United States v. Correa-Alicea
Full Opinion (html_with_citations)
After a six-day trial, appellants Emilio Correa-Alicea and Geraldo Santiago-Torres were convicted of conspiracy to possess with intent to distribute certain narcotics as part of a drug point operating in the Aristedes Charier public housing project in Ponce, Puerto Rico. 1 See 21 U.S.C. §§ 841(a)(1), 841(b)(1)(A), 846, 860. Correa-Alicea was sentenced to a 360-month term of imprisonment, and Santiago-Torres was sentenced to a 324-month term. In this consolidated appeal, Correa-Alicea raises several challenges to his sentence, and Santiago-Torres argues that the cumulative effect of a number of trial errors requires a new trial. We affirm in all respects.
I.
We recite the facts in the light most favorable to the verdict. United States v. Flores-de-JesĂșs, 569 F.3d 8, 16 (1st Cir.2009).
Between November 2005 and November 2006, approximately six drug points operated at the Aristedes Chavier Housing *487 Project. Appellant Correa-Alicea was âin chargeâ of one of the drug points, and appellant Santiago-Torres was a runner at that point, although he also sold drugs on occasion. Correa-Aliceaâs drug point operated every day of the year, from 6:00 or 7:00 a.m. until 11:00 p.m., and had a large number of customers daily. The drug point sold cocaine base as well as other narcotics.
At trial, two government informants testified as to controlled purchases they had made from Correa-Aliceaâs drug point, and the court admitted into evidence the audio recordings of those purchases. AngĂ©lica ColĂłn-GonzĂĄlez, a lifelong resident of the housing project, testified that she had seen appellants selling drugs to others behind her house in the housing project. She stated that she had known appellants her entire life, although on cross-examination, she clarified that she had known Santiago-Torres for three years. She began cooperating with the DEA in early 2006, and on February 16, 2006, she made a controlled drug purchase of six large packages, each of which contained twenty-five to twenty-six individual packages of cocaine base. Although she and the seller, Sergio Martinez, had agreed the previous day that she would purchase fifteen large packages, he had only four packages available on February 16. Martinez then called appellant Santiago-Torres to see whether Santiago-Torres had any additional drugs, and later that day Santiago-Torres brought ColonGonzĂĄlez two more packages of cocaine base. The parties stipulated that the net weight of the cocaine base purchased in the February 16 transaction was 17.1 grams.
Another long-time resident of the housing project, Javier Ortiz-Cruz, had grown up with Correa-Alicea and had known him for most of his life, and had known Santiago-Torres for more than fifteen years. Ortiz-Cruz sold heroin and crack at Correa-Aliceaâs drug point in the fall of 2005, during which time Santiago-Torres also acted as a drug runner and seller. Ortiz-Cruz began cooperating with the DEA in 2006, and, on April 12, 2006, he made a controlled purchase of fifteen packets of heroin. 2 On May 4, 2006, Ortiz-Cruz made a purchase of eight packets of cocaine base from Santiago-Torres, whom he identified as one of Correa-Aliceaâs sellers. Correa-Alicea was also present at this purchase, and the witness identified both Correa-Alicea and Santiago-Torres on the audio recording of the purchase. The parties stipulated that the net weight of the cocaine base purchased in the May 4 transaction was 24.3 grams.
The government also presented the testimony of Eddie Vidal-Gil, a Puerto Rico Police Department agent in charge of the investigation conducted at the housing project. The district court accepted Agent Vidal-Gil as an expert on drug trafficking. Vidal-Gil testified that Correa-Aliceaâs drug point sold cocaine base in small plastic bags, each with a net weight of .07 grams. He further testified, over defense counselâs objection, as to the quantity of drugs sold during his surveillance of the drug point:
*488 Q Okay. And during the times that you actually performed surveillance at the Defendantsâ drug point, how many buys of those little baggies did you see occur during the time period that you were looking at the drug point?
