ROGERS, J., delivered the opinion of the court in which SENTELLE, J., joined, and MOORE, J., joined in part. MOORE, J. (pp. 736-37), delivered a separate opinion concurring in Parts I, II, III, and V, and in the judgment as to Part TV.
OPINION
ROGERS, Circuit Judge.In this bankruptcy case, Trustee Marcia Meoli seeks to recover allegedly fraudulent transfers of funds from the now-bankrupt company, Teleservices, to Huntington National Bank. Huntington lent money to, and maintained the deposits of, another company called Cyberco, which had created Teleservices to perpetuate a ‘ Ponzi scheme. As a part of the scheme, Cyberco and Teleservices shuttled the fraud’s proceeds between their bank accounts.
The Trustee sought to recover from Huntington three types of transfers from Teleservices: (i) direct loan repayments, which Teleservices sent directly to Huntington to pay down Cyberco’s debt to Huntington; (ii) indirect loan repayments, which Teleservices sent to Cyberco’s deposit account at Huntington, and which Cyberco later used to repay its debt to Huntington; and (iii) excess deposits, which Teleservices sent to Cyberco’s deposit account at Huntington, and which Cyberco later withdrew or the government later seized. The bankruptcy court held that the Trustee could recover all three *720types of transfers from Huntington. The district court agreed. Huntington appeals, arguing that it is not a transferee of the excess deposits, and that it received the loan repayments in good faith. The Trustee cross-appeals, challenging as unreasonably low the bankruptcy court’s calculation of prejudgment interest on the recoverable transfers.
Because Cyberco was free to withdraw its money from its deposit account, Cyber-co retained “dominion and control” over the deposits, despite Huntington’s security interest in them. Huntington gained dominion and control only over the money that it received in satisfaction of Cyberco’s debt to it; Huntington is a transferee of the direct and indirect loan repayments, but not of the excess deposits. Furthermore, because Huntington’s investigator discovered a critical clue to Cyberco’s fraud on April 30, 2004, and because that investigator failed to share that discovery with Huntington’s manager who oversaw the Cyberco account, Huntington has failed to prove that it continued to receive transfers from Teleservices in good faith after that date. But this court’s precedent does not necessarily require recoverability of transfers before that date, just because Huntington had earlier inquiry notice of fraud. Finally, the bankruptcy court did not err in calculating prejudgment interest using the statutory rate.
I.
Teleservices Group, Inc. and Cyberco Holdings, Inc. are two related bankrupt companies. Marcia Meoli is Teleservices’s bankruptcy trustee. Huntington National Bank is a bank headquartered in Columbus, Ohio. Barton Watson was the chairman and chief executive of Cyberco.
Barton Watson masterminded a Ponzi scheme. As the chairman and chief executive of Cyberco, Watson represented widely that Cyberco needed money to buy more computer equipment to support its growing computer-services business. Watson also represented widely that Cyberco purchased its equipment from a company called Teleservices. It was not widely known, though, that Teleservices was a paper company that Watson had created to perpetuate his fraud. Teleservices had no separate officers, directors, or employees, and it acted only through Cyberco’s executives, who assumed fake names for their fake Teleservices jobs. Watson and his accomplices would fabricate Teleser-vices invoices that purported to document the purchase of computer equipment. Under that pretense, Watson borrowed money from several equipment financing companies and instructed them to send the money directly to Teleservices to pay for the computer equipment. Once they paid Teleservices, Watson moved the money from Teleservices’s bank account to Cy-berco’s account at Huntington. Watson then used that money to pay his and others’ salaries and to pay Cyberco’s earlier debts to financing companies for previous “purchases.”
From September 2002 to October 2004, Huntington was Cyberco’s bank. A team of Huntington employees in Michigan managed the Cyberco account. Huntington maintained Cyberco’s deposit account and lent money to Cyberco. Originally, Huntington had lent about $9 million to Cyber-co, but by 2004 the loan would grow to $16 million.
In September 2003, Cyberco tried to deposit into its Huntington account a check for $2.3 million, but the check bounced. The check was from Teleservices. Gail White, a Huntington employee, was responsible for processing the bounced check. Smelling danger, she decided not to credit the check to Huntington’s account, despite knowing that without it, Cyberco’s *721checks would begin bouncing, too. White had discovered that, in the immediately preceding months, Cyberco had deposited a series of checks from Teleservices, all in large amounts. Neither White nor her superiors knew what Teleservices was.
In October 2003, several Huntington employees met with Watson. When asked about Teleservices, Watson explained that Teleservices was a recent addition to Cy-berco’s holdings, that Teleservices was not yet operational, and yet that Teleservices was already collecting Cyberco’s receivables before sending them to Cyberco.
Although the Huntington employees did not know it, Watson’s explanation contradicted his previous representation that Teleservices was Cyberco’s supplier of computer equipment. While White and her superiors were unaware, Huntington’s files actually recorded that previous representation. If they had been aware of that previous representation, they would easily have dismissed Watson’s explanation of Teleservices at the meeting. If indeed Tel-eservices were Cyberco’s supplier, then Cyberco should have been paying Teleser-vices for the purchased computer equipment — not the other way around — and Teleservices would not have suddenly transformed into a collector and aggregator of Cyberco’s receivables.
Though unaware of that smoking gun, White was still suspicious. Watson stated that Teleservices was not yet operational, but also that Teleservices had already begun transferring millions of dollars to Cy-berco, as White had seen. White suspected foul play also because Cyberco refused to use the lockbox that Huntington had set up. A lockbox is a service offered by a bank, where the bank receives and processes checks on behalf of a company. Through the lockbox, Huntington had hoped to monitor Cyberco’s revenues. But Cyberco did not use the lockbox, allegedly because its customers refused to use it, and thereby blinded White to Cyberco’s source of money. White continued to investigate Cyberco’s account with what information she had.
Other Huntington employees were apparently satisfied with Watson’s explanation — at least for the time being.
Nevertheless, in January 2004, Huntington asked Cyberco to find a new bank. Some Huntington employees explained that Huntington simply did not understand Cyberco’s complex business. But other Huntington employees clearly worried that Cyberco might not pay back its debt. John Kalb, one of White’s superiors, wrote at the time that “the ‘red flags’ continue,” that there was in “recent times ... the heightened risk of financial misinformation (as well as fraud),” and that he hoped Huntington would not “lose money” in the process of ending its relationship with Cy-berco.
By April 2004, Kalb was even more certain that there was foul play. Kalb believed that, in general, the computer services business was risky and that there was a tendency in the industry to exaggerate receivables. Kalb also knew that Cyberco had previously overspent its deposits and that Huntington had covered the excess spending by extending Cyberco’s loan to offset the overdraft. Finally, Kalb knew that Cyberco never gave audited financial statements to Huntington, even though the loan agreement required Cyberco to do so. Cyberco just kept promising that it would give the statements as soon as the auditors signed off on sales and mergers of Cyber-co’s subsidiaries.
