OWEN, Circuit Judge:

Dimetriace Eva-Lavon John was found guilty by a jury on all counts of a seven-count indictment arising out of her involvement in a scheme to incur fraudulent charges on accounts held by various Citigroup customers. John challenges her convictions and sentence in this appeal. We affirm the convictions but vacate her sentence and remand for further proceedings.

I

Dimetriace Eva-Lavon John was employed as an account manager at Citigroup for approximately three years. By virtue of her position, she had access to Citigroup's internal computer system and customer account information contained in it. In September 2005, John provided Leland Riley, her half-brother, with customer account information enabling Riley and other confederates to incur fraudulent charges.

John accessed and printed information pertaining to at least seventy-six corporate customer accounts and provided it to Riley. The information was in the form of either scanned images of checks written by the account holders or printouts of computer screens containing detailed account information. Before he was apprehended, Riley and cohorts used information John had provided to incur fraudulent charges on four different accounts.

A grand jury returned a seven-count indictment against John. Count 1 charged John with conspiracy to commit access device fraud in violation of 18 U.S.C. § 371. Counts 2 through 5 charged John with fraud in connection with an access device and aiding and abetting, in violation of 18 U.S.C. §§ 1029(a)(5) and (2). Counts 6 and 7 charged John with exceeding authorized access to a protected computer in violation of 18 U.S.C. §§ 1030(a)(2)(A) and (C). A jury found John guilty on all seven counts.

A Presentence Report (PSR) concluded that the Sentencing Guideline applicable to the conspiracy count was § 2X1.1(a), which provides that the base offense level is that applicable to the substantive offense. The substantive offense underlying the conspiracy count — a violation of 18 U.S.C. § 1029(a)(5) — is governed by § 2B1.1 of the Guidelines, which provides for a base offense level of six. However, in calculating the advisory Guidelines sentencing range, the PSR recommended that the base offense level be increased by 16 levels because the PSR concluded that John intended to cause a loss of approximately $1,451,865. The PSR also determined that John intended to obtain account holders' personal information and accordingly added two levels pursuant to § 2B1.1(b)(14)(A)(i)(II). After other adjustments that are not at issue in this appeal, the PSR arrived at a final base offense level of thirty. John had no criminal history, and the resulting advisory Guidelines range of imprisonment was 97-121 months. The district court ultimately sentenced John to 108 months' imprisonment.

II

John has raised several issues regarding her convictions. Her first contention is that the evidence was insufficient to support her convictions on Counts 6 and 7 under 18 U.S.C. § 1030(a)(2) for exceeding authorized access to Citigroup's comput ers. She candidly acknowledges that at trial her counsel failed to renew a motion for acquittal at the close of the evidence and that we therefore may only reverse her convictions on these counts "if there was a `manifest miscarriage of justice,' which would occur if there is no evidence of the defendant's guilt or `the evidence on a key element of the offense was so tenuous that a conviction would be shocking.'"

Whether John's convictions on Counts 6 and 7 may be sustained depends on the proper interpretation of "exceeds authorized access" as used in § 1030(a)(2) and defined in § 1030(e)(6).

John was convicted of violating § 1030(a)(2), which provides:

(a) Whoever —

. . .

(2) intentionally accesses a computer without authorization or exceeds authorized access, and thereby obtains —

(A) information contained in a financial record of a financial institution, or of a card issuer as defined in section 1602(n) of title 15, or contained in a file of a consumer reporting agency on a consumer, as such terms are defined in the Fair Credit Reporting Act ( 15 U.S.C. 1681 et seq.); . . .

shall be punished as provided in subsection (c) of this section.

The term "exceeds authorized access" is defined in § 1030(e)(6): "the term `exceeds authorized access' means to access a computer with authorization and to use such access to obtain or alter information in the computer that the accesser is not entitled so to obtain or alter. . . ."

John argues that she was authorized to use Citigroup's computers and to view and print information regarding accounts in the course of her official duties. The evidence, she contends, reflects only that she was not permitted to use the information to which she had access to perpetrate a fraud, she could make changes to account information only in compliance with a customer's request, and she was not permitted to take material she printed regarding accounts from her office building. She asserts that her mental state or motive at the time she accessed or printed account information cannot determine whether she violated 18 U.S.C. § 1030(a)(2). Specifically, she argues that the statute does not prohibit unlawful use of material that she was authorized to access through authorized use of a computer. The statute only prohibits using authorized access to obtain information that she is not entitled to obtain or alter information that she is not entitled to alter, John contends.

We first note; hat John was not charged in Counts 6 or 7 with altering information in Citigroup's computer system. She was charged with "exceeding authorized access" and obtaining confidential Citigroup and Home Depot customer account information.

The statute at issue prohibits both accessing a computer "without authorization" and "exceeding] authorized access" to obtain specified information. The statute does not deiine "authorized," or "authorization," which is used in the definition of "exceeds authorized access." The question before us is whether "authorized access" or "authorization" may encompass limits placed on the use of information obtained by permitted access to a computer system and data available on that system. We conclude that it may, at least when the user knows or reasonably should know that he or she is not authorized to access a computer and information obtainable from that access in furtherance of or to perpetrate a crime.