A Well, so many that I couldnât really count them. I would say in a period of 25 minutes or half an hour that I would be surveilling the drug point, many more than 10, 15, of these small bags containing crack would be sold in that period of time.
The jury found appellants guilty of conspiracy to possess with intent to distribute fifty grams or more of cocaine base and detectable amounts of cocaine and heroin in the housing project, within 1,000 feet of a public school. See 21 U.S.C. §§ 841(a)(1), 841(b)(1)(A), 846, 860. At sentencing, the district court found Correa-Alicea accountable for in excess of 4.5 kilograms of cocaine base, and sentenced him to a prison term of 360 months. In a separate sentencing proceeding, the court found Santiago-Torres accountable for in excess of 1.5 kilograms of cocaine base, and sentenced him to a 324-month term.
On appeal, Correa-Alicea challenges his sentence, and Santiago-Torres contends that his conviction should be reversed. We address each appellantâs claims in turn.
II.
A. Correa-Aliceaâs Sentence
Correa-Alieeaâs Presentence Investigation Report (PSR) recommended that he be held accountable for approximately six kilograms of cocaine base over a one-year period. The PSR did not explain how it reached this drug-quantity determination, but instead merely stated that the finding was based on âthe evidence presented at trial, and the testimony of drug expert Eddie Vidal.â
Correa-Alicea filed objections to the PSR, contending that the drug-quantity finding was incorrect because it was based on unreliable expert witness testimony from Agent Vidal-Gil. At the sentencing hearing, after argument from counsel on this point, the district court adopted the PSRâs recommendation and found that Correa-Alicea was responsible for in excess of 4.5 kilograms of crack cocaine. Like the PSR, the district court did not explain how it reached its drug-quantity determination, but instead simply stated that its finding was based on âthe evidence presented at trial.â
Based on the finding that Correa-Alicea was responsible for in excess of 4.5 kilograms of cocaine base, the district court assigned him a base offense level of 38. U.S.S.G. § 2D1.1(c)(1). The court applied a one-level enhancement because the offense was committed in a protected location, U.S.S.G. § 2D1.2(a)(2), and a three-level enhancement because Correa-Alicea was a leader in the criminal activity, U.S.S.G. § 3B1.1(b), for a total offense level of 42. Combined with a Criminal History Category of I, this offense level resulted in a guidelines imprisonment range of 360 months to life, and the court sentenced Correa-Alicea to a prison term of 360 months.
B. Calculating Drug Quantity
Under the sentencing guidelines, the base offense level largely depends upon the total drug quantities involved in the offense. U.S.S.G. § 2D1.1(e). If the quantity of drugs seized does not accurately reflect the scale of a drug-distribution conspiracy, the district court must âapproximate the [total] quantity of the controlled substance.â U.S.S.G. § 2D1.1, cmt. n. 12. Because Correa-Alicea was convict *489 ed of conspiring to distribute controlled substances, he is responsible ânot only for the drugs he actually handled but also for the full amount of drugs that he could reasonably have anticipated would be within the ambit of the conspiracy.â United States v. Santos, 357 F.3d 136, 140 (1st Cir.2004).
A district courtâs finding as to drug quantity âneed only be by a preponderance of the evidence and is not required to be an exact determination but rather only a reasoned estimate.â RodrĂguez, 525 F.3d at 107; see also United States v. Laboy, 351 F.3d 578, 584 (1st Cir.2003) (holding that when âit is impossible or impractical to obtain an exact drug quantity for sentencing purposes, a reasoned estimate will sufficeâ). The sentencing court has âbroad discretion in determining which data are sufficiently dependable for sentencing purposes.â United States v. Rivera-Maldonado, 194 F.3d 224, 228 n. 2 (1st Cir.1999). Although the court âmay rely on reasonable estimates and averagesâ to reach its drug-quantity determination, those estimates must possess adequate indicia of reliability and demonstrable record support. Id. at 228. We âcannot uphold a drug quantity calculation on the basis of hunch or intuition.â United States v. Marrero-Ortiz, 160 F.3d 768, 780 (1st Cir.1998).