Also around April 2004, White found proof of Cyberco’s foul play. White had accessed Cyberco’s receivables aging report, which is a standard accounting report that lists unpaid customer invoices. Cyber-*722co had claimed in that report that its customers included many companies in the Fortune 500 — companies that would never object to sending their checks to a lockbox. Furthermore, Cyberco listed companies like Electronic Data Systems as customers, even though such companies were also in the computer services business and were Cyberco’s competitors, not customers.
Kalb and White contacted Huntington’s security department. Larry Rodriguez, Huntington’s regional head of security, found out that the FBI was investigating Cyberco. Rodriguez also learned that Watson had a fraudulent past: Watson had been permanently blacklisted by the National Association of Securities Dealers; he had confessed to a bank fraud in Michigan and another fraud in California; and he had served three years in jail for a fraud-related crime. Rodriguez relayed that information to the FBI. But he did not share the information with Kalb or White. According to Kalb, Kalb asked Rodriguez about Rodriguez’s findings, and Rodriguez was reluctant to share them.
Meanwhile, Kalb suggested to Watson that Huntington contact Cyberco’s customers to verify that they were indeed Cyber-co’s customers. Watson fumed at the idea. As he explained, Cyberco’s customers would “smell a rat of some kind” if a bank called them for verification. As a compromise, Watson suggested that Grant Thornton, a renowned accounting firm, independently audit Cyberco. Huntington agreed. In early May 2004, Grant Thornton reported that Cyberco’s customers were real. Later, it would turn out that Watson had deceived Grant Thornton by providing it with fake responses from Cyberco’s fake customers.
In the following months, Cyberco paid its debt to Huntington. By July 2004, Cy-berco had paid down around half of its debt. Cyberco promised that more checks would arrive from Teleservices for the remaining debt. Although those checks from Teleservices did not arrive on time, causing Huntington to demand the full balance due, the checks did finally arrive in the following months. The last check cleared on October 29, 2004; that was also the largest check. Upon hearing the news, Kalb remarked: “All I can say is ‘whew’!!”
Later in 2004, the FBI raided Cyberco’s offices. Watson committed suicide shortly thereafter.
Creditors of Cyberco commenced an involuntary Chapter 7 proceeding against it. By then, a state court had already ordered a receiver to take control of both Cyberco and Teleservices. That receiver also filed for Teleservices’s bankruptcy.
In this suit, the trustee for Teleservices sought to recover from Huntington all of the direct loan repayments, the indirect loan repayments, and the excess deposits. The parties fully litigated the case in the bankruptcy court. The bankruptcy court conducted two trials and issued multiple opinions. On the issues on appeal alone, the bankruptcy court issued five opinions, some of which are published: (1) Bench Opinion (9/25/2009); (2) Bench Opinion (10/7/2009); (3) Bifurcated Issues Opinion (3/17/2011), Meoli v. The Huntington National Bank (In re Teleservices Group, Inc.), 444 B.R. 767 (Bankr. W.D. Mich. 2011); (4) Summary Judgment Opinion (3/30/2012), Meoli v. The Huntington National Bank (In re Teleservices Group, Inc.), 469 B.R. 713 (Bankr. W.D. Mich. 2012); and (5) Bench Opinion (7/23/2012).
The bankruptcy court concluded first that the Trustee could potentially recover $72 million in loan repayments and excess deposits from Huntington, as Huntington was those transfers’ “transferee” under the Bankruptcy Code. Reasoning that a *723bankruptcy trustee may recover certain transfers out of the bankrupt company only if the recipient of those transfers was the transfers’ “transferee,” the court applied the dominion-and-control test to determine whether Huntington was a transferee of the loan repayments and excess deposits. The court concluded that Huntington was a transferee of the direct and indirect loan repayments because, once received, the loan repayments were Huntington’s to spend however Huntington liked. The bankruptcy court also concluded that Huntington was a transferee of Cyberco’s excess deposits, even though Huntington never owned those deposits, because Huntington was free to use those deposits while they remained in Cyberco’s deposit account, subject only to Cyberco’s right to withdraw them.
The bankruptcy court then concluded that Huntington had received those transfers in good faith until April 80, 2004, but not after. The court noted that the Bankruptcy Code carves out an affirmative defense for transferees who received the transfers in good faith; they are immune from the bankruptcy trustees’ efforts to recover those transfers. To determine whether Huntington, after beginning its relationship with Cybereo as an innocent third-party bank, continued to receive transfers from Teleservices in good faith, the bankruptcy court analyzed whether Huntington “ever reach[ed] the point where it could no longer legitimately cling to its belief that the Teleservices transfers were [as Cybereo claimed they were] only Cyberco’s collected receivables.” Meoli v. The Huntington National Bank (In re Teleservices Group, Inc.), 444 B.R. 767, 818 (Bankr. W.D. Mich. 2011). The bankruptcy court concluded that Huntington reached that point on April 30, 2004, id. at 880, when Rodriguez, one of its investigators, discovered that Watson was a convicted fraudster, but failed to share that discovery with Kalb, who was overseeing Huntington’s banking relationship with Cybereo, id. at 785-86. By then, the bankruptcy court found, Kalb already had serious suspicions about Cyberco’s activities, id. at 779-80, and, the bankruptcy court also found, if Rodriguez had alerted Kalb to Watson’s fraudulent past, Kalb would have acted differently than to continue to receive the transfers from Teleservices in seeming good faith, id. at 829. That critical breakdown in communication, the bankruptcy court concluded, ended Teleser-vices’s good faith.
The bankruptcy court next concluded that Huntington’s affirmative defense ceased to apply earlier — before that breakdown of good faith on April 30, 2004 — because Huntington gained inquiry notice of Cyberco’s fraud on September 25, 2003, and because that inquiry notice sufficed to end Huntington’s affirmative defense under Sixth Circuit precedent. The bankruptcy court found that Huntington gained inquiry notice of Cyberco’s fraud on September 25, 2003, because that was when Watson explained to Huntington that Teleservices was Cyberco’s recently acquired subsidiary, and because at that time Huntington’s files documented Watson’s prior and contradictory statement to Huntington that Teleservices was Cyber-co’s supplier, not subsidiary. That contradiction, the bankruptcy court concluded, constituted inquiry notice of fraud. The bankruptcy court furthermore reasoned that it was bound by Sixth Circuit precedent to conclude that that inquiry notice destroyed Huntington’s affirmative defense, even as it concluded that Huntington’s good faith continued for months thereafter. In re Teleservices Group, Inc., 444 B.R. 767, 835-40.
Huntington’s liability amounted to $72 million in loan repayments and excess deposits. The bankruptcy court awarded pre*724judgment interest to the Trustee on a portion of that amount and under the federal statutory rate for postjudgment interest. In applying that statutory rate, the bankruptcy court noted that the rate was low, but nevertheless concluded that the rate reflected the sort of conservative investment that a bankruptcy trustee would have had to make.
