hope to reduce any prison
           sentence that you may be facing; correct?
      A.   Yes.
      Q.   And in fact, you signed a plea agreement. And the plea
           agreement calls for the government’s assistance in the event
           that it deems your cooperation substantial; right?
      A.   Yep.
      Q.   And you know as well as I do that there’s only one person
           in this courtroom that can even make a motion to reduce
           your sentence; correct?
      A.   Yes, I do.
      Q.   And that’s [the prosecutor] Mr. Thaggard; right?
      A.   Yep.
   We note that, although defense counsel did not attempt to question
Lamere further on the issue of his mandatory life sentence, Defendants did
not forfeit their Confrontation Clause claims in light of the district court’s
clear ruling. Moreover, Defendants moved for the court to reconsider its
ruling at the close of the Government’s case, arguing that their Confronta-
tion Clause rights were violated by the limitation.
                   UNITED STATES v. LARSON                 9159
mony, as it would reveal to the jury Lamere’s potential biases
and motivations for testifying against Defendants.

   [9] In evaluating whether other legitimate interests out-
weighed Defendants’ interest in presenting the evidence, “we
begin by considering the probative value of the evidence.”

. The probative value of a mandatory
life sentence is significant. A cooperating witness who faces
a statutorily mandated sentence of life in prison unless the
government moves for reduction of the sentence has a com-
pelling incentive to testify to the government’s satisfaction.
Thus, the mandatory nature of the potential sentence, the
length of the sentence, and the witness’ obvious motivation to
avoid such a sentence cast considerable doubt on the believa-
bility of the witness’ testimony.

   The district judge’s stated reason for denying cross-
examination of this area of inquiry was that “all matters
related to sentencing are the decision of the court and the
court only.” In the context of this case, that statement was
inaccurate. There was no dispute that, in the absence of a
motion by the Government, the court would be required to
sentence Poitra and Lamere to the statutory minimum sen-
tences for the offenses to which they pled guilty. The Govern-
ment had the discretion to control the potential sentences that
Poitra and Lamere could receive, and the district court’s ratio-
nale cannot form the basis for the exclusion of relevant evi-
dence.

   [10] We recognize, however, the risk that a jury could infer
the potential sentence faced by a defendant from the admis-
sion of testimony regarding a witness’ mandatory minimum
sentence, and that such information could have some influ-
ence on the jury’s deliberative process. See United States v.
Frank, 

, 879 (9th Cir. 1991) (“It has long been
the law that it is inappropriate for a jury to consider or be
informed of the consequences of their verdict.”). Here, there
was added risk because Lamere and Poitra were part of the
9160                  UNITED STATES v. LARSON
same conspiracy to distribute methamphetamine as Defen-
dants. Cf. United States v. Alvarez, 

, 82 (1st Cir.
1993) (recognizing the concern that “such testimony might
mislead or confuse the jury; particularly where, as here, the
witness sought to testify to the same penalties faced by the
defendants”). However, while the Government has an interest
in preventing a jury from inferring a defendant’s potential
sentence, any such interest is outweighed by a defendant’s
right to explore the bias of a cooperating witness who is fac-
ing a mandatory life sentence.9 See United States v. Chandler,

, 223 (3d Cir. 2003) (“[W]hile the government
had a valid interest in keeping from the jury information from
which it might infer [the defendant’s] prospective sentence
were she to be convicted, that interest . . . had to yield to [the
defendant’s] constitutional right to probe the possible biases,
prejudices, or ulterior motives of the witnesses against her.”)
(internal quotation marks omitted).

   Finally, we consider whether the jury was provided with
sufficient information otherwise to assess Lamere’s credibil-
ity. The jury learned that Lamere was both a drug addict and
dealer, and that he had sold at least ten pounds of metham-
phetamine. Lamere admitted that he had been convicted of
seven felonies and that he had served 2-1/2 years in prison.
Lamere testified that he had pleaded guilty and was cooperat-
ing with the Government as a witness against Defendants in
the hope that his sentence would be reduced. He acknowl-
edged that the prosecutor was the only person in the court-
room who could move to reduce his sentence.