To give but one example, an employer may "authorize" employees to utilize computers for any lawful purpose but not for unlawful purposes and only in furtherance of the employer's business. An employee would "exceed[] authorized access" if he or she used that access to obtain or steal information as part of a criminal scheme.

In United States v. Phillips, this court analyzed whether a criminal defendant had accessed university computers "without authorization" in violation of § 1030(a)(5)(A)(ii), as distinguished from "exceed[ing] authorized access," and we recognized that "[c]ourts have . . . typically analyzed the scope of a user's authorization to access a protected computer on the basis of the expected norms of intended use or the nature of the relationship established between the computer owner and the user." We applied this "intended-use analysis" to conclude that a student who used his privilege of access to a university's computer was not authorized to access parts of the system to which he had not been given a password. John's situation differs from that of the student in Phillips because John was authorized to view and print all of the information that she ac cessed and that she provided to Riley. However, John's use of Citigroup's computer system to perpetrate fraud was not an intended use of that system.

John's use of Citigroup's computer system to perpetrate a fraud was also contrary to Citigroup employee policies, of which she was aware. The First Circuit has held that an employment agreement can establish the parameters of "authorized" access. In EF Cultural Travel BV v. Explorica, Inc., the plaintiffs brought a civil action under the Computer Fraud and Abuse Act (CFAA) seeking injunctive relief against former employees who had become competitors. The former employees used their knowledge of codes that they had obtained while in their former employment to create a high-speed computer program to mine their former employer's public website for pricing information. The former employees had entered into a broad confidentiality agreement with their former employers protecting proprietary information. The First Circuit held "that because of the broad confidentiality agreement [the former employees'] actions `exceed[ed] authorized access'" within the meaning of § 1030(a)(4). The court reasoned, "[the former employees'] wholesale use of EF's travel codes to facilitate gathering EF's prices from its website reeks of use — and, indeed, abuse — of proprietary information that goes beyond any authorized use of EF's website." The court continued, "[i]f EF's allegations are proven, it will likely prove that whatever authorization [former employees] had to navigate around EF's site (even in a competitive vein), [they] exceeded that authorization by providing proprietary information and know-how to [a programmer] to create the scraper."

While we do not necessarily agree that violating a confidentiality agreement under circumstances such as those in EF Cultural Travel BV would give rise to criminal culpability, we do agree with the First Circuit that the concept of "exceeds authorized access" may include exceeding the purposes for which access is "authorized." Access to a computer and data that can be obtained from that access may be exceeded if the purposes for which access has been given are exceeded. In other words, John's access to Citigroup's data was confined. She was not authorized to access that information for any and all purposes but for limited purposes.

In the present case, the Government demonstrated at trial that Citigroup's official policy, which was reiterated in training programs that John attended, prohibited misuse of the company's internal computer systems and confidential customer information. Despite being aware of these policies, John accessed account information for individuals whose accounts she did not manage, removed this highly sensitive and confidential information from Citigroup premises, and ultimately used this information to perpetrate fraud on Citigroup and its customers.

We recognize that the Ninth Circuit may have a different view of how "exceeds authorized access" should be construed. In LVRC Holdings LLC v. Brekka, a civil proceeding, the Ninth Circuit construed 18 U.S.C. § 1030(a)(2) to mean that "a person who `intentionally accesses a computer without authorization,' §§ 1030(a)(2) and (4), accesses a computer without any permission at all, while a person who `exceeds authorized access,' id., has permission to access the computer, but accesses information on the computer that the person is not entitled to access." That court stated that "[t]he definition of the term `exceeds authorized access' from § 1030(e)(6) implies that an employee can violate employer-placed limits on accessing information stored on the computer and still have authorization to access that computer." In Brekka it was alleged that an employee e-mailed to his and his wife's personal computers proprietary documents to which his employer had given him access with the intention of using the information to compete with his employer once he resigned. The court rejected the argument that one who is authorized to obtain information stored in a computer exceeds authorized access within the meaning of 18 U.S.C. § 1030(a)(2) "if the defendant breaches a state law duty of loyalty to an employer" in accessing and using that information "to further his own competing business."

The Ninth Circuit's reasoning in Brekka was influenced by its recognition that "[f]irst, and most important, § 1030 is primarily a criminal statute, and §§ 1030(a)(2) and (4) create criminal liability for violators, of the statute." The court explained its view that, "[a]lthough this case arises in a civil context, our interpretation of [the statute] is equally applicable in the criminal context," and that "ambiguity concerning the ambit of criminal statutes should be resolved in favor of lenity." The Ninth Circuit explained that "[i]f the employer has not rescinded the defendant's right to use the computer, the defendant would have no reason to know that making personal use of the company computer in breach of a state law fiduciary duty to an employer would constitute a criminal violation of the CFAA. It would be improper to interpret a criminal statute in such an unexpected manner."

