SUPREME COURT OF CANADA

Appeal Heard: February 15, 2017 Judgment Rendered: December 20, 2017 Docket: 36875

Between:

Deloitte & Touche (now continued as Deloitte LLP)

Appellant

and

Livent Inc., through its special receiver and manager Roman Doroniuk

Respondent

- and -

Canadian Coalition for Good Governance and Chartered Professional Accountants of Canada

Interveners

Coram: McLachlin C.J. and Karakatsanis, Wagner, Gascon, Côté, Brown and Rowe JJ.

Joint Reasons for Judgment: (paras. 1 to 115)

Gascon and Brown JJ. (Karakatsanis and Rowe JJ. concurring)

Reasons Dissenting in Part: (paras. 116 to 178)

McLachlin C.J. (Wagner and Côté JJ. concurring)

Note: This document is subject to editorial revision before its reproduction in final form in the Canada Supreme Court Reports .

deloitte & touche v. livent inc.

Deloitte & Touche (now continued as Deloitte LLP) Appellant

v.

Livent Inc., through its special receiver and

manager Roman Doroniuk Respondent

and

Canadian Coalition for Good Governance and

Chartered Professional Accountants of Canada Interveners

Indexed as: Deloitte & Touche v. Livent Inc. (Receiver of)

2017 SCC 63

File No.: 36875.

2017: February 15; 2017: December 20.

Present: McLachlin C.J. and Karakatsanis, Wagner, Gascon, Côté, Brown and Rowe JJ.

on appeal from the court of appeal for ontario

Torts — Duty of care — Negligence — Negligent misrepresentation — Auditor failing to discover fraud by company’s directors and company incurring losses — Proper application of analytical framework for establishing tort liability in cases of negligent misrepresentation or performance of service by auditor — Whether auditor breaching duty of care and therefore liable for company’s losses — Appropriate date from which to calculate quantum of damages.

Livent produced and staged performances in theatres that it owned in Canada and the U.S., with its shares listed on Canadian and U.S. stock exchanges. To enhance Livent’s success, its directors manipulated the company’s financial records. Deloitte was Livent’s auditor. Deloitte never uncovered the fraud. In August 1997, however, Deloitte identified irregularities in the reporting of profit from an asset sale. Deloitte did not resign. Instead, for the purpose of helping Livent to solicit investment, Deloitte helped prepare, and approved, a press release issued in September 1997, which misrepresented the basis for the reporting of the profit. In October 1997, Deloitte provided a comfort letter for a public offering. It also prepared Livent’s 1997 audit, which it finalized in April 1998. New equity investors later discovered the fraud. A subsequent investigation and re‑audit resulted in restated financial reports. Livent filed for insolvency protection in November 1998. It sold its assets and went into receivership in 1999. Livent sued Deloitte later in tort and contract.

The trial judge held that Deloitte owed a duty of care to provide accurate information to Livent’s shareholders. He held that Deloitte failed to meet the standard of care under this duty, either when it failed to discover the fraud and act on that discovery in August 1997, or when it signed off on Livent’s 1997 financial statements in April 1998. The trial judge held that the measure of damages was the difference between Livent’s value on the date on which Deloitte should have resigned and Livent’s value at the time of insolvency. He reduced this by 25 percent to account for contingencies or trading losses, which he held were too remote to make Deloitte liable. The trial judge consequently awarded damages to Livent for breach of its duty of care, and alternatively for breach of contract, in the amount of $84,750,000. The Court of Appeal upheld the trial judge’s award and dismissed Deloitte’s appeal and Livent’s cross‑appeal.

Held (McLachlin C.J. and Wagner and Côté JJ. dissenting in part): The appeal should be allowed in part.