Finally, citing concerns that it lacked the constitutional authority to enter final judgment, Meoli v. The Huntington National Bank (In re Teleservices Group, Inc.), 456 B.R. 318 (Bankr. W.D. Mich. 2011), the bankruptcy court summarized those findings of fact and conclusions of law in a Report and Recommendation.
The district court rejected all objections to the Report without much discussion.1
On appeal, Huntington challenges its transferee liability to Cyberco’s excess deposits, arguing that it never gained dominion and control over those funds, as Huntington was merely maintaining those funds, subject to Cyberco’s right to remove those funds at its will. Huntington also argues that the bankruptcy court applied the wrong standard to assess its good faith, pointing primarily to the fact that Huntington continued to cooperate beyond April 30, 2004, with the FBI’s investigation of Cyberco, and suggesting that its good faith therefore also continued beyond that date. Huntington also contends that Sixth Circuit precedent did not require the bankruptcy court to conclude that its affirmative defense ended earlier, when it gained inquiry notice of Cyberco’s fraud.
On cross-appeal, the Trustee challenges the bankruptcy court’s good-faith determination from the other side, arguing that the bankruptcy court applied the wrong test, and that had it applied the right test, it would have found that Huntington’s good faith ended far before April 30, 2004. The Trustee also challenges the bankruptcy court’s application of the federal post-judgment rate to calculate prejudgment interest, complaining that that rate is unreasonably low.
II.
The Trustee cannot recover Cy-berco’s excess deposits (those deposits not applied to pay back debts to Huntington) under the Bankruptcy Code provision for recovery of avoidable transfers from “transferees” because banks are not “transferees” with respect to ordinary bank deposits. The Trustee may only recover from Huntington the transfers as to which Huntington is a “transferee” under the Bankruptcy Code.2 For the following *725reasons, Huntington is not a “transferee” of the excess deposits.
Huntington is not a “transferee” of Cyberco’s excess deposits because Huntington did not gain “dominion and control” over them.3 After some initial confusion, Huntington, the Trustee, and the lower courts agree that this court applies the dominion-and-control test to determine transferee liability — for both initial transferees and subsequent transferees. “An initial transferee must have ‘dominion’ over the funds to be an ‘initial transferee’ under the statute.” Taunt v. Hurtado (In re Hurtado), 342 F.3d 528, 533 (6th Cir. 2003). We have repeatedly and approvingly quoted the Seventh Circuit’s original articulation of the test: “[T]he minimum requirement of status as a ‘transferee’ is dominion over the money or other asset, the right to put the money to one’s own purposes.” Id. (quoting Bonded Fin. Servs., Inc. v. European Am. Bank, 838 F.2d 890, 893 (7th Cir. 1988)); First Nat’l Bank of Barnesville v. Rafoth (In re Baker & Getty Fin. Servs., Inc.), 974 F.2d 712, 722 (6th Cir. 1992) (quoting the same). We have thus distinguished “mere possession” from “ownership,” so that “a party is not to be considered an initial transferee if it is merely an agent who has no legal authority to stop the principal from doing what he or she likes with the funds at issue.” In re Hurtado, 342 F.3d at 534.
First, Huntington did not gain dominion and control over Cyberco’s excess deposits by virtue of its being Cyberco’s depository bank. As our sister circuits have explained, the account-holder’s right to withdraw the deposits keeps the bank from obtaining dominion and control. In Bonded, the Seventh Circuit squarely held that a depository bank lacks dominion and control over customers’ deposits. In that case, as in this case, a bank provided two sendees to a customer: the bank lent money to the customer and maintained the customer’s deposit account. See Bonded Fin. Servs., Inc. v. European Am. Bank, 838 F.2d 890, 891 (7th Cir. 1988). When that bank received a check for $200,000 with instructions to deposit that amount into the customer’s account, and when the bank did so, the Seventh Circuit held that the bank “received nothing .... that it could call its own ... and [the customer] owed the [blank as much as ever.” Id. at 893-94. It was only when the customer instructed the bank to use that $200,000 to reduce its debt to the bank that the bank gained dominion over the money. Id. at 894.
Other circuits agree. Citing Bonded, the Eleventh Circuit has also squarely held the same: “When banks receive money for the sole purpose of depositing it into a customer’s account ..., the bank never has actual control of the funds and is not a section 550 initial transferee.” Nordberg v. Societe Generate (In re Chase & Sanborn Corp.), 848 F.2d 1196, 1200 (11th Cir. 1988). Also citing Bonded, the Ninth Circuit has noted the same: “[A bank] will not have dominion over [deposits] because it has no discretion over the uses to which the depositor’s money is to be put,” and therefore, “the bank is not the transferee, but the conduit or agent for a general deposit.” Universal Serv. Admin. Co. v. PosP-Confirmation Comm, of Unsecured Creditors of Incomnet (In re Incomnet, Inc.), 463 F.3d 1064, 1074 (9th Cir. 2006). Commentators have agreed. For instance, Collier on Bankruptcy § 5-550.02(4)(a) (Alan N. Resnick & *726Henry J. Sommer eds., 16th ed.) includes “a financial institution that received funds for deposit” among “[pjarties that have been held not to be transferees because they did not have sufficient dominion or control over the property transferred.”
The Eleventh Circuit has further explained that depository banks lack dominion and control over deposits even though the banks may use the deposits until the depositor withdraws them. Depository banks turn profits by employing customers’ deposits to earn money at a higher rate than it pays interest to the customers. In other words, depository banks have “legal rights to put deposited funds to use.” Menotte v. United States (In re Custom Contractors, LLC), 745 F.3d 1342, 1350 (11th Cir. 2014). But still depository banks lack dominion and control over deposits. In re Chase & Sanborn Corp., 848 F.2d at 1200. “[W]hen a bank receives funds in the form of a deposit, the attendant obligations owed to the transferor — namely to return the funds upon request — are sufficiently important that ... the bank [is not] liable as an initial transferee in spite of the significant control it exercises over the funds.” In re Custom Contractors, LLC, 745 F.3d at 1350; see also id. at 1350 n.6.
Our decision in Hurtado, if not squarely on point, supports that reasoning. In Hur-tado, we held that when someone deposits money in a bank account, that person exercises “dominion and control” over the deposits. Taunt v. Hurtado (In re Hurtado), 342 F.3d 528 (6th Cir. 2003). In that case, the debtors tried to hide their assets from their creditors by writing checks to Hurta-do’s mother: Barbara Hurtado. In re Hurtado, 342 F.3d at 530. Hurtado deposited the checks into her savings account at TNC Credit Union. Id. Hurtado then maintained those deposits at the debtors’ direction, withdrawing some funds for them from time to time. Id. Three sets of parties arguably exercised dominion and control over those deposits: (i) the debtors; (ii) Hurtado; and (iii) TNC Credit Union. We held that Hurtado exercised dominion and control over her deposits, explaining that “Hurtado was not under any legal obligation to follow the debtor’s directions,” that “Hurtado here was given legal title to the funds,” and that “Hurtado had legal authority to do what she liked with the funds.” Id. at 535. Thus we held that the depositor maintains dominion and control over her deposits.