We review the district courtâs factual finding as to drug quantity for clear error. RodrĂguez, 525 F.3d at 107. In reviewing the courtâs drug-quantity finding, our task is to âdetermine whether the government presented sufficient reliable information to permit the court reasonably to conclude that [Correa-Alicia was] responsible for a quantity of drugs at least equal to the quantity threshold for the assigned base offense level.â United States v. Barnett, 989 F.2d 546, 553 (1st Cir.1993); see also United States v. Pizarro-BerrĂos, 448 F.3d 1, 8 (1st Cir.2006) (holding that if a district court makes an erroneous factual finding under the sentencing guidelines, yet âthere is enough evidence to support the alternative explanation for the courtâs finding, the error would be harmless and there would be no reason to remand to the district court when the result will be the same.â).
C. Correa-Aliceaâs Challenge to the Drug-Quantity Determination
Correa-AIicea argues that the district court erroneously found that he was individually responsible for more than 4.5 kilograms of cocaine base. Relatedly, he challenges the admission of Agent Vidal-Gilâs testimony, which he claims provided the basis for the courtâs drug-quantity determination. He argues that Vidal-Gilâs testimony was unreliable and not based on recognized scientific techniques, and therefore was improperly admitted as expert testimony. The government responds that, even if Vidal-Gilâs testimony was unreliable, an alternate calculation based on the stipulated quantities of cocaine base purchased in two controlled transactions adequately supports the district courtâs 4.5 kilogram finding. We agree with the governmentâs contention. Because we conclude that the district courtâs drug-quantity determination is sufficiently supported by reliable evidence independent of Vidal-Gilâs testimony, we need not reach the question of whether the district court erred in admitting Vidal-Gilâs testimony. See Barnett, 989 F.2d at 553; see also Pizarro-BerrĂos, 448 F.3d at 8.
At trial, the parties stipulated to the quantity of cocaine base recovered from each of two controlled purchases: on February 16, 2006, ColĂłn-GonzĂĄlez purchased cocaine base with a net weight of 17.1 grams, and on May 4, 2006, Ortiz-Cruz purchased cocaine base with a net weight *490 of 24.3 grams. 3 These controlled purchases were recorded on audio tape, and the audio recordings were admitted as evidence at trial. According to the testimony of Ortiz-Cruz and ColĂłn-GonzĂĄlez, longtime residents of the housing project, Correa-Alicea was âin chargeâ of the drug point and was involved in the conspiracy from November 2005 until November 2006. The drug point operated for at least sixteen hours every day, and a large number of people visited the drug point daily.
Based on this evidence, it is reasonable to conclude that, over the course of the year-long conspiracy, Correa-Aliceaâs drug point made at least one cocaine base sale per day in an amount comparable to the controlled purchases. Using the smaller controlled purchase of 17.1 grams as a sample, 6.2415 kilograms of cocaine base are reasonably attributable to Correa-Alicea over the one-year conspiracy (17.1 grams of cocaine base sold per day, multiplied by 365 days per year). Using the larger controlled purchase of 24.3 grams, he is accountable for 8.8695 kilograms (24.3 grams of cocaine base sold per day, multiplied by 365 days per year). Even the more conservative estimate of 6.24 kilograms is significantly more than the 4.5 kilograms of cocaine base attributed to Correa-Alicea by the district court. The courtâs finding as to drug quantity was not a mere âhunch or intuition,â see MarreroOrtiz, 160 F.3d at 780, but was a âreasoned estimateâ based on reliable evidence in the record. See RodrĂguez, 525 F.3d at 107.
Relying on our decision in Riverar-Maldonado, 194 F.3d 224, Correa-Alicea contends that the quantity of drugs purchased in the two controlled buys does not provide a reasonable basis for the district courtâs 4.5 kilogram finding, and asserts that any estimate of drug quantity based on these two buys âwould not have been arrived at in a scientifically sound and methodologically reliable manner.â We disagree.