   [11] We conclude that, although the above evidence did
cast doubt on Lamere’s credibility, it did not reveal the mag-
nitude of his incentive to testify to the Government’s satisfac-
  9
   To avoid potential confusion, the district court has discretion to pro-
vide the jury with an appropriate limiting instruction or to modify Ninth
Circuit Model Criminal Jury Instruction 7.4, regarding jury consideration
of punishment, as warranted.
                       UNITED STATES v. LARSON                       9161
tion. The permitted cross-examination was therefore
insufficient to allow the jury to assess Lamere’s credibility.
As we explained in Schoneberg:

       Where a plea agreement allows for some benefit or
       detriment to flow to a witness as a result of his testi-
       mony, the defendant must be permitted to cross
       examine the witness sufficiently to make clear to the
       jury what benefit or detriment will flow, and what
       will trigger the benefit or detriment, to show why the
       witness might testify falsely in order to gain the ben-
       efit or avoid the 

. Although the jury here learned that Lamere
stood to benefit by testifying to the Government’s satisfac-
tion, which went a long way “toward satisfying the constitu-
tional requirement,” 

the jury did not learn the
extent to which he stood to benefit from testifying in a man-
ner that satisfied the Government.10 Obviously, an individual
facing life in prison without the possibility of release has an
extremely strong incentive to testify to the Government’s satis-
faction.11 Thus, the jury was deprived of evidence that would
have allowed it to determine just how strong Lamere’s moti-
vation was to please the Government. As in Van Arsdall, “[a]
reasonable jury might have received a significantly different
impression of [the witness’] credibility had . . . counsel been
  10
      We agree with the dissent that the jury’s exposure to information
regarding Poitra’s mandatory minimum sentence “may have actually com-
pounded the court’s error in not allowing any cross-examination of
Lamere on this subject.”
   11
      The concurrence argues that a reasonable jury would have inferred
that Lamere faced “substantial prison time.” There is, however, a funda-
mental difference between a sentence involving “substantial prison time”
with a likelihood of release and life in prison without the possibility of
release, which ensures that no matter how long a person lives, he will live
out the remainder of his days behind bars. The fact that Lamere faced a
mandatory life term therefore cannot be dismissed as being “of slight mar-
ginal utility” even in light of the other information given to the jury.
9162                   UNITED STATES v. LARSON
permitted to pursue his proposed line of 

.

   [12] We have previously held that it is not error for the dis-
trict court to prohibit cross-examination regarding the poten-
tial maximum statutory sentence that the witness faces. See
United States v. Dadanian, 

, 1449 (9th Cir.
1987), rev’d on reh’g on other grounds, 

(9th
Cir. 1988); cf. United States v. Mussare, 

(3d
Cir. 2005) (holding that the district court did not abuse its dis-
cretion in excluding testimony regarding the maximum penal-
ties the government’s witness would have faced in the
absence of government cooperation). In Dadanian, we
referred to such evidence as “at best marginally relevant” to
a witness’ “potential bias and motive in 

.

   [13] The potential maximum statutory sentence that a coop-
erating witness might receive, however, is fundamentally dif-
ferent from the mandatory minimum sentence that the witness
will receive in the absence of a motion by the Government.
The former lacks significant probative force because a defen-
dant seldom receives the maximum penalty permissible under
the statute of conviction.12 In contrast, the fact that here a
  12
     Other cases from outside our circuit are distinguishable on similar
grounds. Most involve the exclusion of testimony regarding the potential
sentence that the witness faced in the absence of cooperation with the gov-
ernment. See, e.g., United States v. Arocho, 

, 636 (7th Cir.
2002), superseded by statute on other grounds as recognized in United
States v. Rodriguez-Cardenas, 

, 959-60 (7th Cir. 2004)
(affirming exclusion of cross-examination of government witnesses
regarding specific sentences and sentencing guideline ranges they faced
before and after cooperation with the government, where the jury had
learned that in exchange for their testimony, the government had dis-
missed several charges against them); United States v. Cropp, 

, 359 (4th Cir. 1997) (affirming exclusion of cross-examination regard-
ing the potential sentences faced by cooperating witnesses); United States
v. Mulinelli-Navas, 

