In its recently released decision in Teva Canada Ltd. v. TD Canada Trust, the Supreme Court of Canada missed a golden opportunity to update controversial and antiquated interpretations of section 20(5) of the Bills of Exchange Act. In so doing, the Court refused to allocate liability for a cheque fraud scheme on to the company who failed to supervise a fraudster employee. Instead, in a 5-4 split decision, the majority decided to allocate liability to the bank who processed the cheques, basing their decision upon who "benefits" most from the bills of exchange system, rather than who is in the best position to prevent the fraud.
While the Court framed its decision as upholding precedent, for banks Teva was a missed opportunity to establish a more objective and predictable test. While it is never easy to determine liability between two innocent parties, the Court's decision sets out a system where banks, despite having limited information to protect themselves, bear a disproportionate amount of the risk when negotiating cheques.
TD Bank and the Bank of Nova Scotia (the "Collecting Banks") and Teva Canada Limited, a genetic pharmaceutical manufacturer, were innocent victims of a fraud totalling over $5M.
An employee in Teva's financing department carried out the fraud, requisitioning 63 cheques made payable to six different entities; four actual customers of Teva and two entities that he made up. The cheques were issued by Teva's accounts payable department and given to the employee, who had no authority to requisition cheques or approve payments to customers.
No signing officer or directing mind of Teva turned their mind to the cheques, and nobody examined them. The cheques were mechanically processed with no internal approval. None of the cheques in the fraud scheme were for any money owed by Teva. Accounts were opened up with the Collecting Banks in the name of the six entities, the cheques were negotiated, and funds were deposited into those accounts. In banking, the funds to cover those cheques are actually transferred to the Collecting Banks from Teva's account at a separate bank (the "drawer bank"). The scheme continued for three years until the fraud was discovered. The Collecting Bank would have had no reason to question the validity of the cheques, though Teva's internal accounting system should have flagged the transactions.
Teva sued the Collecting Banks, arguing the banks were liable for the amount Teva was defrauded because they had converted Teva's property without its consent. Teva could not sue its own bank, the drawer bank from where the money had been withdrawn, as that bank would have been protected by its customer contract. These banking customer agreements typically require the customer to review its accounts within 30 days of receipt and alert the bank of any discrepancies. Teva obviously failed to do so, as the fraud continued for three years. The Collecting Banks, however, did not have the benefit of any such protective clause as they had no contractual relationship with Teva.
The last time the Supreme Court of Canada considered this specific issue, over twenty years ago in Boma Manufacturing Ltd v Canadian Imperial Bank of Commerce ("Boma"), the Court described conversion as involving "a wrongful interference with the goods of another, such as taking, using or destroying these goods in a manner inconsistent with the owner's right of possession". Conversion is a strict liability tort. Therefore, if it can be shown that money was paid to a party that was not entitled to receive those funds, it has been converted. Because it is strict liability, the banks cannot plead negligence by Teva to limit their liability. The only shield they can rely on is s. 20(5) of the BEA, which gives a statutory defence to the banks if the cheque was payable to a "fictitious or non-existing" entity.
The interpretation of this statutory defence has resulted in the use of some awkward legal gymnastics. The Court in Boma adopted the cumbersome "Falconbridge" principles that consider the intention of the writer of the cheque to determine if the payee would or could have been real, even if in reality the payee was not. This has spawned a lengthy line of cases, and placed the onus on banks to disprove that a payee was an intended real person or recipient, even if the name of the payee was invented by a fraudster.
Both sides brought motions for summary judgment, and the motion judge awarded Teva the $5.4M it lost in the fraud. The banks appealed the decision to the Ontario Court of Appeal. The Court of Appeal reversed the motion judge's decision, finding that Teva was in the best position to prevent the fraud, and that the cheques could validly be categorized as fictitious and non-existing. The banks were thereby shielded from being liable for the defrauded sum. Our detailed analysis of the Court of Appeal's reasons can be found here.
The Supreme Court of Canada
Teva appealed to the Supreme Court of Canada, where the central issue was whether the Collecting Banks were entitled to the protection provided by s. 20(5) of the BEA. The bank would be protected if it were found that the fraudulent cheques were made out to a "fictitious" or "non-existing" entity and therefore could be treated as payable to bearer. If the instrument is payable to bearer, like a cheque made out to cash, the bank only needs to take delivery to become the lawful holder. Irrespective of whether a cheque was obtained by fraud, a bank would not be held liable for conversion.