In Riverar-Maldonado, we recognized that â[generally speaking, the smaller the sampling, the less reliable the resulting probability estimate,â and an estimate of drug quantity may be unreliable if based on an extrapolation from too small a sample. Id. at 231. We held that it was improper for the sentencing court to estimate the average drug transaction size using a very small sample of controlled buys over the course of a six-month investigation. Id. at 233. The sampling in that case was âminuscule,â âtwelve controlled buys drawn from a set of 86,400 transactions (20 transactions per hour, times 24 hours per day, times 180 days),â and there was no evidence that the twelve controlled buys were representative of ordinary drug transactions at the drug point. Id. at 231-32. Furthermore, the other estimates upon which the total drug-quantity finding was based, including the average number of transactions per hour and average operating hours per day, were also unreliable, and therefore âthe risk of error was compounded by pyramiding unreliable inferences.â Id. at 233. We concluded that the sentencing courtâs drug-quantity determination was not based on sufficiently reliable information, and remanded the case for resentencing. Id. at 233; see also United States v. Culps, 300 F.3d 1069, 1078 (9th Cir.2002) (holding that an estimate of average drug quantity for 60,250 transactions based on nine controlled buys was statistically and legally unreliable); United States v. Butler, 41 F.3d 1435, 1447 (11th Cir.1995) (rejecting an estimate of sixty-six drug transactions per day based solely on a videotape of transactions occurring on a single day, where there was no *491 evidence that single day was âtypicalâ or âaverageâ).
In this case, however, the sample used for calculating average drug transaction size is proportionately much larger, and therefore more reliable, than in RiverarMaldonado: two controlled buys drawn from a set of 365 transactions (one per day for a year), rather than twelve buys drawn from 86,400 transactions. Furthermore, unlike in Riverar-Maldonado, the other estimates upon which the total drug-quantity finding is based are reliable, even conservative. The estimate of one transaction per day, or 365 transactions per year, is highly conservative in light of testimony that the drug point operated for sixteen to seventeen hours per day, 365 days per year, and had a large number of customers daily. The district courtâs ultimate finding that Correa-Alicea was accountable for in excess of 4.5 kilograms of cocaine base is likewise conservative, given that even calculations based on the smaller of the two controlled buys would yield a figure of 6.24 kilograms. The courtâs finding, although not an âexact determinationâ of drug quantity, see RodrĂguez, 525 F.3d at 107, is a reasonable estimate with âdemonstrable record supportâ and âadequate indicia of reliability.â See Rivera-Maldonado, 194 F.3d at 229. 4
In sum, because there is clear record support for the district courtâs finding that Correa-Alicea was accountable for more than 4.5 kilograms of cocaine base during the charged conspiracy, we affirm the sentence imposed.
III.
Santiago-Torres contends that the cumulative effect of several errors requires a new trial. We disagree and affirm his conviction.
A. Voice Identification Expert
Santiago-Torresâs principal argument is that the district court erred in denying his request to retain a voice identification expert and his related motion for a continuance to secure that expertâs presence at trial. We review a denial of a request to fund expert services for abuse of discretion. United States v. Quñones-Medina, 553 F.3d 19, 25 (1st Cir.2009). The denial of a continuance is likewise reviewed for abuse of discretion. United States v. RodrĂguez-DurĂĄn, 507 F.3d 749, 762 (1st Cir.2007).
