Court File and Parties

Court of Appeal for Ontario

Date: 20220508 Docket: M53421 (C69439)

Before: Lauwers J.A. (Motion Judge)

Between:

SS&C Technologies Canada Corp. Respondent (Responding Party)

and

The Bank of New York Mellon Corporation and CIBC Mellon Global Securities Services Company Appellants (Moving Parties)

Counsel: J. Thomas Curry, Brian Kolenda, Christopher Yung and Brandon Kain, for the moving parties Chris Paliare, Ren Bucolz, Glynnis Hawe and Catherine Fan, for the responding party

Heard: May 6, 2022 by video conference

Reasons for Decision

[1] The appellants, The Bank of New York Mellon Corporation (“BNY”) and CIBC Mellon Global Securities Services Company (“CIBC Mellon”), seek an order staying the judgment of Koehnen J., dated April 14, 2021, and amended on March 15, 2022, pending appeal. The appellants submit that the amendments the application judge made to his reasons for judgment give rise to a reasonable apprehension of bias on his part.

[2] For the reasons that follow, the motion is dismissed.

I. The Context

[3] SS&C Technologies Canada Corp. sold data to BNY under a data services agreement dated September 1, 1999. CIBC Mellon (a Canadian joint venture between Mellon Financial Corporation and CIBC) had its own data services agreement with SS&C, dated April 1, 1999, which it terminated on April 1, 2011. At that point CIBC Mellon began taking SS&C’s data from BNY. In 2017 SS&C brought this application claiming damages from BNY, on the basis that BNY was sharing data in breach of its data sharing agreement with SS&C, and that CIBC Mellon was in knowing receipt of the data. BNY defended, asserting that it was entitled to share the data it purchased from SS&C with all BNY entities, including CIBC Mellon.

[4] The application judge found BNY’s data sharing to be in breach of its contract with SS&C. He directed a trial on damages because he found that he needed the assistance of witnesses to understand the paper record.

II. The Procedural History

[5] BNY has unsuccessfully sought on several occasions to have the liability appeal heard by this court before the damages trial, as the chronology of proceedings leading to this motion shows:

• April 14, 2021: The application judge finds BNY liable and directs a trial of the damages issue (2021 ONSC 2657).

• The damages trial is scheduled for May 18-27, 2022 and June 21-24, 2022, on consent.

• May 14, 2021: BNY serves a Notice of Appeal in respect of the liability decision.

• June 30, 2021: The application judge refuses BNY’s request to adjourn or stay the damages trial until after the liability appeal (2021 ONSC 4682).

• September 3, 2021: Thorburn J.A. refuses to grant BNY’s motion in this court for an order expediting the appeal (2021 ONCA 601).

• September 10, 2021: BNY requests a panel review of Thorburn J.A.’s decision.

• December 22, 2021: The panel (Simmons, Pepall and Roberts JJ.A.) refuses to interfere with Thorburn J.A.’s discretionary decision not to expedite the appeal (2021 ONCA 913). The appeal is perfected the same day.

• January 21, 2022: BNY asks for argument of the appeal to be scheduled on May 16 or May 18, 2022, despite the scheduling of the damages trial.

• February 1, 2022: In my capacity as list judge, I refuse the appellants’ request to schedule the appeal for argument on May 16 or 18. The appeal on both liability and damages is set for September 22, 2022.

• March 15, 2022: The application judge issues an endorsement amending his liability reasons.

• April 21, 2022: The application judge dismisses the appellants’ recusal motion.

III. The Decision(s) under appeal

(1) The Liability Decision

[6] The application judge made two key findings in his original reasons for judgment: first, that BNY had breached the data services agreement with SS&C by sharing the data covered by the agreement with unauthorized users in BNY’s corporate group; and, second, that CIBC Mellon breached its agreement with SS&C.

[7] In his original reasons, at para. 19, the application judge made a finding that the “authorized users” under the BNY data sharing agreement with SS&C consisted of all “custodial entities of Mellon Financial Corporation as of 1999”. But the application judge also found, at para. 13, that “CIBC Mellon is a Canadian joint venture formed in 1996 by Mellon Financial Corporation and CIBC”. The appellants argue that this finding qualified CIBC Mellon as an entity with which BNY could share data, so that the application judge erred in finding that SS&C could recover damages from BNY Mellon for sharing data with CIBC Mellon.