, 988 (1st Cir. 1997) (same); United
States v. Alvarez, 

, 82 (1st Cir. 1993) (affirming exclusion of
cross-examination regarding potential sentence faced by witness where he
“had not actually pleaded guilty and . . . might or might not be found
guilty at a later trial”).
                      UNITED STATES v. LARSON                       9163
cooperating witness faced a mandatory life sentence without
the possibility of release in the absence of a government
motion is highly relevant to the witness’ credibility. It is a
sentence that the witness knows with certainty that he will
receive unless he satisfies the government with substantial
and meaningful cooperation so that it will move to reduce his
sentence.13

   [14] These circumstances are similar to those in Chandler,
an instructive case from the Third Circuit. In Chandler, one
government witness faced a minimum 97-month sentence
under the Sentencing Guidelines, but received only one month
of house arrest in exchange for his guilty plea and 

. The jury heard only that the first witness
pleaded guilty to an offense with a twelve- to eighteen-month
Guidelines sentence range, that he could have been charged
with a greater offense, and that he received one month of
house arrest and probation. See 

witness faced a
Guidelines minimum sentence of twelve years, and testified
that she expected the government to move for a reduced sen-
tence in exchange for her cooperation. See 

court sustained the government’s objections regarding the
penalties that the witnesses faced in the absence of coopera-
tion. Chandler held that the district court violated the defen-
dant’s Confrontation Clause rights because:

       The jury would have had little reason to infer from
       that information that [the first witness’] cooperation
       with the government might have meant the differ-
       ence between more than eight years in prison, on the
       one hand, and the modest sentence he in fact
  13
     Furthermore, unlike a potential sentence range under the Sentencing
Guidelines, a mandatory minimum sentence is readily ascertainable and is
not difficult for a jury to understand. Cf. 

(affirming limitation on cross-examination of potential sentences and sen-
tencing guideline ranges that cooperating witnesses faced, and recognizing
that such a “detailed inquiry could place in dispute many side issues, and
could also confuse the jury as to the real issue at hand”).
9164                  UNITED STATES v. LARSON
       received, on the other. The limited nature of [the wit-
       ness’] acknowledgment that he had benefitted from
       his cooperation makes that acknowledgment insuffi-
       cient for a jury to appreciate the strength of his
       incentive to provide testimony that was satisfactory
       to the prosecution.

the court reasoned, although the jury was aware
of the witnesses’ incentives to lie, they were not aware of the
magnitude of those incentives, which would likely have
affected the jury’s assessment of their testimony. Applying
that reasoning here, any reduction from a mandatory life sen-
tence is of such a significant magnitude that excluding this
information denied the jury important information necessary
to evaluate Lamere’s credibility.14

   [15] Taking the above factors into account, we conclude
that the district court abused its discretion, violating Defen-
dants’ Sixth Amendment constitutional right to effective
cross-examination when it prevented defense counsel from
exploring the mandatory life sentence that Lamere faced in
the absence of a motion by the Government.

B.     Harmless Error Analysis

   Having determined that Defendants’ Confrontation Clause
rights were violated, we next determine whether the error was
harmless beyond a reasonable doubt. Van 

(citing Chapman v. California, 

, 24 (1967)).
  14
    The fact that Lamere had not yet been sentenced is therefore of no
consequence to this determination. We further note that Lamere would
only be sentenced following the Government’s evaluation of the quality
and significance of his testimony at this trial and its subsequent recom-
mendation to the sentencing court regarding a sentence reduction. Thus,
the more incriminating Lamere’s testimony against Larson and Laverdure,
the more he would stand to gain from the Government’s motion on his
behalf.
                   UNITED STATES v. LARSON                  9165
    Whether such an error is harmless in a particular
    case depends upon a host of factors, all readily
    accessible to reviewing courts. These factors include
    the importance of the witness’ testimony in the pros-
    ecution’s case, whether the testimony was cumula-
    tive, the presence or absence of evidence
    corroborating or contradicting the testimony of the
    witness on material points, the extent of cross-
    examination otherwise permitted, and, of course, the
    overall strength of the prosecution’s case.

Id.; see also 

(citing and apply-
ing Van Arsdall factors).