Prior to trial, upon defense counselâs request, the district court ordered that the defendants and their counsel be permitted to listen on March 6, 2007, to the audio recordings of the controlled purchases. Nearly two months later, on April 30, Santiago-Torres and other defendants filed a motion to continue the trial date set for May 14, citing ongoing investigations and plea negotiations. On May 1, Santiago-Torres submitted a sealed, ex parte motion requesting funds to retain a voice identification expert, asserting that it was not his voice on the incriminating audio recordings, and the only way he could present a defense against the evidence contained in the audio recordings was through such an expert. Two days later, in an order deny *492 ing the April 30 motion for continuance of trial, the court referenced, but did not expressly rule on, Santiago-Torresâs ex parte request for a voice identification expert: âFurthermore, defendantsâ counsel listened to the United Statesâ evidence on March 6, 2007 and only now request[s] an expert on voice identification.â 5
Santiago-Torres renewed his request for a voice identification expert in a second sealed ex parte motion on May 7. The following day, the court issued sealed orders denying both ex parte motions without prejudice pending completion of the governmentâs case in chief. The district court did not further explain the basis for its denial. On May 14, the day trial was scheduled to start, Santiago-Torres filed a motion again requesting funds to retain a voice identification expert and a continuance to ensure the expertâs availability for trial. The court orally denied the motion. At trial, audio recordings of the February 16 purchase by government witness ColĂłn-GonzĂĄlez and the April 12 and May 4 purchases by government witness Ortiz-Cruz were admitted into evidence over defense counselâs objections.
As an indigent, Santiago-Torres was entitled to the benefits of the Criminal Justice Act (CJA), 18 U.S.C. § 3006A. The CJA provides that âa person who is financially unable to obtain ... expert ... services necessary for an adequate defenseâ may obtain them after demonstrating in an ex parte hearing that such services are ânecessary.â 18 U.S.C. § 3006A(e)(l). âGenerally, expert services have been found necessary when the proffered expert testimony was pivotal to the indigent defendantâs defense.â United States v. Manning, 79 F.3d 212, 218 (1st Cir.1996). For example, âcourts have appointed a fingerprint expert when a fingerprint, alleged to be the defendantâs, was the primary means of connecting the defendant to the crime, and a psychiatrist when the defendantâs sanity at the time of the offense was at issue.â Id. (internal citations omitted).
We discern no abuse of discretion in the district courtâs denial of Santiago-Torresâs ex parte motion for funds to retain a voice identification expert, or his related motion for a continuance to secure that expertâs presence at trial. Santiago-Torresâs proffered expert testimony was not pivotal to his defense. The audio recordings were not the âprimary means of connecting [Santiago-Torres] to the crime,â see id., but instead merely served as corroboration for the ample eyewitness testimony that Santiago-Torres had acted as a runner and seller for the drug point.
ColĂłn-GonzĂĄlez testified that she had known Santiago-Torres for three years, and had seen him acting as the runner at the drug point and selling drugs to others behind her house. She further testified that during the February 16 controlled transaction, the seller, Martinez, called Santiago-Torres to see whether he had any additional drugs, and Santiago-Torres brought her two more packages of cocaine base later that day. Ortiz-Cruz testified that he had known Santiago-Torres for more than fifteen years, and had sold drugs at Correa-Aliceaâs drug point in late 2005, during which time Santiago-Torres also worked as a drug runner and seller. Ortiz-Cruz also testified that he purchased cocaine base directly from Santiago-Torres in the May 4 controlled transaction. In light of this eyewitness testimony, none of which depended on identification of San *493 tiago-Torresâs voice on the audio recordings, the voice identification expert services were not critical or necessary to his defense.
Moreover, as the district court noted, Santiago-Torresâs counsel had the opportunity to listen to the audio recordings at issue on March 6, but did not request a continuance until April 30 and did not request funds to retain a voice identification expert until May 1, less than two weeks before his trial was scheduled to begin. Santiago-Torres contends on appeal, as he did in his motion to the district court, that his requests were delayed because he spent weeks trying to locate a voice identification expert available to come to Puerto Rico. However, he does not explain why he could not have made his requests for expert services and for a continuance prior to locating an available expert, and the district court could reasonably have found that the requests were unreasonably dilatory. In light of these facts, we cannot conclude that the court abused its discretion in denying Santiago-Torresâs requests for a continuance and for funds to retain the proffered expert.