There are no such concerns in the present case. An authorized computer user "has reason to know" that he or she is not authorized to access data or information in furtherance of a criminally fraudulent scheme. Moreover, the Ninth Circuit's reasoning at least implies that when an employee knows that the purpose for which she is accessing information in a computer is both in violation of an employer's policies and is part of an illegal scheme, it would be "proper" to conclude that such conduct "exceeds authorized access" within the meaning of § 1030(a)(2).

III

At trial, the Government presented testimony from an expert witness who opined that John's fingerprints were on Citigroup documents that were found in Riley's possession. John contends that the district court erred in admitting this testimony because it was never established that the evidence was reliable. John asserts that the error is not harmless because this evidence was necessary to connect John to the fraudulent scheme.

The admission of expert evidence is reviewed for abuse of discretion; however, evidentiary rulings are subjected to heightened scrutiny in criminal cases. Federal Rule of Evidence 702 provides that testimony by a qualified expert is admissible if (1) it will assist the trier of fact; (2) it is based upon sufficient facts or data; (3) the testimony is a product of reliable methods; and (4) the witness has applied those principles reliably to the facts. It is the gatekeeping responsibility of the trial judge to ensure that any admitted scientific testimony or evidence is both relevant and reliable.

The Supreme Court has articulated a non-exclusive list of factors that a district court may consider in determining whether expert evidence is reliable: (1) whether the expert's technique can be tested; (2) whether the technique has been subject to peer review; (3) known or potential rate of error associated with the technique; (4) the existence of standards or controls; and (5) whether the technique or theory has been generally accepted in the scientific community. This test is flexible, and the proponent of the expert testimony need not satisfy each factor. Further, a district court has broad latitude in deciding how to determine reliability, as well as in its ultimate reliability determination.

John's threshold argument is that the district court "abdicated its gatekeeping function" by failing to hold a Daubert hearing on the matter. However, we agree with a number of our sister circuits that have expressly held that in the context of fingerprint evidence, a Daubert hearing is not always required. As the Seventh Circuit has noted: "Those [courts] discussing the issue have not excluded fingerprint evidence; instead, they have declined to conduct a pretrial Daubert hearing on the admissibility of fingerprint evidence or have issued brief opinions asserting that the reliability of fingerprint comparison cannot be questioned."

We agree that in most cases, absent novel challenges, fingerprint evidence is sufficiently reliable to satisfy Rule 702 and Daubert. "Fingerprint identification has been admissible as reliable evidence in criminal trials in this country since at least 1911." In terms of specific Daubert factors, the reliability of the technique has been tested in the adversarial system for over a century and has been routinely subject to peer review. Moreover, as a number of courts have noted, the error rate is low. The district court did not err in dispensing with a Daubert hearing.

John also asserts that the fingerprint evidence was inadmissible because the Government's expert did not explain how many matching points were required to determine that the prints were hers. She asserts that because of the expert's silence on the quantitative standard, the Daubert factors cannot be satisfied because the expeit's technique is by definition standardless, untestable, and its error rate indeterminable.

We are unpersuaded by this argument because there is no universally accepted number of matching points that is required for proper identification, as this varies depending on the quality of the print. A number of circuits have determined that this "sliding-scale" procedure is testable, generally accepted, and sufficiently reliable and that its known error rate is essentially zero.

Moreover, although the Government's fingerprint expert did not testify to a precise number of matching points, contrary to John's assertion, he gave extensive testimony regarding the uniqueness of finger-prints generally, as well as the particular identification methodology used. In fact, the expert provided a detailed step-by-step description of the identification process focusing on one specific print. As to this particular example, the expert also pointed out eight specific matching points to the jury and noted that more were found.

John challenges the reliability of the Government's fingerprint evidence because it was not subject to "blind verification," which requires a second expert to match the prints without being told the results of the original test. However, we have not located any case law supporting John's assertion that blind verification is required.

John has not demonstrated that the district court abused its discretion by admitting the fingerprint expert's testimony. Issues regarding the accuracy of fingerprint evidence in a particular case generally go "to the weight and credibility of the evidence" and are "best left to the finder of fact, not an appellate court." John had the opportunity to analyze the fingerprint evidence herself and question its validity, and she availed herself of this opportunity by introducing her own expert at trial who testified regarding all the issues John now raises on appeal.

Even if the district court erred in this respect, the error was harmless. The Government demonstrated at trial that all the account information in question was printed from John's computer on days that John was at work. Along with other evidence presented at trial, the jury could have reasonably reached the same verdict.

IV

During trial, the defense sought to introduce evidence that allegedly would have demonstrated that John's half-brother Riley had other inside sources at Citigroup. John contends that this evidence would have created reasonable doubt as to whether John was the informant who provided Riley with customer account information. To raise this doubt, the defense sought to have police officers who interrogated Riley testify to statements he allegedly made. The Government objected on the basis that this testimony would be hearsay, and those objections were sustained.

John does not argue that this testimony would not have been hearsay or that it would have fallen within a hearsay exception. More importantly, John has not offered any proof as to what the officers' testimony would have been and accordingly cannot prevail on a direct challenge to the district court's ruling. Nevertheless, John attempts to frame her evidentiary challenge as a constitutional issue. She argues that adherence to the hearsay rule in this case violated her right to "present a complete defense" under the Due Process Clause and the Compulsory Process Clause of the Sixth Amendment.