Per Karakatsanis, Gascon , Brown and Rowe JJ.: The general framework set out in Anns v. London Borough of Merton , [1977] 2 All E.R. 492 (H.L.) , and later refined in Cooper v. Hobart , 2001 SCC 79 , [2001] 3 S.C.R. 537, applies in cases of pure economic loss arising from an auditor’s negligent misrepresentation or performance of a service . Comprising two stages, the Anns/Cooper framework asks whether a prima facie duty of care exists between the parties, and if so, whether there are any residual policy considerations that may negate the imposition of a duty of care. At the first stage, a prima facie duty of care is recognized where proximity and reasonable foreseeability of injury are established. When assessing proximity, if a relationship falls within a previously established category, or is analogous to one, then the requisite close and direct relationship is shown. If a risk of reasonably foreseeable injury can also be shown, the first stage of the Anns/Cooper framework is complete and a duty of care may be identified. In such circumstances, the second stage of the framework will seldom be engaged because any residual policy considerations will have already been taken into account when the proximate relationship was first identified.

Where an established proximate relationship cannot be found, courts must undertake a full proximity analysis. To determine whether the close and direct relationship exists, courts must examine all relevant factors arising from the relationship. In cases of pure economic loss arising from negligent misrepresentation or performance of a service, two factors are determinative in the proximity analysis: the defendant’s undertaking and the plaintiff’s reliance. Where the defendant undertakes to provide a representation or service in circumstances that invite the plaintiff’s reasonable reliance, the defendant becomes obligated to take reasonable care and the plaintiff has a right to rely on the defendant’s undertaking. These corollary rights and obligations create a relationship of proximity. Any reliance on the part of the plaintiff, which falls outside of the scope of the defendant’s undertaking, necessarily falls outside the scope of the proximate relationship and, therefore, of the defendant’s duty of care. This properly limits liability on the basis that the defendant cannot be liable for a risk of injury against which he did not undertake to protect.

As for assessing reasonable foreseeability in the prima facie duty of care analysis, this entails asking whether an injury to the plaintiff was a reasonably foreseeable consequence of the defendant’s negligence. Reasonable foreseeability concerns the likelihood of injury arising from the defendant’s negligence. In cases of negligent misrepresentation or performance of a service, the proximate relationship informs the foreseeability inquiry. The purpose underlying the undertaking and the corresponding reliance limits the type of injury that could be reasonably foreseen to result from the defendant’s negligence. An injury to the plaintiff will be reasonably foreseeable if the defendant should have reasonably foreseen that the plaintiff would rely on his or her representation and such reliance would, in the particular circumstances of the case, be reasonable. Both the reasonableness and the reasonable foreseeability of the plaintiff’s reliance will be determined by the relationship of proximity between the parties.

At the second stage of the Anns/Cooper framework, the question is whether there are residual policy considerations outside the relationship of the parties that may negate the imposition of a duty of care. This stage is not concerned with the relationship between the parties, but with the effect of recognizing a duty of care on other legal obligations, the legal system and society more generally. Factors to be considered include whether the law already provides a remedy, the spectre of unlimited liability to an unlimited class and whether there are other reasons of broad policy that suggest that the duty of care should not be recognized. The place within the Anns/Cooper framework of this policy inquiry is significant. It follows the proximity and foreseeability inquiries. The policy inquiry assesses whether, despite the proximate relationship between the parties and the reasonably foreseeable quality of the plaintiff’s injury, the defendant should nonetheless be insulated from liability. That it would limit liability in the face of findings of both proximity and reasonable foreseeability makes plain how narrowly it should be relied upon.

No proximate relationship has previously been established as between an auditor and its client for the purposes of soliciting investment. This case therefore requires a full proximity analysis. From August to October 1997, the services which Deloitte provided to Livent — particularly its ongoing assistance in relation to the press release and the provision of the comfort letter — were undertaken for the purpose of helping Livent to solicit investment. Given this undertaking, Livent was entitled to rely upon Deloitte to carry out these services with reasonable care. It follows that a relationship of proximity arose but only in respect of the content of Deloitte’s undertaking. Losses outside the scope of this undertaking are not recoverable from Deloitte. With respect to the press release and the comfort letter, Deloitte never undertook to assist Livent’s shareholders in overseeing management; it cannot therefore be held liable for failing to take reasonable care to assist such oversight. Given that Livent had no right to rely on Deloitte’s representations for a purpose other than that for which Deloitte undertook to act, Livent’s reliance was neither reasonable nor reasonably foreseeable. Consequently, the increase in Livent’s losses or liquidation deficit, which arose from that reliance, was not a reasonably foreseeable injury. Because no prima facie duty of care arose, there is no need to consider residual policy considerations.