[8] The pertinent ground of appeal of the liability decision focused on the breach of contract finding against CIBC Mellon that was not pleaded or argued by the respondent. The appellants dispute adamantly the application judge’s comment in the recusal reasons that: “In response to the request for directions, the Court corrected what both parties agreed was an error…” This error, which was the basis of their appeal, was a matter, say the appellants, for this court, not the application judge.

(2) SS&C’s Request for Amendments to the Reasons

[9] At a February 11, 2022 case conference, SS&C asked the application judge to make two changes to his liability reasons, on the basis that they were technical errors: first, to amend para. 19 of his reasons so that “authorized users” under the data services agreement would include “none of” the custodial entities of Mellon Financial Corporation as they existed in 1999; and second, to substitute for the finding that CIBC Mellon breached its data services agreement, which SS&C had not pleaded, the actual pleaded claim of knowing receipt on CIBC Mellon’s part.

[10] In response, the application judge made several amendments to his original reasons. At para. 3 of his March 15, 2022 amendment reasons, the application judge stated:

I will vary my reasons to remove the finding that CIBC Mellon is liable for breach of contract and will add a statement to the effect that CIBC Mellon was not entitled to receive Market Data provided by SS & C. I decline, however, to make any finding to the effect that CIBC Mellon is liable for knowing receipt because that relief was not sought before me on the initial application.

[11] As to the first request, para. 19 of the liability reasons originally provided:

BNY submits that courts should interpret agreements by looking to the words of the contract as informed by, but not overwhelmed by, the factual matrix in which the contract was made. If that analysis leads a court to believe that the contract is ambiguous, the court may look at the course of conduct of the parties for guidance. I adopt that approach below and look first to the terms of the agreement, then to the factual matrix and then to the conduct of the parties. In my view, each of those three elements demonstrates that the agreements were intended to benefit only the entity named in it and did not allow data sharing beyond the entity named in the agreement. In the case of the Mellon Trust agreement, that meant the custodial entities of Mellon Financial Corporation as they existed in 1999.

[12] SS&C’s specific request was that the last line of para. 19 of the liability reasons be changed to state: “In the case of the Mellon Trust agreement, that meant none of the custodial entities of Mellon Financial Corporation as they existed in 1999” (the underlined portion being the addition). This would have had the effect of excluding CIBC Mellon as a permitted data sharer.

[13] The application judge refused to make that change. Instead, he amended para. 19 in the following manner:

In the case of the Mellon Trust agreement, that meant the custodial entities of Mellon Financial Corporation as they existed in 1999 with the exception of CIBC Mellon and CIBC Mellon Trust Company (the underlined portion being the addition).

[14] The application judge also deleted the breach of contract finding against CIBC Mellon from para. 7 of the liability reasons. Paragraph 7 originally provided:

I have concluded that the respondents did breach their agreements. The language of the agreements, the factual matrix and the conduct of the parties are all consistent with an understanding that the data was to be used by the entity named in the contract.

[15] The application judge amended para. 7 to add the finding that “CIBC Mellon Global was not entitled to receive Market Data provided by SS&C under the Mellon Trust Agreement”.

[16] The application judge varied his liability reasons by making an additional order at paras. 33 and 34 of the amendment reasons:

The relief requested in paragraph 81 of SS & C’s [original factum in the application] specifically asks for an order to the effect that BNY breached the Mellon Trust agreement and seeks damages of $881,752,087 from BNY for that breach. The only relief the factum seeks against CIBC Mellon is found in paragraph 81 (b) which asks for an order: that CIBC Mellon Global was not entitled to receive the Market Data provided by SS & C under the Mellon Trust Agreement;

I grant that order.

(3) The Recusal Motion

[17] In light of the application judge’s amendments to the liability reasons, the appellants brought a motion asking the application judge to recuse himself and to declare a mistrial based on a reasonable apprehension of bias. He heard the recusal motion on April 12, 2022, and dismissed it on April 21, 2022. He found that the appellants had not established a reasonable apprehension of bias.