The Supreme Court has stated that a defendant's right to present a complete defense "is abridged by evidence rules that infringe upon a weighty interest of the accused and are arbitrary or disproportionate to the purposes they are designed to serve." However, "the accused . . . must comply with established rules of procedure and evidence designed to assure both fairness and reliability in the ascertainment of guilt and innocence."

John relies on Chambers v. Mississippi, in which the Supreme Court held that, given the particular facts of that case, exclusion of trustworthy hearsay evidence critical to the defense, coupled with an unrelated error, resulted in a denial of due process. However, the facts of Chambers are distinguishable. In Chambers, the district court excluded testimony, on hearsay grounds, by three different parties who would have testified that an individual other than the defendant admitted to committing the murder in question. The Court found that these particular statements, which were offered at trial, were made in "circumstances that provided considerable assurance of their reliability." In particular, the Court noted that (1) the confessions were made spontaneously to close acquaintances shortly after the murder; (2) each statement was corroborated by other evidence; (3) the same confession was made three times; (4) the confessions were against interest; and (5) the person who allegedly made the confession was available to testify. No such evidence is present in this case. John offers only speculation as to what the witnesses in question would have said had they been allowed to testify

Moreover, the Court's decision in Chambers did not rest solely on the hearsay issue. In Chambers, the defendant called the individual who confessed to the murder as a witness, but when cross-examined by the State, the witness repudiated his confession. The defendant, in turn, could not redirect because of the state's "voucher" rule, which barred parties from impeaching their own witnesses. Thus, in combination with the hearsay rule, the voucher rule — which the Court noted bore little relationship to the realities of the criminal process — completely prevented the defendant from impeaching a highly damaging witness. The Court's ultimate holding that the defendant was denied a fair trial rested on the combined effects of the hearsay and voucher rules. John's situation is not analogous to Chambers. The district court did not err.

V

John contends her convictions must be reversed because the Vice President of Citigroup for Risk and Commercial Operations, Mitch Raymond, improperly speculated as to what John's mental state must have been, suggesting to the jury that she acted with fraudulent intent. John also argues that Raymond's testimony was un-supported lay opinion in violation of Federal Rule of Evidence 701. John takes issue with the following exchange:

Prosecution: But assume [the account] had [an administrative block code], that date of 8/1/2003, and assume a call came in from somebody who was unknown to your person answering the phone, was really a criminal engaged in a fraud scheme and they knew how they manipulate your system, would an honest employee getting such a call who looked on the screen and saw [the block code] and this GE acquisition date of 8/1/03, would they just automatically go in and change the block code from A to blank?

Raymond: No. They would know, and it was in our procedures that they would have to have the customer reapply. So that call probably would have been transferred to our credit group.

Prosecution: So an honest employee getting such a call that came in on 11/1/2005, would they have taken the action that [John] took [and removed the block from the account]?

Raymond: That would not have been an appropriate action.

We review the district court's evidentiary rulings for abuse of discretion, subject to the harmless-error rule. Pursuant to Rule 701, "a lay opinion must be based on personal perception, must be one that a normal person would form from those perceptions, and must be helpful to the [fact finder]." A lay opinion of a corporate officer may be based on knowledge acquired as a result of an individual's employment position.

Although Raymond's testimony was based on his experience in a supervisory position, he did express an opinion regarding what an employee acting properly within the scope of employment would have done in a particular situation. To the extent John argues that this opinion testimony was speculative because it involved the probable actions of an unidentified "honest" employee, her argument may be well taken. However, we need not decide this issue because any error was harmless. Raymond testified that Citigroup's procedures and training programs clearly stated that accounts that had been dormant for a certain period of time could not be altered until the account holder renewed its application with the credit department because such accounts were not considered active. From this testimony, the jury could reasonably infer that an employee acting within the scope of his or her duties would have followed this procedure. Additionally, given other evidence presented at trial, including the fact that John altered these accounts shortly after receiving calls from Riley and that she never returned to work after Riley's first arrest, the jury could reasonably infer that John was acting with fraudulent intent.

VI

John raises a number of issues with regard to her sentence. She contends that the district court erred in determining that the intended loss was $1,451,865. This intended loss amount resulted in a sixteen-level increase to her base offense level under § 2B1.1(b)(1) of the Guidelines. The actual amount of loss was $78,750, which would have resulted in an eight-level increase. The district court adopted the PSR, which arrived at the intended loss by aggregating the credit limits on all seventy-six accounts for which John printed and disseminated information and then added the amount by which the credit limit was exceeded on one account. Charges totaling $78,750 were actually incurred on only four of these accounts. John suggests several different loss amounts that the district court could alternatively have calculated, all resulting in less than a sixteen-level increase in her base offense level.