   [16] With respect to Laverdure, we conclude that the Con-
frontation Clause error was harmless beyond a reasonable
doubt. Even without Lamere’s testimony, the Government
offered significant evidence of Laverdure’s involvement in a
conspiracy to possess methamphetamine with intent to distrib-
ute. Poitra’s testimony alone was sufficient to establish his
involvement in the conspiracy. She testified that in 2003,
Laverdure — who knew that Poitra was both a methamphet-
amine user and seller — offered to obtain methamphetamine
for her. She testified that he sold her methamphetamine on a
number of occasions, totaling about 3.5 ounces. In addition,
Gilstrap and Sergeant Kohm testified that Laverdure had sold
methamphetamine to Gilstrap on two occasions.

   Furthermore, while it was constitutional error to prevent
defense counsel from cross-examining Lamere about the man-
datory minimum life sentence that he faced in the absence of
Government cooperation, on cross-examination, defense
counsel did explore Lamere’s criminal past, history as a drug
user and seller, and desire to obtain a lesser sentence through
his testimony against his co-conspirators. We note, too, that
the court instructed the jury that it should view the testimony
of the cooperating witnesses with greater caution than that of
other witnesses.
9166               UNITED STATES v. LARSON
   [17] We also conclude that the Confrontation Clause error
was harmless with respect to Larson. The Government’s case
against Larson was strong even without Lamere’s testimony.
Poitra testified that Larson was the source of the methamphet-
amine that Laverdure obtained for her. Poitra overheard
Laverdure calling “Patty” Larson to obtain the drugs, and
Poitra accompanied Laverdure to an alley by Larson’s house
while he went to Larson’s home and returned with metham-
phetamine. In addition, Poitra testified that the twenty-one
grams of methamphetamine that she sold to the confidential
informant were indirectly obtained from Larson. Furthermore,
Riggs and Detective Gerhart testified as to Riggs’ controlled
purchase from Larson. As discussed above, the cross-
examination of Lamere, and the court’s instruction, also gave
the jury reason to doubt Lamere’s credibility.

  [18] We therefore conclude that the constitutional error was
harmless beyond a reasonable doubt. Accordingly, we
AFFIRM the convictions of Larson (No. 05-30076) and
Laverdure (No. 05-30077).



GRABER, Circuit Judge, with whom RYMER, TALLMAN,
CLIFTON, BEA, M. SMITH, and IKUTA, Circuit Judges,
join, concurring in part and specially concurring in part:

   I agree with the majority opinion except in one major
respect: Defendants Patricia Ann Larson and Leon Nels
Laverdure suffered no Confrontation Clause violation when
the court disallowed a defense question to cooperating co-
conspirator Rick Lee Lamere concerning the statutory mini-
mum prison sentence that he would face in the absence of the
prosecutor’s help in exchange for his testimony. Accordingly,
the district court did not abuse its discretion. I would affirm
on that ground.

   The Supreme Court consistently has held that a Confronta-
tion Clause violation occurs when a trial judge prohibits any
                    UNITED STATES v. LARSON                  9167
inquiry into why a witness may be biased, Delaware v. Van
Arsdall, 

, 679 (1986); Davis v. Alaska, 

, 318 (1974), because “exposure of a witness’ motivation
in testifying is a proper and important function of the constitu-
tionally protected right of cross-examination,” 

-17. The defense must be able to expose through
cross-examination a witness’ motive to lie, because “[a] rea-
sonable jury might . . . receive[ ] a significantly different
impression of [the witness’] credibility” than it would if it had
no basis for understanding why a witness may be biased. Van

. At the same time, the Supreme
Court has taken care to caution that, although a defendant
must be permitted to explore a witness’ biases, “trial judges
retain wide latitude insofar as the Confrontation Clause is
concerned to impose reasonable limits on such cross-
examination based on concerns about, among other things, . . .
prejudice, confusion of the issues, . . . or interrogation that is
. . . only marginally relevant.” 

see also 

(noting that the extent of cross-examination of a
witness is “[s]ubject always to the broad discretion of [the]
trial judge”).