Santiago-Torres further contends that the district court erred in disclosing his ex parte request for a voice identification expert in its May 3, 2007 denial of his motion for a continuance. The government correctly concedes error on this point. See United States v. Abreu, 202 F.3d 386, 391 (1st Cir.2000) (holding that district court erred in not handling entire application for expert services under 18 U.S.C. § 3006A(e)(l) on an ex parte basis, and explaining that the âmanifest purpose of requiring that the inquiry be ex parte is to insure that the defendant will not have to make a premature disclosure of his caseâ (internal quotation marks omitted)). However, voice identification was not pivotal to Santiago-Torresâs defense. Even absent the audio recordings, the government presented overwhelming evidence of his guilt. Therefore, we find that the courtâs error in not handling his expert request on an ex parte basis was harmless.
B. Alleged Discovery Violations
Santiago-Torres further contends that the district court failed to remedy two alleged discovery violations by the government.
1. Disclosure of Photograph
First, Santiago-Torres contends that the court erred in not ordering the government to produce a photograph shown to ColĂłn-GonzĂĄlez prior to trial, which he argues was subject to disclosure under Rule 16 of the Federal Rules of Criminal Procedure. We review the courtâs determination under Rule 16 for abuse of discretion. United States v. Caro-Muñiz, 406 F.3d 22, 29 (1st Cir.2005). We find no such abuse here.
ColĂłn-GonzĂĄlez testified that she had personally identified both defendants to DEA agents while she and the agents were riding in an undercover car, and she had later identified Santiago-Torres in a photograph shown to her by DEA agents. Santiago-Torres objected on the ground that, during discovery, he had asked the prosecution for information about any identification procedures used in the investigation, such as photographic line-ups, and the government had responded that no such procedures had occurred. Santiago-Torres contended that the government should have disclosed the photograph during discovery so that he could cross-examine ColĂłn-GonzĂĄlez about her identification, and requested that the court order production of the photograph. The government argued that the photograph was irrelevant because ColĂłn-GonzĂĄlez knew Santiago-Torres from personal interac *494 tions, and did not identify him based on the photograph. The court denied Santiago-Torresâs request, accepting the governmentâs assertion that the photograph was not the basis for her identification of Santiago-Torres because she knew him personally.
The district court did not abuse its discretion in refusing to order production of the pretrial photograph. Rule 16 requires disclosure of documents and objects within the governmentâs possession, custody or control upon a defendantâs request if â(i) the item is material to preparing the defense; (ii) the government intends to use the item in its case-in-chief at trial; or (iii) the item was obtained from or belongs to the defendant.â Fed.R.Crim.P. 16(a)(1)(E). Santiago-Torres contends that the photograph was material to his defense because it could have been used during cross-examination of ColĂłn-GonzĂĄlez and could have formed the basis for a âpretrial motion to suppress the reliability of her identification.â However, nothing in the record indicates that ColĂłn-GonzĂĄlez identified Santiago-Torres on the basis of the photograph. To the contrary, she testified that she had known Santiago-Torres for three years and purchased drugs from him, and that, before being shown the photograph, she had identified him in-person while riding in an undercover car with DEA agents. Thus, the photograph was not material and the government was not required to disclose it.
2. Disclosure of Audio Recording
Santiago-Torres contends that the district court erred in denying his request for a mistrial based on the governmentâs alleged failure to produce the April 12, 2006 audio recording during pretrial discovery. We review the courtâs denial of a mistrial motion for a manifest abuse of discretion. United States v. Van Anh, 523 F.3d 43, 54 (1st Cir.2008).