The Sentencing Guidelines applicable to fraud offenses are contained in § 2B1.1. The Guidelines' commentary explains that under § 2B1.1, "loss serves as a measure of the seriousness of the offense and the defendant's relative culpability and is a principal factor in determining the offense level under this guideline." Accordingly, § 2B1.1(b)(1) creates a sliding scale that increases the defendant's base offense level by zero to thirty points depending on the amount of loss. The Guidelines commentary explains that for purposes of that section, "loss is the greater of actual or intended loss."

In ascertaining the intended loss, the district court must determine the defendant's actual intent. A determination of the loss amount is a factual finding reviewed for clear error. Under this standard, "as long as the determination is plausible in light of the record as a whole, clear error does not exist." Moreover, the Guidelines provide that in determining the amount of loss, "[t]he court need only make a reasonable estimate." Accordingly, we give district courts "wide latitude" in this regard. "The method used to calculate the amount of loss, however, must bear some reasonable relation to the actual or intended harm of the offense."

We have held that in estimating intended loss in fraud cases, the district court may look beyond the monetary amount the defendant actually obtained or clearly attempted to obtain prior to being apprehended. For example, in United States v. Chappell, the defendant, who was convicted for cashing a number of counterfeit checks, was also found with fifty-one counterfeit blank checks in his possession. In calculating the intended loss, the district court accounted for the blank checks by assigning to each the average value of the checks that were actually forged and cashed. We held that the district court's assessment "manifestly was not clearly erroneous" in light of the fact that the defendants had already attempted to negotiate at least 25 counterfeit checks and evidence indicated that they had planned to negotiate over 150 more.

Similarly, in United States v. Sowels, we held that a district court's determination that intended loss was the combined credit limit of all credit cards stolen by the defendants was not clearly erroneous, even though the defendants were apprehended before they could use any of the cards. We noted the district court's "difficult task of projecting into the future Sowels's intent as to the extent to which he would use the cards" and stated that because Sowels had not completed or withdrawn from the offense before apprehension, the district court did not clearly err in concluding that he intended to charge the cards to their limit.

More generally, in this context we have often emphasized that the defendant's actions placed the victim at risk of loss for the full amount and relatedly, we have considered the defendant's ability to access the entirety of the funds in question. John argues that merely obtaining information regarding a credit account does not create a risk of loss as to the entire account limit or establish sufficient access to infer intended loss up to the account limit as a matter of law. Specifically, she argues that obtaining information is far more preparatory in nature than the actual possession of credit cards or checks, which allows for immediate access to the account. However, immediate access and risk of loss are not dispositive in all circumstances; rather, the critical determination is the intent of the defendant. We have expressly rejected the argument that a defendant could not be held accountable for attempted credit-card charges declined for exceeding the credit limit because the defendant never had actual access nor created any risk of loss.

Nevertheless, John maintains that the district court's determination of intended loss was clearly erroneous in light of the record because the Government did not carry its burden of proving John's intent to utilize all seventy-six accounts that she accessed and about which she disseminated information. She points out that her confederates actually obtained funds from only four accounts and that out of these accounts only one was used to its maximum credit limit. John urges that the Government at most demonstrated that each of the seventy-six accounts was a potential candidate for fraud, not that the defendants ever intended to utilize all of the accounts. She asserts that additional action would have been required to allow the co-conspirators to incur any charges on the accounts and that because such action had not yet been taken, the district court was clearly erroneous in its determination of intended loss.

John has not demonstrated that the district court clearly erred. For forty-four of the seventy-six relevant accounts, she obtained printouts of account computer screens, which contained account numbers, names of account holders and their telephone numbers, account holder titles, billing addresses, credit lines, last payment dates, the existence of any administrative blocks on the account and other information that could be used in the fraudulent scheme. These printouts contained the same type of information from which John's confederates were able to have other accounts altered, obtain access, and incur charges. The district court's conclusion that John's co-conspirators simply did not have time to take the requisite actions before they were discovered is plausible in light of the record, and John "should not be rewarded simply because law enforcement officials thwarted [her] plans."

John argues somewhat more persuasively with respect to the thirty-two accounts about which the information disseminated involved only images of checks written in payment of the account balances. John asserts that these check images provide limited account information and that several additional steps would have to be taken before charges could be incurred upon these accounts. John also points out that the check images were printed well before most of the account screens were printed and that the check images were retained for over three months without use, and that when only check images regarding an account were obtained, the account was not accessed, evidencing a lack of intent to ever do so.

John's arguments, however, offer no explanation of why the check images were printed in the first place if they were indeed useless in furthering the fraudulent scheme. John's intentions in providing this information to her partner in crime were not benign. John has not rebutted the evidence that she intended to maximize charges on these accounts as well. In light of the record, it is plausible that the co-conspirators simply had not yet mined these accounts. Accordingly, we cannot conclude that the district court clearly erred in making its intended loss determination.