That holding strongly implies that the bank lacks dominion and control over the same deposits. When a bankrupt entity transfers property to a third party, and that third party deposits the property into its deposit account, either the third party or the bank exercises dominion and control over the deposits, but not both. If both did, then both would be transferees of the deposits, and the trustee of the bankrupt estate would potentially recover two times the property that was transferred out of the bankruptcy.
It is not sufficient that Huntington could use the cash as it pleased as long as it had enough liquidity to pay back deposits on demand. As the Eleventh Circuit explained, the depository bank’s obligation to maintain liquidity is “sufficiently important” to defeat any dominion and control that the depository bank might otherwise have exercised over those funds. In re Custom Contractors, LLC, 745 F.3d at 1350. The Tenth Circuit has noted the same, stating that “[t]he bank was obligated to make the funds available to [the depositor] upon demand and, therefore, it acted only as a financial intermediary,” In re First Sec. Mortg. Co., 33 F.3d 42, 44 (10th Cir. 1994), and explaining that “the fundamental banking principle that money deposited in a bank account becomes property of the bank, offset by a debt to the *727account holder, ... does nothing to aid [trustee’s] argument regarding identity of the initial transferee,” id. at 43 n.5.
Huntington did not become a transferee of Cyberco’s deposits simply by maintaining Cyberco’s deposit account.
Second, the Trustee argues that even if a bank is not always a transferee of bank deposits, Cyberco’s was not a typical deposit account and that cases like Bonded therefore do not apply. The Trustee argues that Huntington’s security interest in Cyberco’s deposits, which Huntington perfected to secure its loan to Cyberco, granted dominion and control over those deposits to Huntington. The district court did not rely on this theory, but Huntington raises it as an alternative basis for affir-mance.4
That security-interest theory fails. Huntington’s security interest, perfected to secure the $16 million loan to Cyberco, could not have granted dominion over $64 million in deposits: a security interest cannot grant dominion over $48 million more than the underlying debt. The security interest did not even grant dominion and control over $16 million of the deposits, in light of plain statements in the texts of the loan agreements that Cyberco owned the deposits, despite Huntington’s security interest in them. Even though Huntington was the recipient of a transfer of a security interest, Huntington was not a “transferee” of that transfer under the Bankruptcy Code.
The Trustee’s main security-interest argument proves far too much; Huntington’s security interest did not yield dominion and control over $64 million, the entirety of Cyberco’s deposits. The law of secured transactions limits recovery upon default to the underlying debt.' Under Michigan law, when the secured party liquidates the collateral to satisfy the debt, then “the secured party shall account to and pay a debtor for any surplus.” Mich. Comp. Laws § 440.9615(4)(a). Similarly, the Uniform Commercial Code states that “[a] secured party shall account to and pay a debtor for any surplus.” U.C.C. § 9-608(a)(4). The Supreme Court has explained: “A mortgage can have no separate existence,” and “[w]hen the note is paid the mortgage expires.” Carpenter v. Longan, 83 U.S. 271, 275, 16 Wall. 271, 21 L.Ed. 313 (1875). The security interest ex*728ists only to ensure that the secured creditor gets its money back; once it does, nothing is left of the security interest. Huntington’s security interest in Cyberco’s deposits therefore did not result in dominion and control over funds in excess of the underlying debt of $16 million.5
The Trustee insists that Huntington gained dominion and control over all of Cyberco’s deposits because they were deposited in pieces. Under that theory, Huntington gained dominion and control over every deposit not exceeding the outstanding debt. Suppose Cyberco owed Huntington $10 million. Suppose Cyberco daily deposited $1 million for 80 days, and Huntington never applied the deposits to reduce Cyberco’s debt. Then, under the Trustee’s theory, Huntington gained dominion and control over $80 million because every $1 million deposit did not exceed the outstanding $10 million debt, and Huntington therefore had a right to take every $1 million deposit. But that theory ignores a critical difference between the tenth $1 million deposit and the eleventh $1 million deposit. Unlike Huntington’s “right” to take the tenth $1 million deposit, Huntington’s “right” to take the eleventh $1 million deposit was potential and conditional on Huntington’s not having taken the previous deposits to satisfy the debt. The cap on Huntington never lifted; Huntington’s “rights” were limited to the amount of the debt.
Furthermore, even the more limited version of the Trustee’s security-interest argument fails. Huntington’s security interest in Cyberco’s assets did not grant dominion and control over Cyberco’s deposits at all — not even over the $16 million in debt. The text of the security agreement between Huntington and Cyberco reveal that Cyberco retained dominion and control over its deposits, despite Huntington’s security interest in them. The Security Agreement states at the outset that, despite the security interest, Cyberco “owns” the collateral. Furthermore, the Security Agreement states that, despite the security interest, Cyberco may use and dispose of its collateral in the ordinary course of business. Under Michigan law, such “floating liens” are valid, even though they leave the debtor with “unfettered dominion or control over the collateral.” Mich. Comp. Laws § 440.9206 cmt. 2. In short, Cyberco retained dominion and control over its deposits, despite Huntington’s security interest in them, and Cyberco could use its money and other assets however it wanted to.
Finally, the Trustee erroneously argues that because the grant of a security interest is a “transfer” under the Bankruptcy Code, the Code therefore makes the recipient of that “transfer” a “transferee.” The Code defines a “transfer” very broadly to include the grant of a security interest. See 11 U.S.C. § 101(54).6 But “[t]he Bankruptcy Code ... generally sep*729arates the issue of whether a transfer is avoidable from the issue of the liability of a transferee.” 4 Norton Bankruptcy Law and Practice 3d § 70:2 (emphases added). Indeed, the cases that apply the dominion- and-control test to determine transferee liability hold that not all recipients of transfers are transferees. At least one case states so explicitly. “[The statutory text] is not so plain as to compel, or persuasively argue for, the principle that every conduit is an initial transferee. The statutory term is ‘transferee’ — not ‘recipient’ — and is not self-defining.” Christy v. Alexander & Alexander of N.Y. Inc. (In re Finley, Kumble, Wagner, Heine, Underberg, Manley, Myerson & Casey), 130 F.3d 52, 56 (2d Cir. 1997). Other courts necessarily hold so implicitly, because the dominion-and-control test narrows the set of transferees to exclude recipients of transfers of limited possessory interests.7
Huntington is therefore not a transferee of Cyberco’s excess deposits, even though it could use those funds subject only to Cyberco’s right to withdraw them, and even though it perfected a security interest in them.8
III.