A split bench of the Supreme Court of Canada restored the motion judge's decision. The reasons of the majority and the dissent reflect the strong disagreement over how s. 20(5) of the BEA should be applied. While the majority upheld the previous framework recognized in Boma, this provides little clarity to how banks can anticipate and mitigate risks when it comes to negotiating fraudulent cheques.
In a strongly worded dissent, four justices advocated deviating from precedent and applying a purely objective test for assigning loss in cases of cheque fraud. The dissent noted that the current test is problematic, and a simplified purely objective test would be easier to implement. The dissent suggested a two-step inquiry where a payee would be non-existing if they factually do not exist. If the payee does exist, a payee would be fictitious if there was no underlying transaction or debt to justify the transaction- easy to ascertain on evidence.
The main modification the dissent sets out was to the second-stage, and it arguably found the right balance. By allowing banks to rely on s. 20(5) where there was no underlying valid transaction, this analysis sidesteps the difficult inquiry of establishing "intention", particularly in larger corporate contexts. It would protect banks from liability in cases where a drawer failed to adequately monitor its accounting, as was the case with Teva. Requiring a subjective analysis of intent creates more uncertainty, and makes the application of s. 20(5) unpredictable.
The fraud in Teva had gone on for three years, with funds fraudulently transferred over sixty times. There was a systemic failure by Teva to catch millions of dollars in misappropriated funds. Due to the nature of this particular fraud, the Collecting Banks lacked the information to ascertain if the cheques were appropriately obtained. If potential fraud is not flagged by the drawer early on, there is no reason for banks to scrutinize cheques that are regularly being deposited into the fraudster's accounts.
The dissent, in addition to noting that the current test is less predictable, also acknowledged that the emphasis should be on incentivizing "upstream" controls by the drawer, who is best positioned to detect fraud. Importantly, both business and banks benefit greatly from the bills of exchange system. Banks should not be vulnerable to liability because of information it cannot be expected to be privy to. The risk should follow the party who is best able to prevent it.
The majority disagreed. They reviewed prior decisions by the Court and upheld a two-step framework that considers subjectively if a payee is fictitious, and objectively whether the payee is non-existing. The first step is a subjective inquiry to whether the drawer intended to pay the payee. Regardless of whether the payee is real or not, this test considers whether there was an intention to satisfy some debt or otherwise transfer funds. The second step is an objective inquiry into whether the payee was a legitimate payee of the drawer or one who could reasonably be mistaken for a legitimate payee. If either step is satisfied, the bank cannot rely on the s. 20(5) defence.
A payee is fictitious when the drawer did not intend to pay the payee. Intent to pay is presumed, and a specific intention does not need to be given for any particular cheque. The Court noted that modern commercial realities are such that cheques are often produced en masse and mechanically, and the guiding minds of the corporation cannot be expected to direct their attention to every cheque that is issued. The Court placed the burden on the bank to rebut the presumption by showing there was no intention to pay the payee on behalf of the drawer. If that can be established, then s. 20(5) operates to make the cheques payable to bearer, relieving the bank from liability.
As for a non-existing payee, the bank must establish that the payee lacks an established relationship with the drawer. Significantly in this case, if the name of the payee is close enough to an entity that does have an established relationship such that a person could reasonably mistake the two, the payee is not considered non-existing for the purpose of s. 20(5).
Neither Teva nor the banks were complicit in the fraud. They were both innocent parties left dealing with the loss. But the Court's affirmation of the two-step test for s. 20(5) allocates more of the risk on banks than had previously been made clear. The majority acknowledges as much, but also notes that the current system has been in place for over forty years without serious complaint. More importantly, the majority held that banks significantly benefit from the bills of exchange system and should have to bear some of the risks and losses associated with the system.
The majority decision expressly allocates the risk of loss from fraud based on who "benefits" most from the system rather than who is in the best position to prevent the loss in the first place. This may be an unfortunate choice, as it does not incentivize businesses to be vigilant to prevent fraud. One can only hope that, without waiting another twenty years, a further opportunity will arise for the Court to revisit this issue where hopefully the wisdom of the dissenting judges in Teva will then prevail.
 Boma Manufacturing Ltd v Canadian Imperial Bank of Commerce,  3 SCR 727 at para 31.