[18] On April 29, 2022, the appellants brought this motion to: (a) stay the liability judgment pending the appeal; and (b) obtain leave to file a Supplementary Notice of Appeal.

[19] The appellants served a Supplementary Notice of Appeal: (a) to remove the appeal ground that the breach of contract finding against CIBC Mellon was neither pleaded nor argued by SS&C; and (b) to add new grounds of appeal, including that the entire judgment and all other orders made by the application judge in the proceeding should be set aside due to the reasonable apprehension of bias created by his cumulative conduct.

IV. The Governing Principles

[20] The governing principles in three areas of law intersect in this motion: the first is on a judge’s authority to amend reasons after they are released; the second is on staying a judgment pending appeal; and the third is on judicial bias. I address each in turn.

(1) A Judge’s Authority to Amend Reasons for Decision

[21] As Goudge J.A. noted in Montague v. Bank of Nova Scotia (2004), 69 O.R. (3d) 87, leave to appeal refused, [2004] S.C.C.A. No. 79, at para. 34: “There can be no doubt that until a judgment is formally entered in the court record, the judge has a very broad discretion to change it”. However, he warned, at para. 40:

I have no doubt that in some cases the very wide discretion a judge has to change his or her judgment before it is entered could be abused if it were exercised for an improper purpose. Any change to a judgment once given, no matter how soundly based, runs the risk of evoking suspicions of abuse on the part of those adversely affected. It is at the least disquieting, and to that extent can put a cloud over the administration of justice. A judge exercising this discretion bears a significant onus to explain the change. Giving clear reasons for decision is always a profoundly important part of judging…It is never more important than in this circumstance. [Emphasis added.]

[22] See also, to the same effect, 1711811 Ontario Ltd. v. Buckley Insurance Brokers Ltd., 2014 ONCA 125, 371 D.L.R. (4th), per Gillese J.A. and Brown v. The Municipal Property Assessment Corp., 2014 ONSC 7137 (Div. Ct.), per Nordheimer J. (as he then was). Nordheimer J. acknowledged that there is a “fairly broad power” in a judge to change an order before it is signed and entered, but noted, at para. 20, that such a change should only be made if it is “technical” or “to avoid a miscarriage of justice”.

(2) Staying a Judgment Pending Appeal

[23] The applicable law on staying a judgment pending appeal is found in RJR-MacDonald Inc. v. Canada (A.G.), [1994] 1 S.C.R. 311, applied by analogy. The test for a stay is whether: (1) there is a serious issue to be tried; (2) the appellants will suffer irreparable harm if the stay is not granted; and (3) the balance of convenience favours a stay pending the appeal. “These three criteria are not watertight compartments,” Laskin J.A. observed in Circuit World Corp. v. Lesperance (1997), 33 O.R. (3d) 674, at p. 677, because: “The strength of one may compensate for the weakness of another”. He noted that “[g]enerally, the court must decide whether the interests of justice call for a stay”, citing International Corona Resources Ltd. v. LAC Minerals Ltd. (1986), 21 C.P.C. (2d) 252 (Ont. C.A.).

[24] The overarching consideration is the interests of justice. I am mindful of the caution expressed by Robert J. Sharpe in his text Injunctions and Specific Performance, loose-leaf, (Toronto: Canada Law Book, 2021), at para. 2:6: “The weight to be placed upon the preliminary assessment of the relative strength of the plaintiff's case is a delicate matter which will vary depending upon the context and circumstances”. This caution applies with necessary modifications to a stay pending appeal in the context of an allegation of judicial bias.

(3) Establishing an Apprehension of Judicial Bias

[25] The applicable law regarding judicial bias is found in Yukon Francophone School Board, Education Area #23 v. Yukon (A.G.), 2015 SCC 25, [2015] 2 S.C.R. 282, at para. 20, per Abella J. The test for establishing a reasonable apprehension of bias is: “[W]hat would an informed person, viewing the matter realistically and practically − and having thought the matter through − conclude. Would he think that it is more likely than not that [the decision-maker], whether consciously or unconsciously, would not decide fairly”, citing the classic formula in Committee for Justice and Liberty v. National Energy Board, [1978] 1 S.C.R. 369, at p. 394, per de Grandpré J. (dissenting). Abella J. added: “The objective of the test is to ensure not only the reality, but the appearance of a fair adjudicative process”: Yukon, at para. 22 (emphasis in original). She also added this important caution: “Because there is a strong presumption of judicial impartiality that is not easily displaced, the test for a reasonable apprehension of bias requires a ‘real likelihood or probability of bias’ and that a judge’s individual comments during a trial not be seen in isolation” (internal citations omitted).