VII

John contends that even if the district court did not clearly err in finding the amount of the intended loss, the district court should have applied a three-level reduction for a "partially completed offense" based on note 17 in the Commentary to § 2B1.1 of the Guidelines. John asserts that she was convicted of discrete credit card transactions, not of a broad scheme to defraud such as mail fraud or wire fraud, and that only eight Citigroup accounts were either altered or accessed by her or her confederates. With regard to the sixty-eight other accounts, the district court found only an intent to access the credit limits. She asserts that as to these sixty-eight accounts, the offenses were not completed. The Government's briefing does not address John's contentions.

Note 17 in the Commentary accompanying § 2B1.1 provides:

Partially Completed Offenses. — In the case of a partially completed offense ( e.g. an offense involving a completed theft or fraud that is part of a larger, attempted theft or fraud), the offense level is to be determined in accordance with the provisions of § 2X1.1 (Attempt, Solicitation, or Conspiracy) whether the conviction is for the substantive offense, the inchoate offense (attempt, solicitation or conspiracy), or both. See Application Note 4 of the Commentary to § 2X1.1.

Application Note 4 of the Commentary to § 2X1.1 provides in pertinent part:

In certain cases, the participants may have completed (or have been about to complete but for apprehension or interruption) all of the acts necessary for the successful completion of part, but not all, of the intended offense. In such cases, the offense level for the count (or group of closely related multiple counts) is whichever of the following is greater: the offense level for the intended offense minus 3 levels (under § 2X1.1(b)(1), (b)(2), or (b)(3)(A)), or the offense level for the part of the offense for which the necessary acts were completed (or about to be completed but for apprehension or interruption). For example, where the intended offense was the theft of $800,000 but the participants completed (or were about to complete) only the acts necessary to steal $30,000, the offense level is the offense level for the theft of $800,000 minus 3 levels, or the offense level for the theft of $30,000, whichever is greater.

Additionally, the Commentary provides:

Background: In most prosecutions for conspiracies or attempts, the substantive offense was substantially completed or was interrupted or prevented on the verge of completion by the intercession of law enforcement authorities or the victim. In such cases, no reduction of the offense level is warranted. Sometimes, however, the arrest occurs well before the defendant or any conspirator has completed the acts necessary for the substantive offense. Under such circumstances, a reduction of 3 levels is provided under § 2X1.1(b)(1) or (2).

The PSR prepared prior to John's sentencing, which the district court adopted, concluded, "[t]he defendant and codefendants completed all acts necessary for the successful completion of the substantive offense, therefore a 3-level decrease is not applicable under USSG § 2X1.1(b)(2)." John made no objection in the district court to this conclusion or the failure to apply a three-level decrease, and we accordingly review for plain error. As part of that review, the district court's interpretation and application of the Guidelines is considered de novo.

In determining whether an offense was completed when a conspiracy has been alleged, "we consider the degree of completion of the underlying offense, and not of the conspiracy itself." The underlying offense of fraud and related activity in connection with access devices requires an actual loss, and it requires use of an access device. There were actual losses with regard to only four Citigroup accounts, and access devices were used only with regard to those accounts. The conspiracy indictment included as overt acts John's actions in providing information regarding not only those accounts but other accounts that were not accessed. The question is whether the offense with respect to the seventy-two accounts that were not accessed was "partially completed" within the meaning of note 17 under § 2B1.1(b)(2) or "completed" within the meaning of note 4 and other Commentary to § 2X1.1.

We observed in United States v. Waskom that there are several considerations in determining whether a reduction under § 2X1.1(b)(2) is appropriate. This Guidelines section has been amended since our decision in Waskom, and the conspirators' subjective belief as to what was "necessary on their part for the successful completion of the substantive offense" is no longer part of the equation. However, four of the considerations set forth in Waskom remain pertinent. They are (1) a focus "on the substantive offense and the defendant's conduct in relation to that specific offense," (2) "§ 2X1.1(b)(2) does not require the reduction for a conspirator who has made substantial progress in his criminal endeavor simply because a significant step remains before the commission of the substantive offense becomes inevitable," (3) "the circumstances must demonstrate that the balance of the significant acts completed and those remaining tips toward completion of the substantive offense," which "requires that the district court consider the quality of the completed and remaining acts, not simply the relative quantities of each," and (4) "a sentencing court should consider the temporal frame of the scheme and the amount of time the defendant would have needed to finish his plan, had he not been interrupted." In Waskom, we concluded that the district court had clearly erred in concluding that the defendants were "about to" obtain the materials they needed to carry out the robbery of an armored car or were "on the verge" of completing the acts they believed were necessary to carry out the offense.

In the present case, John was convicted in Count 1 of conspiracy to violate 18 U.S.C. § 1029(a)(5). The advisory Guidelines sentencing range was calculated based on this offense. With regard to this count, the indictment alleged and the evidence reflected that about six weeks after John had provided Riley with printed images of checks and printed images of the "best customer" account computer screen contents, Riley called Citigroup and asked that the mailing address on the account of Carolyn Baker Real Estate be changed. Approximately a month after that, he called and asked that the name of a confederate be added to the list of authorized users, and Citigroup complied. The following day, the confederate who had been added as an authorized user purchased gift cards having a value of $5,000 at Home Depot. With regard to the other three accounts that were actually accessed, the process was similar. Accordingly, in order to complete the scheme with regard to the accounts for which Riley had the detailed "best customer" account computer information, someone would have had to have requested a change of address, Citigroup would have had to have complied, and a confederate would have to have been added as an authorized user by Citigroup before the account could have been accessed. With regard to the accounts for which Riley had only copies of checks, it is unclear whether Citigroup customer accounts could have been accessed with just this information, and if so how.