Huntington concedes that it is a transferee of the loan repayments. Or. Arg. at 2:50. When Teleservices sent a check to Huntington with instructions to use the check to pay down Cyberco’s debt, and Huntington did so, Huntington came to own that money exclusively, and could use that money as it wished. Huntington is therefore an initial transferee of the direct loan repayments. When Cyberco instructed Huntington to use funds in its deposit account, which Teleservices had sent to Cyberco, to pay down Cyberco’s debt, and Huntington did so, Huntington came to own that money exclusively, too, and could use that money as it wished.9 Huntington is therefore a subsequent transferee of those indirect repayments.
Although Huntington is not a transferee of Cyberco’s excess deposits, it is a transferee of both direct and indirect loan repayments. Huntington concedes as much. But Huntington claims that it has an affirmative defense to the Trustee’s recovery of the loan repayments.
IV.
To its transferee liability for the loan repayments, Huntington raises affirmative defenses under §§ 548(c) and 550(b)(1). An initial transferee is not liable for the transferred property if the transferee: (i) took the property “in good faith”; and (ii) “gave value to the debtor in exchange for such transfer.” 11 U.S.C. § 548(c). Huntington *730is an initial transferee of the direct loan repayments from Teleservices, which Tel-eservices . gave directly to Huntington; Huntington argues that it took those transfers “for value” and “in good faith.” A subsequent transferee is not liable for the transferred property if the transferee took the property: (i) “for value,” (ii) “in good faith,” and (iii) “without knowledge of the voidability of the transfer avoided.” 11 U.S.C. § 550(b)(1). Huntington is a subsequent transferee of the indirect loan repayments, which Teleservices sent to Cy-berco, and which Cyberco later gave to Huntington to pay down its debt; Huntington argues that it received those transfers “for value,” “in good faith,” and “without knowledge of the[ir] voidability.”10
The bankruptcy court did not err in finding that a critical breakdown in Huntington’s internal communications ended its proven good faith on April 30, 2004. The Trustee may therefore recover all subsequent loan repayments, which include some of the indirect loan repayments and all the direct loan repayments. With respect to the earlier indirect loan repayments, the bankruptcy court erred in concluding that our precedents necessarily ended Huntington’s affirmative defense earlier — on September 25, 2003 — when Huntington gained inquiry notice of Cy-berco’s fraud. The Trustee may therefore recover indirect loan repayments made between the two dates — September 25, 2003, and April 30, 2004 — only if the court concludes on remand that Huntington’s proven lack of knowledge of the voidability of the transfers ceased during that period.
A.
Despite Huntington’s arguments, the bankruptcy court did not err in concluding that Huntington’s proven good faith ended on April 30, 2004. On that day, Huntington’s investigator discovered Watson’s fraudulent past, but failed to disclose that past to Huntington’s manager who oversaw the bank’s relationship with Cyberco. After that critical breakdown in communication, Huntington as a corporate entity could not prove that it continued to receive transfers from Teleservices in good faith.
After reviewing the conduct of Huntington’s employees, Meoli v. The Huntington National Bank (In re Teleservices Group, Inc.), 444 B.R. 767, 817-30 (Bankr. W.D. Mich. 2011), the bankruptcy court properly concluded that Huntington had not proven its good faith after April 30, 2004, id. at 830. On April 30, 2004, Larry Rodriguez, to whom Huntington had delegated the investigation into Cyberco, discovered Watson’s fraudulent past. Id. at 785. Watson had been permanently blacklisted by the National Association of Securities Dealers; he had confessed to a bank fraud in Michigan and another fraud in California; and he had served three years in jail for a fraud-related crime. Id. Rodriguez then disclosed that past to the FBI, but failed to share that discovery with John Kalb, to whom Huntington had delegated the responsibility of managing the bank’s relationship with Cyberco. Id. at 785-86. The bankruptcy court found that that failure was critical: “Had Rodriguez only done his duty — that is, had Rodriguez shared his discovery about Watson with Kalb as he had with the FBI [...] — Kalb would have recognized Watson for the liar he was and' reacted accordingly.” Id. at 829. Indeed, by then, Kalb had already discovered that Watson had lied about Cyberco’s customers, id. at 779, had already spotted many “red flags” in Cyberco’s financials, id. at 780, and had worried that Hunting*731ton may “lose money” because of Cyberco, id. Because “a corporation cannot feign ignorance ... when it has delegated responsibilities to a group of individuals and an inexcusable breakdown of communication then occurs within the group,” id. at 828, and because Rodriguez’s failure to share his discovery “cannot be excused,” id. the bankruptcy court properly concluded that Huntington’s proof of good faith ended on April 30, 2004, id. at 830.
While Huntington argues that the bankruptcy court committed legal error by failing to consider Huntington’s continuing cooperation with the FBI, that argument fails on its face. As explained above, the bankruptcy court found sufficient facts— that Rodriguez discovered, but failed to inform Kalb about, Watson’s fraudulent past, In re Teleservices Group, Inc., 444 B.R. 767, 785-86, and that, eounterfactually, had Rodriguez informed Kalb about Watson’s fraudulent past, Kalb would have done differently than continue Huntington’s relationship with Cyberco as-is, including receiving continued transfers from Teleservices, id. at 829 — to support its conclusion that Huntington’s proven good faith ended on April 30, 2004. While the bankruptcy court did not also find the negative fact — that Huntington’s continued cooperation with the FBI did not cure the corporate bad faith embedded in that breakdown in communication — the bankruptcy court was well aware of that cooperation, see id. at 785-86, and the bankruptcy court was under no obligation to specify a negative factfinding for all the possible reasons that could have, but did not, trump Huntington’s failure.
Huntington also complains that the bankruptcy court committed legal error by amalgamating Huntington’s employees’ scattered information. Relatedly, Huntington complains that the bankruptcy court found that Huntington’s proven good faith ended on April 30, 2004, without finding that any Huntington employee’s proven good faith ended on April 30, 2004. Neither argument works. Huntington’s good faith may end while its employees’ good faiths continued, if its employees failed to share information — innocently but critically— with the person whom Huntington charged with managing Huntington’s relationship with Cyberco. That is what the bankruptcy court concluded happened here. According to the bankruptcy court, it was Rodriguez’s failure to communicate with Kalb that ended Huntington’s proven good faith; and the bankruptcy court expressly declined to find either way whether Rodriguez’s failure was in good faith. In re Teleservices Group, Inc., 444 B.R. at 813, 829-30. It did not matter to the bankruptcy court whether Rodriguez’s failure was, as Huntington argues it was, an “innocent miscommunication.” Rodriguez failed to communicate his findings to Kalb, as the bankruptcy court found Rodriguez was supposed to, and that failure may cease Huntington’s good faith, regardless of whether it ceased Rodriguez’s good faith: “Huntington cannot hold up Kalb as the measure of its good faith yet not accept responsibility for Rodriguez withholding from Kalb information that would have truly put Kalb to the test.” In re Teleservices Group, Inc., 444 B.R. at 829.