   Applying those principles, we have found a violation of the
Confrontation Clause when a witness had a motive to lie but
the trial judge refused to allow cross-examination concerning
that motive, leaving the jury to infer that the witness had no
motive to lie. United States v. Schoneberg, 

,
1040-43 (9th Cir. 2005) (holding that the limitations placed
on the defendant’s cross-examination, coupled with the dis-
trict court’s “emphatic admonitions,” “vitiated” the defen-
dant’s attempt to establish that the witness had a motive to
lie); see also Fowler v. Sacramento County Sheriff’s Dep’t,

, 1041 (9th Cir. 2005) (concluding that there
was a violation of the Confrontation Clause where all inquiry
into a witness’ bias was foreclosed by the trial judge). Once
a defendant has had an opportunity to question a witness
about bias, we have permitted “reasonable limits,” Van Ars-

, on the scope of cross-examination. No
9168               UNITED STATES v. LARSON
Confrontation Clause violation occurs “as long as the jury
receives sufficient information to appraise the biases and
motivations of the witness.” United States v. Shabani, 

, 403 (9th Cir. 1995) (internal quotation marks omitted).

   According to the majority, a defendant must be permitted
to reveal to the jury not only the fact that the witness has
made a deal with the government to testify in the hope of
receiving a lesser sentence, but also the precise length of the
statutory minimum sentence faced by the witness. “Obvious-
ly,” according to the majority, a lengthy statutory minimum
sentence creates “an extremely strong incentive to testify to
the Government’s satisfaction”; without such evidence, the
jury in this case was unable “to determine just how strong
Lamere’s motivation was to please the Government.” Maj. op.
at 9161. Despite its superficial appeal, the majority’s argu-
ment suffers from two serious flaws.

   First, the majority downplays the significant evidence
before the jury that suggested Lamere’s strong incentive to
lie. The jury knew the following facts to assist them in assess-
ing Lamere’s credibility:

    •   Lamere was 31 years old and had five children;

    •   Lamere was a drug addict and a dealer who had
        sold at least 10 pounds of methamphetamine;

    •   Lamere previously had been convicted of seven
        felonies and had served 2-1/2 years in prison;

    •   Lamere had pleaded guilty to the drug conspiracy
        in the present case;

    •   He faced another prison sentence as a result of
        that guilty plea;

    •   By testifying against Defendants, Lamere hoped
        to reduce his sentence in this case;
                   UNITED STATES v. LARSON                  9169
    •   Lamere had entered into a plea agreement with
        the government. Under his plea agreement, the
        government would assist Lamere in receiving a
        shorter prison term in exchange for his testimony
        against Defendants, but only if the government
        deemed his cooperation to be “substantial”;

    •   As defense counsel colorfully emphasized in
        cross-examining Lamere, “you know as well as I
        do that there’s only one person in this courtroom
        that can even make a motion to reduce your
        sentence”—the prosecutor; and

    •   The other cooperating witness, Joy Lynn Poitra,
        who was merely a novice criminal by contrast to
        Lamere, faced a five-year mandatory minimum
        sentence.

   Knowing those facts, every reasonable juror already would
have inferred that Lamere faced substantial prison time and
that he was extremely eager to please the prosecutor. The
excluded additional fact of the mandatory life term in the
absence of a motion from the prosecutor would not have
given the jury “a significantly different impression of
[Lamere’s] credibility.” Van 

. The
jury knew Lamere’s age, 31, so any lengthy sentence (includ-
ing the 38-year sentence he ultimately received, even with his
substantial assistance to the government) would have been
very bad for him. Depriving the jury of the slight marginal
utility of knowing about the mandatory life term (in the
absence of his cooperation) simply does not equate to a con-
stitutional violation.

   Second, the majority’s reasoning suffers from a logical
flaw. The importance of cross-examination—as the majority
recognizes—is to reveal a witness’ state of mind and, more
particularly, the extent of the witness’ incentive to testify to
the government’s satisfaction. Maj. op. at 9159, 9160. In that
9170                UNITED STATES v. LARSON
regard, the most important piece of the puzzle is the antici-
pated benefit that Lamere expected to receive if his assistance
satisfied the prosecutor. Yet that information was both
unknown and unknowable. The information was unknown
because the district court had not yet sentenced Lamere. The
information was unknowable because defense counsel did not
ask Lamere what benefit he anticipated and, had he been
asked the question, “How much of a reduction in your sen-
tence do you expect to get?,” the only honest answer could
have been, “I don’t know.”