These facts lead us to conclude that the three-level reduction under § 2X1.1(b)(2) applies. The district court found that John and her co-conspirators intended to obtain through their fraudulent scheme goods or cash equivalents worth $1,451,865. The scheme resulted in an actual loss amount of $78,750. The acts necessary to complete the intended offense had not been completed at the time John and others were apprehended. These circumstances are indistinguishable from the example in note 4 of the Commentary to § 2X1.1, which explains:

For example, where the intended offense was the theft of $800,000 but the participants completed (or were about to complete) only the acts necessary to steal $30,000, the offense level is the offense level for the theft of $800,000 minus 3 levels, or the offense level for the theft of $30,000, whichever is greater.

The district court clearly erred in failing to apply the three-level reduction. This clear error satisfies the first two prongs of the plain-error standard of review.

To satisfy the third prong of plain-error review, the error must have affected the defendant's substantial rights, which ordinarily requires the defendant to show that the error "affected the outcome of the district court proceedings." "When the rights acquired by the defendant relate to sentencing, the `outcome' he must show to have been affected is his sentence." A sentencing error affects a defendant's substantial rights if he "can show a reasonable probability that, but for the district court's misapplication of the Guidelines, [he] would have received a lesser sentence."

A three-level reduction in John's offense level would reduce the advisory sentencing Guidelines range from 97-121 months of imprisonment to 70-87 months of imprisonment. These two sentencing ranges do not overlap. The difference between a potential minimum sentence under these ranges is 27 months of imprisonment, and the difference between a potential maximum sentence under each is 34 months of imprisonment. John's sentence of 108 months of imprisonment exceeds the high end of the 70-87-month range by 21 months and exceeds the low end by 38 months. Our court has held in similar circumstances that such a disparity affects substantial rights. Furthermore, it is not apparent from the record that John would have received an above-Guidelines sentence of 108 months if the district court had applied the three-level reduction in the offense level. Accordingly, as there is a reasonable probability that, but for the district court's misapplication of the Guidelines, John would have received a lesser sentence, we conclude that the district court's error affected John's substantial rights.

As we have determined that the first three elements of plain error are present, we may exercise our discretion to correct the error if it "seriously affect[s] the fairness, integrity or public reputation of judicial proceedings." We are also cognizant that "[m]eeting all four prongs" of plain-error review "is difficult, `as it should be.'" Our court has concluded in prior decisions that when a district court's error clearly affects a defendant's sentence, that error seriously affects the fairness, integrity, or public reputation of judicial proceedings, particularly when the disparity between the Guidelines' range applied by the district court and the correctly calculated range is significant. We have not explained our reasoning in any detail. We do not attempt today to extrapolate a "test" or "factors" that should be applied in every application of the fourth prong of plain-error review when sentencing error is present because we are mindful of the Supreme Court's admonition in Puckett that "[t]he fourth prong is meant to be applied on a case-specific and fact-intensive basis. We have emphasized that a ` per se approach to plain-error review is flawed.' "

In the present case, our analysis of whether the sentencing error seriously affects the fairness, integrity, or public reputation of judicial proceedings is influenced by the fact that the sentence imposed is significantly (21 months) outside the advisory Guidelines range after applying the three-level deduction. We also bear in mind that before reaching the fourth prong of plain-error review, we determined with respect to the third prong that there is a reasonable probability that, but for the district court's misapplication of the Guidelines, John would have received a lesser sentence. Under these circumstances, the perception of fairness in sentencing is marred by at least two considerations.

The first is that a sentence has been imposed without the district court's understanding that the chosen sentence was, in reality, an above-Guidelines sentence and without the district court's consideration of the correct advisory Guidelines sentencing range. The sentence of 108 months of imprisonment was in the middle of the 97-121 months range that the district court erroneously applied. Absent remand, the defendant's sentence will be imposed without the district court's consideration of a lower Guidelines range, even though the Supreme Court has said that district courts should consider the properly calculated Guidelines range as "the starting point and the initial benchmark."

The second consideration is that if we were to affirm John's sentence, it would not be subjected to the process otherwise applicable to above-Guidelines sentences, which is that if a judge "decides that an outside-Guidelines sentence is warranted, he must consider the extent of the deviation and ensure that the justification is sufficiently compelling to support the degree of the variance." "[A] major departure should be supported by a more significant justification than a minor one. After settling on the appropriate sentence, he must adequately explain the chosen sentence to allow for meaningful appellate review and to promote the perception of fair sentencing." John's significantly above-Guidelines sentence has been imposed without the accompanying justification from the district court that the Supreme Court has recognized is necessary "to promote the perception of fair sentencing." If we vacate and remand, and the district court determines on remand that 108 months of imprisonment or another sentence above the advisory Guidelines range is an appropriate sentence, the reasons for such a sentence will be part of the public record and subject to review on appeal.