Because Huntington’s proven good faith ceased on April 30, 2004, Huntington is liable for all Teleservices transfers after that date of which it is the transferee. The Trustee may recover all direct loan repayments, of which Huntington is an initial transferee, because Huntington received them all after April 30, 2004.11 The Trustee *732may also recover indirect loan repayments, of which Huntington is a subsequent transferee, that were made after April 30, 2004.
B.
With regard to indirect loan repayments between September 25, 2003, and April 30, 2004, the district court found good faith, but that recoverability was nonetheless required because Huntington gained inquiry notice of Cyberco’s fraud on the earlier date. The Trustee successfully argued below that that inquiry notice constituted “knowledge of the voidability of the transfers,” 11 U.S.C. § 550(b)(1), which defeated Huntington’s defense to its subsequent-transferee liability. That argument should not have succeeded without more, however, because inquiry notice, although sometimes enough to constitute “knowledge of the voidability of the transfer,” is not necessarily enough in every case. Taken together, Nordic Village and First Independence invite a holistic review of the facts to determine whether a reasonable person would have been alerted to a transfer’s voidability.
In Nordic Village, a divided panel of this court held that, under the facts of that case, “inquiry notice” sufficed for “knowledge of the voidability.” IRS v. Nordic Village, Inc. (In re Nordic Village, Inc.), 915 F.2d 1049, 1056 (6th Cir. 1990), rev’d on other grounds, United States v. Nordic Village Inc., 503 U.S. 30, 112 S.Ct. 1011, 117 L.Ed.2d 181 (1992). The transferee there was the IRS. It received a check, issued by the debtor company doing business as “Swiss Haus,” delivered by an individual who was an officer of that company, and with instructions to credit the check against the outstanding tax liabilities of the delivering individual, doing business as “Josefs Hair Design.” Nordic Village, 915 F.2d at 1050-51. Facing those facts, this court held that the IRS had “inquiry notice” of the voidability of that check, even though it was not apparent that the IRS had “actual notice.” Id. at 1056. This court explained that “[i]t is not in an ordinary business practice for corporate entities to pay one another’s taxes” and that that irregularity “is sufficient to place a reasonable person on notice that the transfer was illegitimate, and by extension, that it was voidable.” Id. That inquiry notice of voidability sufficed for “knowledge” of voidability. Id.
We distinguished Nordic Village in First Independence. In First Independence, the checks were issued by First Independence, the bankrupt debtor, to the Baumhafts, who owned First Independence, and the Baumhafts cashed the checks into their personal accounts at Merrill Lynch. In re First Independence, 181 Fed.Appx. 524, 526 (6th Cir. 2006). Merrill Lynch obliged. Id. In finding that Merrill Lynch accepted those checks without knowledge of their voidability, this court distinguished Nordic Village by pointing to two factual differences. First, there was “a range of legitimate scenarios” in which the Baumhafts could have cashed their company’s checks into their personal accounts — maybe the checks were salaries, loan repayments, or distributions to shareholders. Id. at 529. Second, any inquiry into the legitimacy of the checks would have been futile, because Merrill Lynch would have nobody else other than the Baumhafts, the owners of First Indepen*733dence, to ask about the legitimacy of First Independence’s checks. Id. at 530. We therefore concluded that, on those facts, “nothing ... would alert a reasonable person to the transfers’ recoverability.” Id.
In First Independence, we reasoned:
Certainly “some facts suggest ... other facts,” and “[i]f a transferee possesses knowledge of facts that suggest a transfer may be fraudulent, and further inquiry by the transferee would reveal facts sufficient to alert him that the property is recoverable, he cannot sit on his heels, thereby preventing a finding that he has knowledge.” But here the facts surrounding the Baumhafts’ deposits would not lead a reasonable person to believe that the transfers were voidable, even though the Baumhafts’ use of corporate checks may have placed Merrill Lynch on implied notice of the transfers’ illegitimacy under Michigan law.
In re First Independence, 181 Fed.Appx. at 529-30 (internal citations omitted).
Taken together, Nordic Village and First Independence invite a holistic factual determination of whether a reasonable person, given the available information, would have been alerted to a transfer’s voidability. This court’s holding in Nordic Village was that the transferee failed to prove lack of knowledge of voidability because the facts would have “place[d] a reasonable person on notice that the transfer was illegitimate, and by extension, that it was voidable.” In re Nordic Village, Inc., 915 F.2d at 1056. It was only in discussing what would suffice to put that reasonable person on notice that this court discussed inquiry notice. Id. While inquiry notice sometimes suffices to “alert” a reasonable person to voidability, In re First Independence, 181 Fed.Appx. at 529, on different facts, and viewing those facts holistically, a reasonable person may not be alerted to a transfer’s voidability even if there was inquiry notice, see id. at 529-30. What a reasonable person would be alerted to depends not just on whether there was inquiry notice, but also on what investigative avenues existed, whether a reasonable person would have undertaken those avenues given the situation, and what findings the reasonable investigations would have yielded. Taken together, Nordic Village and First Independence encourage these inquiries.
We do not read Nordic Village to foreclose rigidly those inquiries. The bankruptcy court concluded that it could not distinguish Nordic Village in the way that First Independence did. Unlike in First Independence, here the checks were sent from one company to another, and a full investigation by Huntington would have revealed that Teleservices was Cyberco’s alleged equipment provider, not Cyberco’s agent collecting Cyberco’s receivables before transferring them in lump sum to Cyberco. In re Teleservices Group, Inc., 444 B.R. 767, 839—40. But the controlling question remains: were the facts such that they would have “plaee[d] a reasonable person on notice that the transfer was illegitimate, and by extension, that it was voidable,” given the investigative avenues that existed, the reasonableness of pursuing those investigations, and the findings that those reasonable investigations would have yielded? See In re Nordic Village, Inc., 915 F.2d 1049, 1056 (6th Cir. 1990). That question remains unanswered.
The bankruptcy court determined that, under Nordic Village, Huntington failed to prove lack of knowledge of voidability of the transfers to Huntington starting on September 25, 2003, but it did so only with a heavy heart. The bankruptcy court would have found that Huntington proved that it had acted without the requisite knowledge up to April 30, 2004 — the same day until which, the bankruptcy court concluded, *734Huntington proved that it had acted in good faith. In re Teleservices Group, Inc., 444 B.R. 767, 835. The bankruptcy court cited legislative history to conclude, as other courts have, that the without-knowledge-of-voidability requirement is the same as the good-faith requirement. Id. at 830-35. Nevertheless, the bankruptcy court felt bound by this court’s opinions in Nordic Village to hold that inquiry notice constitutes knowledge, and that Huntington had inquiry notice beginning on September 25, 2003. Id. at 835-40. As explained, Nordic Village may permit but does hot demand that result.