   The statutory minimum sentence that a witness faces is rel-
evant for Confrontation Clause purposes only insofar as it
pertains to the sentence that the witness has received or realis-
tically anticipates receiving. So the majority’s focus on a stat-
utory term of life in prison by itself is a red herring and,
worse, that information in isolation could mislead a jury. The
majority asserts that a long statutory minimum sentence nec-
essarily means a proportionately significant psychological
imperative to assist the prosecutor. Maj. op. at 9160-61. That
assertion is incorrect. The important psychological factor—
the extent of the witness’ motive to lie—is not the statutory
minimum, however great or small, but instead is what the wit-
ness hoped to gain and, realistically, could expect to gain.

   This case provides a good example. Although the statutory
minimum sentence was life in prison, as it turned out, Lamere
was sentenced to a staggering 38 years in prison. Realisti-
cally, given his age, a 38-year prison term was nearly a life
sentence anyhow. Is that benefit—the reduction of Lamere’s
sentence from life to 38 years—as great as the benefit to a
witness facing a three-year statutory minimum on whom the
court imposes only probation? Not necessarily. Again, the key
to understanding a witness’ bias is knowing how great a bene-
fit the witness received or reasonably expected to receive.

  The majority’s reliance on United States v. Chandler, 

(3d Cir. 2003), therefore, is misplaced. Chandler in
                    UNITED STATES v. LARSON                  9171
fact supports the proposition that the extent of the benefit, not
the extent of the theoretical sentence, matters. In Chandler,
the Third Circuit correctly focused on the enormous benefit
that one witness actually had received (one month of house
arrest versus more than eight years in prison) and the corre-
sponding benefit that the second witness reasonably could
anticipate based on the magnitude of the benefit that the first
witness had received. 

The court did not find the
potential sentences, in isolation, significant, but instead rested
its decision on the significant benefits that the witnesses either
had received or reasonably could expect to receive. Id.; see
also United States v. Mussare, 

, 170 (3d Cir.
2005) (reiterating that Chandler “requires an examination of
whether the magnitude of reduction [in the witness’ sentence]
would likely have changed the jury’s mind regarding [the wit-
ness’] motive for testifying” (emphasis added)), cert. denied,

(2006). As noted, the record here does not
answer that key question. Neither Poitra nor Lamere was
asked about the magnitude of an expected benefit, and neither
could have known what it was likely to be. And we now
know, in hindsight, that the benefit that Lamere ultimately
received was not substantial; his 38-year sentence effectively
was almost a life sentence, given his age at the time of sen-
tencing. Cf. United States v. Klauer, 

, 1149-50
(8th Cir. 1988) (holding that no Confrontation Clause viola-
tion occurred by the exclusion of testimony about the “reduc-
[tion of] an otherwise five to seven year sentence to three
years” for a testifying co-conspirator).

   Once a testifying co-conspirator’s biases and motivations
have been exposed, the particular number of years faced (in
the absence of concrete information about the actual benefit
that a witness received or realistically can expect to receive)
is only marginally relevant and is outweighed by concerns of
jury confusion and jury prejudice. The circuit courts, includ-
ing ours, agree on that proposition in a number of contexts,
including the context of statutory maxima. See, e.g., United
States v. Dadanian, 

, 1449 (9th Cir. 1987)
9172                UNITED STATES v. LARSON
(holding that knowledge of the specific maximum term of
imprisonment faced by a witness “is at best marginally rele-
vant” in assessing the witness’ “potential bias and motive in
testifying”), modified on reh’g on other grounds, 

(9th Cir. 1988); United States v. Arocho, 

,
636-37 (7th Cir. 2002) (holding that testimony about the spe-
cific sentences the testifying co-conspirators faced was, “at
best, . . . marginally relevant” and outweighed by the potential
for jury prejudice), superseded by statute on other grounds as
recognized in United States v. Rodriguez-Cardenas, 