In the present case, we also take into account the fact that John's base offense level was substantially increased due to the intended-loss amount. As discussed earlier, John received a sixteen-level increase based on the intended-loss amount, whereas the actual-loss amount would have only subjected her to an eight-level increase. The integrity and public reputation of judicial proceedings requires that in this case, the district court consider the appropriate sentence, apprised of the applicability of the three-level reduction.

In conducting our plain-error review of John's sentence, we recognize that our sentencing precedent "has been generous with remand, often finding that errors leading to substantial increases in sentences, even those errors not raised until appeal and thus subject to plain error review, merited remand." However, as was the case in United States v. Ellis, "we are not convinced that the case law on this point is settled or as categorical as language in some cases might make it seem."

We note that other circuits have generally concluded that sentencing error that was likely to have been caused by selection of an incorrect sentencing range seriously affects the fairness, integrity, or public reputation of judicial proceedings when the sentence imposed is significantly above the correctly calculated Guidelines range. At least one court, the Seventh Circuit, has concluded that "even if a sentence imposed is within the correct as well as the incorrect Guideline range, the case must still be remanded for resentencing." Our court has not adopted this latter approach in every case in which there is an overlap between the sentencing range found applicable by the district court and the correctly calculated range. In United States v. Jasso, the district court found the applicable sentencing range to be 46 to 57 months and sentenced the defendant to 46 months of imprisonment. On appeal, this court concluded that the correct sentencing range was 41 to 51 months, but that the sentence of 46 months did not affect substantial rights. However, in United States v. Price, this court held that plain error had been demonstrated, even though the correctly calculated sentencing range overlapped with the range erroneously applied by the district court, because the sentence of 110 months of imprisonment was substantially (18 months) higher than the lowest end of the properly calculated Guidelines range, which was 92 to 115 months of imprisonment.

We conclude that ultimately, whether a sentencing error seriously affects the fairness, integrity, or public reputation of judicial proceedings is dependent upon the degree of the error and the particular facts of the case. In this case, the facts warrant remand.

The dissent disagrees with our assessment of John's sentence. While we share the dissent's concern that district courts should not be "sandbagged" when issues are raised for the first time on appeal, that cannot be the basis for declining to vacate John's sentence. In virtually every case in which plain error is found, a district court has been "sandbagged." The plain-error standard nevertheless remains a part of our jurisprudence, and we are bound to apply it.

We likewise part company with the dissent's suggestion that the nature of the underlying crime gives an appellate panel carte blanche to decline to vacate and remand due to sentencing error irrespective of the disparity between the sentence imposed and the applicable Guidelines range and irrespective of the probability that a lower sentence might be imposed if the district court were apprised of its error. The discretion inherent in the plain-error standard is not tantamount to caprice, nor is it to be exercised because of sympathy or lack thereof for a particular individual or the public's or a judge's opinion as to the seriousness or heinous nature of a particular crime. An appellate court's discretion is cabined by at least some guiding principles. We glean from our precedent and that of other circuits that when there is no indication that the district court would have selected the sentence regardless of the applicable Guidelines range, and the sentence imposed is based on an erroneously calculated Guidelines range, it is appropriate to exercise our discretion to vacate the sentence and remand the proceeding, at least when the sentence is materially or substantially above the properly calculated range. We do not attempt to define today the lower limits of what is "material" or "substantial," but we conclude that the disparity in the present case is both material and substantial, to the extent those terms might not be coextensive in this context, and that the fourth prong of plain-error review supports the exercise of discretion under these circumstances.

VIII

John contends that the district court erred in increasing her offense level under Guidelines § 2B1.1(b)(14)(A)(i), which provides for a two-level increase for a defendant convicted of an offense under 18 U.S.C. § 1030 that involved "an intent to obtain personal information." John asserts that the term "personal information" as used in § 2B1.1(b)(14)(A)(i) "does not include information held only by businesses rather than individuals, and pertaining exclusively to matters of commerce," and therefore that the enhancement does not apply because she only accessed credit information of corporate account holders.

The Guidelines commentary defines "personal information" as: sensitive or private information (including such information in the possession of a third party), including (i) medical records; (ii) wills; (iii) diaries; (iv) private correspondence, including e-mail; (v) financial records; (vi) photographs of a sensitive or private nature; or (vii) similar information.

John did not raise her present objection to the application of this enhancement with the district court. Our review would be for plain error.

However, we are vacating John's sentence, as noted above, on other grounds and are remanding for further proceedings. On remand, John is likely to assert an objection if this enhancement is again applied in calculating the advisory Guidelines sentencing range. We do not know what sentence the district court may impose or whether it will be a sentence within or without the advisory Guidelines. We accordingly express no opinion at this time as to whether the facts of this case permit application of an enhancement because the defendant intended to obtain "personal information."

* * *

We AFFIRM John's convictions. For the reasons considered above, we VACATE John's sentence and REMAND for further proceedings.