The Trustee also argues that it can recover the transfers made between the two dates for a second reason: because the district court applied the wrong test to determine Huntington’s good faith. But the test that the lower court ultimately applied to determine “good faith” was proper. Before articulating the test, the bankruptcy court struggled to decide whether the proper test is objective or' subjective. The court discussed several precedents, which it deemed to have followed “an objective interpretation of good faith,” In re Teleservices Group, Inc., 444 B.R. at 795, then criticized that interpretation as “drawing] attention away from” the traditional markers of good faith, like “integrity, trust, and good conduct,” id. at 816 (quoting In re Bayou Group, 396 B.R. 810, 847 (Bank. S.D.N.Y. 2008)). The court then discussed several precedents and statutes, id. at 795-817, and concluded that the correct standard for good faith is a “subjective” test probing Huntington’s “integrity, trust, and good conduct,” id. at 816. Only then did the bankruptcy court articulate the controlling issue:
Consequently, the question of Huntington’s good faith boils down to simply this: Did Huntington ever reach the point where it could no longer legitimately cling to its belief that the Teles-ervices transfers were only Cyberco’s collected receivables? If the answer is no [i.e., throughout, Huntington legitimately held its belief that the Teleservices transfers were Cyberco’s collected receivables], then Huntington will have established its Section 550(b)(1) good faith. But, if the answer is yes [i.e., Huntington reached a point where it could not legitimately hold that belief], then it is difficult to comprehend how Huntington could have ever accepted any subsequent transfer from Teleservices without also suspecting with good reason that Teleservices was perpetrating a fraud upon its creditors.
Id. at 818.
That ultimate test is not erroneous. This court has observed that courts have “struggled” to define good faith in this context. In re First Independence, 181 Fed.Appx. 524, 528 (6th Cir. 2006). In one case, this court approved a bankruptcy court’s determination of good faith when the court determined that the transferee did not have actual notice of the voidability of the transfers and did not undertake “egregious, vindietive[,] or intentional misconduct.” Id. In another case, this court suggested in passing that when a transferee has actual notice of the voidability of the transfers, then it lacks good faith. In re Nordic Village, Inc., 915 F.2d 1049, 1056 (6th Cir. 1990) (“It is not apparent from the facts that the [transferee] had actual notice — that it did not accept the check in good faith — however,.... ”). Given that minimal guidance, the ultimate test that the bankruptcy court applied to Huntington to determine good faith — whether Huntington legitimately continued to believe that Teleservices’s transfers to Cy-berco’s account were merely Cyberco’s receivables that Teleservices had collected— is not erroneous.
*735In sum, the Trustee may recover from Huntington all direct loan repayments, of which Huntington is an initial transferee, because Huntington received them after its proven good faith ended on April 30, 2004. The Trustee may also recover from Huntington those indirect loan repayments, of which Huntington is a subsequent transferee, that Huntington received after April 30, 2004, and that Huntington received earlier, if the district court concludes on remand that Huntington gained knowledge of the voidability of the transfers before April 30, 2004.
V.
It was not an abuse of discretion for the bankruptcy court to award prejudgment interest according to the rate specified in 28 U.S.C. § 1961, instead of the market interest rate, because the bankruptcy court considered case-specific factors before applying that statutory rate.12 We have repeatedly approved of the statutory rate as within a district court’s discretion. In Ford v. Uniroyal Pension Plan, 154 F.3d 613, 619 (6th Cir. 1998), for instance, we repeated that “the statutory postjudgment framework set forth in 28 U.S.C. § 1961 is a reasonable method for calculating prejudgment interest awards.” We have also held that a “mechanical application” of that rate was an abuse of discretion, in a case addressing highly specific facts. Schumacher v. AK Steel Corp., 711 F.3d 675, 685-86 (6th Cir. 2013). We explained that “[a] proper determination involves a consideration of various case-specific factors and competing interests to achieve a just result.” Id. In that case, two classes of plaintiffs, who were substantially similarly situated, were both awarded prejudgment interest rates under § 1961, but because the treasury rate had changed drastically between when one class was awarded and when the other class was awarded, the two classes ended up with drastically different rates: one class received a 0.12% rate, while the other received a 4.7% rate. Id. at 681. No such idiosyncrasy exists in this case.13 To the *736contrary, as the bankruptcy court noted, the Trustee alone is suing Huntington to make sure that Teleservices’s creditors get their fair share of the bankrupt estate under the Bankruptcy Code. The Trustee is a fiduciary with the duty to invest conservatively. “[W]hile interest rates on treasury bills are disappointingly low at that time, these low rates nonetheless represent the safe type of investment that is expected of a fiduciary.” The bankruptcy court applied the statutory rate after considering case-specific factors and therefore did not abuse its discretion.
While the bankruptcy court did not specifically address the rate of inflation in relation to the treasury bill rate, it was not required to do so: When this court in Schumacher required the lower court to consider case-specific factors before applying the statutory rate, we also provided an illustrative list of case-specific factors, which included the rate of inflation. Schumacher, 711 F.3d at 687. But the case-specific factors that we listed as examples in Schumacher did not require that each example be addressed in every case. The bankruptcy court satisfied its duty to consider case-specific factors when it considered whether the statutory rate was fair in light of the type of conservative investment that a fiduciary like the Trustee would have pursued.
Our decision today, however, may prompt the bankruptcy court to exercise its discretion to apply a different rate. When the bankruptcy court chose the statutory rate to be the prejudgment interest rate, it did so in the context of a judgment that included about $55 million in Cyber-co’s excess deposits. These are funds that Huntington no longer has: Cyberco withdrew them or the government seized them. But we hold that Huntington is not liable for those excess deposits. Huntington remains liable for the loan repayments only, which are funds that Huntington received, as Cyberco’s creditor, in satisfaction of Cyberco’s debt to it. Once Huntington received that money, Huntington was free to invest that money however it wished. Doing so, Huntington may have profited more during this litigation than it will be ordered to pay under the statutory rate. The district court may therefore exercise its discretion to choose a different prejudgment interest rate, should it deem appropriate.
VI.
The judgment of the district court is reversed in part. The case is remanded for further proceedings consistent with this opinion.
. We review the judgment of the district court. The district court below concluded that it "f[ou]nd[] the reasoning in the Report and Recommendation sound and the proposed findings of fact and conclusions of law well supported.” Shorthand references in this opinion to the reasoning of the bankruptcy court are meant technically to refer more specifically to that reasoning as adopted by the district court.
. The Bankruptcy Code generally limits a bankruptcy trustee’s recovery of transfers to the avoidable transfers, see 11 U.S.C. § 548(a), that are received by a "transferee,” 11 U.S.C. § 550(a). The Code states: "The trustee may avoid any transfer ... of an interest of the debtor in property ..., if the debtor voluntarily or involuntarily ... made such transfer ... with actual intent to hinder, delay, or defraud any entity to which the debtor was ... indebted....” 11 U.S.C. § 548(a)(1) (emphasis added). Furthermore, if a trustee may avoid a transfer under that provision, then "the trustee may recover, for the benefit of the estate, the property transferred, or, if the court so orders, the value of such property, from — the initial transferee of such transfer or the entity for whose benefit such transfer was made; or any immediate or mediate [i.e., subsequent] transferee of such initial transferee,” 11 U.S.C. § 550(a) (emphases added). Here, no party asserts that Huntington was an entity for whose benefit the excess deposits *725were made. Thus liability requires in the context of this case that Huntington be a transferee.