At trial, after the audio recording of Ortiz-Cruzâs April 12 controlled purchase was played for the jury, defense counsel moved for a mistrial, claiming that he had not received the recording during discovery. He stated that during pretrial discovery he had received a recording marked âApril 12â along with a transcript, but the recording was inaudible. He then requested and received a second copy of the recording and transcript from the government, also marked âApril 12,â but he claimed that this was not the recording played at trial. The government maintained that it had provided defense counsel with copies of the April 12 recording and the transcript on two occasions well in advance of trial, and defense counsel should have notified the government if the recording did not match the transcript. The district court then listened to the recording at issue and denied Santiago-Torresâs motion, explaining that âwhatever the CD was â that was provided â I mean, it was easily determined that it â it wasnât the proper CD by just looking at the transcript.â
We find no abuse of discretion in the courtâs ruling. The court reasonably concluded that, even if the incorrect recording was mistakenly provided, defense counsel could have easily discovered the error and requested the correct recording from the government. Moreover, Santiago-Torres has not shown that he was prejudiced by the governmentâs alleged failure to produce the correct recording prior to trial. See United States v. Devin, 918 F.2d 280, 290 (1st Cir.1990) (â[I]n cases of delayed disclosure, a courtâs principal concern must be whether learning the information altered the subsequent defense strategy, and whether, given timeous disclosure, a more effective strategy would likely have result *495 ed.â). He contends that the governmentâs alleged discovery failure caused him to rely on the wrong trial strategy, because his strategy âwas to obtain suppression of the April 12 recording because it was inaudible and therefore inadmissible.â He further claims that if he had received the correct recording prior to trial, he might have accepted the governmentâs plea offer, âsince the strategy of having the tape excluded on inaudibility grounds would have been foreclosed.â
Santiago-Torresâs claims of prejudice are not plausible. Defense counsel conceded to the district court that after receiving the first recording, which he claimed was inaudible, he requested and received a second copy of the recording from the government. He did not claim that this second recording was inaudible; instead, he conceded that he listened to the recording and âunderstood that this transcript was not really right, that you could not really read â you could not really hear what is in the transcript.â In addition, defense counsel was permitted to listen to the governmentâs audio recordings in court on March 6, 2007, prior to trial. Defense counsel admitted to the district court that he âwasnât present for the whole session.â Finally, although Santiago-Torres claims that his trial strategy was to file a motion to suppress the audio recording, he did not file any such pretrial motion. Santiago-Torres has failed to credibly demonstrate that the alleged discovery failure caused him to adopt a less effective defense strategy, and we are therefore satisfied that any discovery failure was harmless.
Affirmed.
. Correa-Alicea and Santiago-Torres were indicted along with six co-defendants. The six co-defendants entered into plea agreements, while Correa-Alicea and Santiago-Torres proceeded to a joint trial.
. The government acknowledges in its brief that Ortiz-Cruz provided inconsistent testimony as to whom he purchased heroin from on April 12. On direct examination, he stated that he bought drugs from "Colin [Correa-Alicea] and Geraldo [Santiago-Torres]â on or about April 12. However, he subsequently claimed that he had purchased the drugs on April 12 "[f]rom Colin,â which he then repeated on cross-examination. The government now takes the position that the April 12 purchase did not involve Santiago-Torres at all, noting that the transcript of the audio recording of this purchase does not list Santiago-Torres as a speaker.
. The third controlled purchase, made by Ortiz-Cruz on April 12, 2006, was for heroin rather than cocaine base, and therefore is not relevant to this calculus.
. Although Correa-Alicea does not raise the argument, it could be contended that an estimate of drug quantity based on controlled buys arranged by government agents is particularly unreliable, as there is no evidence that these buys are representative of sales by ordinary customers. See Rivera-Maldonado, 194 F.3d at 232. However, even if the controlled buys stipulated to in this case are somewhat larger than the average buys by ordinary customers at Correa-Alicea's drug point, the calculation outlined above more than compensates for any overestimate in the average drug transaction size by using the extremely conservative estimate of just one drug purchase per day.
. Santiago-Torres also contends on appeal that the district court erred by stating in this order that Santiago-Torres "requests] an expert on voice identification,â thereby disclosing his ex parte request for a voice identification expert. We discuss this claim of error below.