[26] To drill down on this point, I rely on this principle from Marchand v. The Public General Hospital Society of Chatham (2000), 51 O.R. (3d) 97 (Ont. C.A.), at para. 131:

The threshold for a finding of actual or apprehended bias is high. Courts presume that judges will carry out their oath of office. Thus, to make out an allegation of judicial bias, requires cogent evidence. Suspicion is not enough. The threshold is high because a finding of bias calls into question not just the personal integrity of the judge but the integrity of the entire administration of justice.

[27] In Marchand this court summarized the principles for determining judicial bias set out by the Supreme Court of Canada in R. v. R.D.S., [1997] 3 S.C.R. 484, at para. 131. At para. 112 of R. v. R.D.S., Cory J. noted: “Nonetheless the English and Canadian case law does properly support the appellant’s contention that a real likelihood or probability of bias must be demonstrated, and that a mere suspicion is not enough”.

V. The Principles Applied

[28] I consider in turn the application of the three sets of principles, which overlap to some extent.

(1) The Application Judge’s Authority to Amend the Liability Reasons

[29] The application judge amended the reasons within the window for doing so since he had not yet signed the judgment. His reasons for making the amendments include a lengthy explanation for doing so. Whether he nonetheless fell afoul of the principles in Montague v. Bank of Nova Scotia is a matter for the panel hearing the ultimate appeal. The immediate question is whether the changes should be seen as giving rise to an apprehension of judicial bias that would justify a stay pending appeal.

(2) Staying the Judgment Pending Appeal

[30] It is common ground that the first part of the test for a stay, establishing a serious issue to be tried, is a “low bar”. The appellants argue that the low bar is easily met here because the allegation of judicial bias is a very serious one.

[31] I take a more nuanced approach. It cannot be that a mere claim of judicial bias would meet this branch of the stay test. That would discount unreasonably the “strong presumption of judicial impartiality” prescribed in Yukon, and the even stronger words in R. v. R.D.S., which require that “a real likelihood or probability of bias must be demonstrated, and that a mere suspicion is not enough”.

[32] I repeat the caution expressed by Robert J. Sharpe, quoted above: “The weight to be placed upon the preliminary assessment of the relative strength of the plaintiff's case is a delicate matter which will vary depending upon the context and circumstances”. I noted that this caution applies with necessary modifications to a stay pending appeal in the context of an allegation of judicial bias.

[33] The appellants submit that the irreparable harm, the second part of the test, flows from the fact that, unless the application is stayed, the matter will proceed to the damages trial in the face of a reasonable apprehension of bias, which will adversely affect the appearance of a fair adjudicative process.

[34] Finally, the appellants submit that the third part of the stay test, the balance of convenience, weighs strongly in favour of a stay. It would be unfair to require the damages trial to proceed before a biased judge because the judge might be guided by “unconscious instincts” to bolster his liability reasons, which would, inferentially, adversely affect the appellants’ prospects of a successful appeal.

[35] As to prejudice, the appellants submit that only effect of a stay of the judgment pending the appeal will be delay in payment, but, since the claim dates back to 1999, the additional delay is unlikely to prejudice the parties. They add that because a serious issue of judicial bias exists, the interests of justice require that further proceedings be stayed pending appeal.

(3) The Possible Apprehension of Bias

[36] I begin with the first part of the test for a stay, whether the appellants have established that judicial bias in this case constitutes a serious issue to be tried.

[37] In a nutshell, the appellants argue that by “fundamentally amending” the liability reasons, the application judge’s amendments to his reasons were neither merely technical nor the correction of inadvertent errors. The “radical amendments” made by the application judge to the liability reasons “clearly appear designed to favour the Respondent” and are submitted to give rise to an apprehension of bias.