. Nor did Huntington gain dominion and control over the indirect loan repayments before Cyberco instructed Huntington to apply those deposits to pay down the debt.
. Despite the Trustee’s claims, Huntington has not waived its right to refute the security-interest theory of its transferee liability. The Trustee argues that Huntington did waive the right because the bankruptcy court addressed the security-interest theory in an earlier bench opinion, and because Huntington did not challenge the theory in its opening appellate brief. But in that early bench opinion from October 2009, the bankruptcy court merely “forewarned” the parties that the security-interest theory of transferee liability "will need to be addressed,” and described itself as making "only an observation as opposed to a ruling.” Later, in March 2012, the bankruptcy court decided not to address the security-interest theory. Noting that Huntington had briefed the issue, the bankruptcy court plainly stated that it based Huntington’s transferee liability "simply ... upon its relationship as its depository bank” and added that it therefore "has not addres.sed Huntington’s ... arguments regarding [transferee liability] that might have arisen because of the attachment of its lien.” Meoli v. The Huntington National Bank (In re Teleservices Group, Inc.), 469 B.R. 713, 736 n.66 (Bankr. W.D. Mich. 2012). When the district court affirmed the bankruptcy court’s conclusions of law, the district court framed the issue of transferee liability as “whether a bank’s ability to use the funds deposited in a customer’s account constitutes sufficient control.” Under this posture, the Trustee has briefed the security-interest theory as an alternative ground for affirming Huntington’s transferee liability. Because “there is no waiver” when a party ”respond[s] to the alternative basis for affir-mance raised on appeal,” Huntington has not waived challenge to the security-interest theory. Innovation Ventures, LLC v. N.V.E., Inc., 694 F.3d 723, 729 (6th Cir. 2012).
. Huntington's security interest indeed attached not only to Cyberco's deposits, but also to all of Cyberco’s assets. If Huntington's security interest in Cyberco’s assets gave Huntington dominion and control over them, then Huntington was a transferee of not just Cy-berco’s deposits, but also everything that Cy-berco owned. The Trustee's main security-interest argument therefore suggests the absurd result that Huntington gained dominion and control over Cyberco's printers, computers, and phones.
. The Code defines a "transfer” very broadly. Under the Code, a transfer is "the creation of a lien,” 11 U.S.C. § 101(54)(A), "the retention of title as a security interest,” 11 U.S.C. § 101(54)(B), "the foreclosures of a debtor’s equity of redemption,” 11 U.S.C. § 101(54)(C), or "each mode, direct or indirect, absolute or conditional, voluntary or involuntary, of disposing of or parting with — (i) property[] or (ii) an interest in property,” 11 U.S.C. § 101(54)(D). Any transfer of an interest in property is a "transfer” under the Code.
. For example, we have recognized that the avoidability of a transfer and the recoverability of those avoidable transfers from a transferee are "two issues [that] must be kept analytically separate.” In re Hurtado, 342 F.3d 528, 532 (6th Cir. 2003). "An initial transferee must have 'dominion' over the funds to be an 'initial transferee’ under the statute.’ ” Id. at 533.
. We therefore need not address Huntington's alternative arguments against its transferee liability for Cyberco’s excess deposits: that those transfers of funds from Teleservices to Cyberco did not diminish Teleservices's bankrupt estate, and that without that diminution, there is no fraudulent conveyance to be recovered from Huntington.
.The analysis of these indirect loan repayments would mirror the central analysis in Bonded. See Bonded Fin. Servs., Inc. v. European Am. Bank, 838 F.2d 890, 894 (7th Cir. 1988) ("The Bank had no dominion over the $200,000 [in deposits] until January 31, when Ryan instructed the Bank to debit the account to reduce the loan; in the interim, so far as the bank was concerned, Ryan was free to invest the whole $200,000 in lottery tickets or uranium stocks.”).
. Because Huntington is not a transferee of Cyberco’s excess deposits, Huntington need not raise an affirmative defense to those transfers.
. The Trustee also argued that Huntington failéd to demonstrate an affirmative defense to its initial-transferee liability for a second reason: Huntington did not "g[i]ve value to *732[Teleservices] in exchange for [the direct loan repayments],” as required for an affirmative defense to initial-transferee liability under 11 U.S.C. § 548(c). We need not address that argument because Huntington’s affirmative defense to its initial-transferee liability failed for the first reason: Huntington did not prove that it received the direct loan repayments "in good faith.” 11 U.S.C. § 548(c).
. The Trustee raises two challenges to the bankruptcy court’s award of prejudgment interest, but we address only one challenge here. We need not address the other challenge that the bankruptcy court erroneously limited the award of prejudgment interest to the transfers made after April 30, 2004, even though the court also held Huntington liable as a subsequent transferee to the transfers between September 25, 2003, and April 30, 2004. Because we hold that Nordic Village does not necessarily defeat Huntington’s affirmative defense starting September 25, 2003, the challenged discrepancy between the date from which Huntington is held liable and the date from which prejudgment interest is accrued may disappear.
. In particular, we explained:
Here, if the 0.12% rate were to stand, AK Steel and the Plan would essentially be rewarded for their wrongdoing. There are at least three examples that highlight the potential injustice that could result if the district court’s award were affirmed. First, AK Steel’s rate on return (6.55%) and its borrowing costs (7.75%) were much higher than the 0.12% it has been ordered to pay. Allowing such a disparity would result in an unfair economic benefit to Appellants. See Rybarczyk [v. TRW, Inc.], 235 F.3d [975] at 986 [ (6th Cir.2000) ] ("[A] requirement that [defendants] pay the actual rate [on return] merely deprives [defendants] of [their] profit on the wrongfully denied benefits.”). Second, the West and Schumacher plaintiffs are identical plaintiffs. Both sets of plaintiffs worked for the same company, brought the same whipsaw calculation claims, during the same period, and were awarded damages under the same calculation. Nevertheless, the West class members received prejudgment interest of 4.7%, while the Schu-macher plaintiffs received only 0.12%. Such an absurd result amounts to a reversible error. See Ford, 154 F.3d at 618-19. Third, the Consumer Price Index measured the annual inflation rate at approximately 2.75%. In City of Warren, this Court held that a district court's award of pre-judg*736ment interest that only compensated the plaintiffs for the rate of inflation constituted an abuse of discretion where it failed to adequately compensate them for the lost use of their money. [U.S. v.] City of Warren, 138 F.3d [1083] at 1096 [(6th Cir.1998)].
Schumacher, 711 F.3d at 686-87.