, 960 (7th Cir. 2004); United States v. Cropp, 

, 359 (4th Cir. 1997) (holding that “whatever slight addi-
tional margin of probative information gained by quantitative
questions” about the co-conspirators’ potential sentences of
10 and 20 years was outweighed by “the certain [jury] preju-
dice that would result” if the testimony had been allowed);
United States v. Luciano-Mosquera, 

, 1153 (1st
Cir. 1995) (“Any probative value of information about the
precise number of years [the testifying co-conspirator] would
have faced . . . was slight . . . [and] was outweighed by the
potential for [jury] prejudice . . . .”); see also United States v.
Durham, 

, 1334 (10th Cir. 1998) (holding that
additional testimony tending “to show how much prison time
[the testifying co-conspirator] was avoiding through his plea
agreement” was properly excluded by the district court in the
light of the extensive cross-examination that was permitted).

   I see no reason to stray from this universal rule even in the
context of a statutory minimum. The key question is how
much benefit the witness hopes to gain with cooperation, not
how much sentencing discretion the district court retains if the
witness does not cooperate. Consequently, there is no princi-
pled distinction between a statutory maximum and a statutory
minimum faced by a cooperating witness. In either situation,
a witness could receive a huge benefit or a tiny benefit. In
either situation, a district court does not err by ruling that the
probative value of the extra information is substantially out-
weighed by the potential for jury confusion and prejudice.
                    UNITED STATES v. LARSON                 9173
   In summary, I would hold that the jury received enough
information to appraise meaningfully Lamere’s biases and
motives to lie, including his plea agreement and the hoped-for
leniency in sentencing if he succeeded in pleasing the prose-
cutor. The exclusion of the statutory minimum sentence is not
a Confrontation Clause violation in the absence of any means,
other than pure speculation, to assess how much of a benefit
the witness would receive. The district court therefore did not
abuse its discretion when it excluded that portion of Lamere’s
testimony.

   It also bears noting that the majority’s rule is overbroad and
impractical. The rationale that animates its opinion could
apply to any statutory minimum term of years, not just to a
statutory term of life in prison. The strength of a witness’
incentive to assist the prosecutor may be just as great or
greater if she is a young mother facing a 5-year statutory
minimum term, or if he is terminally ill and subject to a 10-
year statutory minimum. Are 5-year and 10-year minima long
enough to require disclosure in cross-examination? The
majority’s opinion does not tell us the answer.

   One practical consequence of that kind of ambiguity is that
prudent trial judges will feel constrained to allow testimony
about cooperating witnesses’ expected statutory minimum
sentences. I question the wisdom of removing significant dis-
cretion from trial judges in the face of the Supreme Court’s
reminders that they are to retain “broad discretion,” 

, and “wide latitude,” Van 

, in controlling cross-examination. I also question the
desirability of that outcome in view of the potential, which the
majority acknowledges, of such testimony to create jury con-
fusion and jury prejudice.

   Additionally, the majority’s rule may have unintended, far-
reaching consequences. What will happen if the court or the
government miscalculates, or a witness misunderstands, the
applicable statutory minimum? If the jury is misinformed,
9174               UNITED STATES v. LARSON
will the defendant have grounds for reversal or collateral
attack if convicted? The majority has, I fear, written a recipe
for unnecessary secondary litigation.

   Finally, even if there were error, it was harmless beyond a
reasonable doubt. Maj. op. at 9165-66. In that respect, as well
as in all others except those discussed above, I agree with the
majority opinion.

  For the foregoing reasons, I concur in part and specially
concur in part.



HAWKINS, Circuit Judge, with whom PREGERSON, REIN-
HARDT, and WARDLAW, Circuit Judges, join, dissenting:

   A conceded street user of drugs is made into a distributor
through the testimony of two witnesses, both facing manda-
tory prison terms. There is nothing particularly new about for-
mer colleagues “singing for their supper” — the problem here
is that the jury did not know of the full extent of the wit-
nesses’ motivation and they were actively misled about what
they did know. The majority properly concludes that what
happened violated the defendants’ Sixth Amendment confron-
tation rights, but ultimately concludes that this constitutional
violation was harmless error. Thus, while I concur in the
majority opinion’s analysis of the standard of review (Section
II) and agree with its conclusion that there was a Confronta-
tion Clause violation with respect to the restrictions on cross-
examining Rick Lamere (Section III.A.), I respectfully dissent
from the majority’s conclusion that the restrictions on the
cross-examination of Joy Poitra did not violate the Confronta-
tion Clause.