[38] The appellants argue that a close parallel is this court’s decision in Stuart Budd & Sons Limited v. IFS Vehicle Distributors ULC, 2016 ONCA 60, 129 O.R. (3d) 37, per Epstein J.A. This court assessed the issue of bias from two perspectives. The first was the motion judge’s evident animus in accusing the party and counsel of engaging in an abuse of process and in making “derisive” comments about them (at paras. 69-78).

[39] I see nothing similar in the application judge’s approach. The recusal reasons directly address several instances the appellants put to the application judge in argument. The appellants claimed that he had disparaged them: “by imputing to them an improper interest in delay” (at para. 37); by attempting to police an appeal before the damages trial (at paras. 40-46); and by treating them and their counsel’s integrity with “unfair scepticism” (at para. 48). In my view these complaints were fairly and fully addressed by the application judge in the recusal reasons. There was nothing close to disparagement on which any sense of animus could be based.

[40] The appellants’ real complaint rests on the amendments to the liability reasons. Epstein J.A. captured this argument in Budd, at para. 86, on which the appellants rely: “In these passages, the motion judge appears to insinuate himself into the appeal process by attempting to defend his actions and his comments. It was an after-the-fact attempt not only to justify but also to bolster his decision”. Epstein J.A. explained, at para. 84:

In my view, the amended endorsement suggests that the motion judge was motivated to respond to the challenges to his decision. I refer to the motion judge's expressly identifying and then responding to grounds of appeal addressed in the reasons on the stay motion. Specifically, the motion judge dealt with two of the three "serious issues" raised in the proposed appeal -- the fact that his second endorsement only addressed jurisdiction in regards to the appellant IFS and the consequences of certain comments he made and actions he took.

[41] Epstein J.A. invoked this court’s decision in R. v. Arnaout, 2015 ONCA 655, 127 O.R. (3d) 241, at para. 23, where the court explained that, “conduct will have the effect of displacing the presumption of integrity if, in all the circumstances, an informed an reasonable observer would think that the amendment was an after-the-fact justification for the decision rather than an articulation of the reasoning that led to the decision”. Epstein J.A. put the question, at para. 83 of Budd:

Would an informed and reasonable observer see the amended endorsement as motivated, at least in part, by a desire of the motion judge to defend his decision to dismiss the motion against the specific attacks made in the appeal that, to the motion judge's knowledge, were identified in this court's reasons for granting the stay? Such motivation may, even subconsciously, lead to reasons designed to justify the challenged decision rather than to reveal the reasoning that led to the decision.

[42] The application judge was alive to this very problem and referred to Budd. He said, at para. 28 of the amendment reasons:

The parties had approached me in earlier case conferences to seek directions about the timing of the damages trial in light of the fact that the respondents were appealing the liability decision. I asked the parties not to advise me of any grounds of appeal so as to avoid even the perception that I might use my findings in the damages trial to shore up anything that a party viewed as an error in my reasons on liability. The parties have now advised me of at least some of the grounds of appeal by the respondents and now, potentially other grounds of appeal by the applicants, thereby giving rise to the risk that I had hoped to avoid. I do not say that critically of either party. Both parties are quite properly and professionally advancing the best interests of their clients. I did not feel, however, that I could do justice to the applicant’s request without providing reasons. I will leave it to others to determine what use or weight, if any, should be given to them on any appeal.

[43] The appellants point to several actions by the application judge as demonstrating his personal investment in his liability decision. The first is in refusing to sign the order before the damages trial in order to prevent an immediate appeal on the liability issue. The application judge candidly acknowledged, in para. 60 of his initial reasons refusing a stay, that he did this in order to avoid an attempt by BNY “to bootstrap itself into a right of appeal”, which he noted again in his recusal reasons, at para. 28. The second is that in the amendment reasons the application judge granted a declaration in paras. 33-34 by reaching back to the original application factum. He did this on his own and not in response to submissions from counsel.

[44] But the primary basis for the appellants’ allegation of judicial bias is in the amendments, described at length above. These amendments to the liability reasons are submitted to have had the effect of eviscerating the appellants’ main ground of appeal in the Notice of Appeal served in May 2021.

[45] The appellants argue that the respondent requested the changes in order to prevent its damages award from being reduced by at least one third. The amendments ensured that the application judge’s obvious error in finding contractual liability against CIBC Mellon would not compromise the respondent’s ability to recover damages for the sharing of SS&C data with CIBC Mellon because liability would shift from CIBC Mellon to BNY. The appellants submit that a reasonable observer, properly informed, would see these changes as exhibiting favouritism to the respondent. I do not accept this argument.

[46] The reasons given by the application judge for the amendments are cogent and responsive to the requests he faced, the arguments he heard, and to the evidence in the case with which he was deeply familiar. The reasons he gave do not reveal indicia of bias such as result-selectivity, tendentiousness, or defensiveness, all of which one might anticipate from a judge who was attempting to dispel an allegation of bias. The appellants point out that the application judge did not refer to the fact that the respondent waited until the appeal was firmly scheduled after the damages trial before asking for changes to the liability decision, which the respondent had never challenged in six prior appearances before the application judge and this court. If anything, this is a criticism of the respondent’s conduct.

[47] The application judge faced a difficult situation. There had to be clear directions for the upcoming damages trial as to what was in issue, and the respondent had raised an obvious error in the reasons. Whether that error and the application judge’s amendments exemplify a deeper problem with his reasons is a matter for the panel hearing the appeal. No one could say that this is an ideal state of affairs, but I cannot find any indicia of bias on the part of the application judge sufficient in these circumstances to warrant a stay of the judgment pending appeal. Applying the measure in R.D.S., the appellants have not demonstrated a real likelihood or probability of bias beyond a mere suspicion. I do not find that the appellants have raised a serious bias issue to be tried.

[48] While this is sufficient to dispose of the motion, I will carry on with the next steps in the stay analysis.

[49] The second part of the stay test is irreparable harm. Assuming that bias was sufficiently made out, this part would weigh against proceeding to the damages trial, but in my view it should not be dispositive. The respondent points out that a damages trial is necessary in any event regardless of the status of the claim that flows from CIBC Mellon’s data sharing. They assert that the appellants “have now admitted that at least 65 different corporate entities were given access to SS&C’s proprietary data”. This is not denied. It is hard to see how the “unconscious instincts” bias argument counsel made bears on the determination of damages related to the 64 entities other than CIBC Mellon, which appear to comprise two-thirds of the amount of damages.

[50] The third part of the stay test, the balance of convenience, does not weigh in favour of a stay, but against it. The damages trial is days away and it can be assumed that the parties are ready to go. Valuable court time has been set aside. If there is a stay pending appeal, it is difficult to predict when the damages trial could be re-scheduled, keeping in mind a predictable application for leave to appeal to the Supreme Court regardless of the outcome.

[51] The balance of convenience shades into the overarching consideration – the interests of justice, which favour both the completion of this application process and a single appeal to hear both liability and damages. The application judge’s decision to order a tailored trial on the damages issue under r. 38.10 (b) of the Rules of Civil Procedure, R.R.O. 1990, Reg. 194, as a continuous part of the application is a discretionary decision on which appellate deference is owed. Our civil justice system generally does not favour appeals between phases of a proceeding: Urbacon Building Groups Corp. v. Guelph (City), 2014 ONSC 3840, 327 O.A.C. 6, at para. 47, Korea Data Systems (USA), Inc. v. Aamazing Technologies Inc., 2012 ONCA 756, at para. 23. This policy consideration led Thorburn J.A. to refuse to expedite the liability appeal in this matter, and my refusal to schedule the liability appeal before the damages trial.

[52] The appellants have raised these issues in their Supplementary Notice of Appeal and can pursue them on the appeal on a full record. I give leave to file the Supplementary Notice of Appeal.

[53] I would be remiss if I did not briefly address the respondent’s argument on this court’s jurisdiction to hear this motion. It is common ground that a judge’s refusal to recuse is an interlocutory decision appealable only to the Divisional Court with leave. This is, the respondent says, an effort to appeal the recusal decision by making “an end run around the high threshold for obtaining leave”. There was more to the appellants’ motion than that and I considered it to be prudent to deal with it on the merits in the interests of clarity and certainty heading into the imminent damages trial.

VI. Disposition

[54] The motion is dismissed, with costs to the respondent in the amount of $12,000 all-inclusive, as agreed.

“P. Lauwers J.A.”