  The majority concludes that the jury was sufficiently
apprised of Poitra’s motivation to lie because defense counsel
asked her if she was going to prison for a minimum of five
                      UNITED STATES v. LARSON                       9175
years and she responded “yes.” A review of the transcript
calls into question whether the jury actually even heard or
heeded Poitra’s affirmative answer, as the government con-
ceded at oral argument, because of the court’s immediate inter-
vention.1 But even if the jury heard Poitra answer “yes,” it
was to the question of whether she faced a “minimum” of jail
time — not mandatory jail time — of five years, which only
the government, and not the judge, could cause to be reduced.

  In any event, any supposed benefit from this admission was
immediately undercut by the court’s sua sponte interruption:

      THE COURT:          Well, just a minute counsel. You
                          know that the sentencing of defen-
                          dants in this court is the responsi-
                          bility of the court. And I will make
                          the decision about the appropriate
                          sentence at the appropriate time.
                          That’s not a proper subject of
                          cross-examination.

When defense counsel tried to at least explore Poitra’s own
understanding of the penalty she was facing — clearly rele-
vant to her motivation to lie, regardless of the sentence the
court actually imposed — the government objected and the
court sustained the objection. Moreover, the court further
(and, in light of the mandatory minimums involved in this
case, incorrectly) admonished the jury:

      THE COURT:          And ladies and gentleman, that is
                          because in this court, all matters
                          related to sentencing are the deci-
                          sion of the court and the court
  1
   The Assistant United States Attorney who also prosecuted Larson and
Laverdure at trial argued the appeal. At oral argument before the en banc
court, he stated that, “With respect to Ms. Poitra, I didn’t have a chance
to object before the district court actually did that sua sponte.”
9176               UNITED STATES v. LARSON
                      only. And that’s the way the law is
                      set up and that’s the way it works
                      in this court.

   Because of the court’s statements, counsel was not effec-
tively allowed to convey the fact that Poitra was facing a
mandatory minimum sentence absent a motion by the govern-
ment. Although the jurors may have been aware of Poitra’s
general incentives to lie, they were not fully aware of the
magnitude of those incentives, which could have significantly
affected their assessment of her testimony. Indeed, the little
testimony about sentencing that was allowed — Poitra’s testi-
mony about five years — may have actually compounded the
court’s error in not allowing any cross-examination of Lamere
on this subject, although Lamere was facing a mandatory
minimum sentence of life without parole. Having heard Poitra
mention five years, and being barred from learning anything
about Lamere’s exposure, a rational juror may well have
thought that Lamere only faced five years.

   Because of these concerns, I also cannot agree with the
majority’s conclusion regarding harmless error (Section
III.B.). In assessing the harmlessness of the Confrontation
Clause error, we must consider the trial as a whole, including
the importance of the improperly-restricted witnesses and the
strength of the other evidence presented by the government.
See Van 

.

   Connie Riggs testified that she had purchased a small
amount of methamphetamine from Larson in a controlled buy,
but that even in that sale Larson did not personally have the
drugs and had to go elsewhere to obtain 1/16th of an ounce
for Riggs. On a second attempted purchase by Riggs, Larson
did not have any drugs and said she could not obtain any for
Riggs. Jason Gilstrap testified that in two controlled buys he
had directly purchased a small amount of methamphetamine
(also about 1/16th of an ounce) from Laverdure.
                   UNITED STATES v. LARSON                 9177
   These small sales, however, do not link Larson or Laver-
dure to a much larger conspiracy to distribute drugs with
Poitra and Lamere, which was the sole charge contained in
the indictment. Poitra and Lamere were critical witnesses to
establish the charged conspiracy, and the jury was not suffi-
ciently apprised of their incentive to curry favor with the gov-
ernment. Thus, I cannot say that the Confrontation Clause
error with respect to these crucial witnesses was harmless
beyond a reasonable doubt. 

I would therefore
reverse the convictions of Larson and Laverdure.