However, the Court has already recognized that a duty is owed by an auditor in preparing a statutory audit and that a claim by a corporation for losses resulting from a negligent statutory audit could succeed. A statutory audit is prepared to allow shareholders to collectively supervise management and to take decisions with respect to the overall administration of the corporation. This describes precisely the function which Livent’s shareholders were unable to discharge by reason of Deloitte’s negligent 1997 audit. Deloitte did not alter the purpose for which it undertook to provide the 1997 audit or disclaim liability in relation to that purpose. Therefore, proximity is established in relation to the statutory audit, on the basis of the previously recognized proximate relationship. In addition, the type of injury Livent suffered was a reasonably foreseeable consequence of Deloitte’s negligence. Through the 1997 audit, Deloitte undertook to assist Livent’s shareholders in scrutinizing management conduct. By negligently conducting the audit, and impairing Livent’s shareholders’ ability to oversee management, Deloitte exposed Livent to reasonably foreseeable risks, including losses that would have been avoided with a proper audit. Because proximity is based on a previously recognized category, there is no need to consider residual policy considerations. Deloitte owed Livent a duty of care, which it breached. Deloitte cannot rely on either the defence of illegality or of contributory fault, because the fraudulent acts of Livent’s directors cannot be attributed to the corporation.

Remoteness is not a bar to Livent’s recovery. Remoteness examines whether the harm is too unrelated to the wrongful conduct to hold the defendant fairly liable. It overlaps conceptually with the reasonable foreseeability analysis but the duty of care analysis is concerned with the type of injury that is reasonably foreseeable as flowing from the defendant’s conduct, whereas the remoteness analysis is concerned with the actual injury suffered by the plaintiff. However, the loss here — stemming from Deloitte’s failure to fulfill the specific undertaking it made to Livent in relation to the 1997 audit — was reasonably foreseeable.

The trial judge assessed Livent’s damages following the 1997 audit at $53.9 million. Applying the trial judge’s 25 percent contingency reduction to this amount results in a final damages assessment of $40,425,000. This is the amount for which Deloitte is liable. At trial, Livent conceded that its losses for negligent performance of a service or breach of contract would be identical. Therefore, the same quantum of liability applies for Deloitte’s concurrent claim in breach of contract.

Per McLachlin C.J. and Wagner and Côté JJ. (dissenting in part): Deloitte owed a duty of care to Livent, which it breached when it failed to discover and expose Livent’s fraud in the audited statements. However, Deloitte is not liable for the loss that befell Livent. The claim in tort must be dismissed. The result is the same with respect to Livent’s action in contract.

Courts have provided two doctrinal approaches for limiting recovery of pure economic loss flowing from negligent misstatement. The first is to hold that the scope of the duty of care of the advice‑giver does not cover the loss claimed. The second is to hold that the loss is too remote from the negligent act and thus was not legally caused by that act. Both inquiries invoke similar considerations and arrive at the same point. The remoteness inquiry looks at the wrongdoing and its proximity to the loss claimed. The factors to be considered are not closed. The advice‑giver’s knowledge of the claimant’s circumstances, the reasonable expectations arising from the relationship, and the presence of intervening factors that led to the loss may figure in the analysis. The scope of the duty of care inquiry looks to the relationship between the defendant’s advice and the plaintiff’s loss. It asks if that relationship was proximate. In cases of economic loss, it inquires into the purpose for which the advice was given and asks whether a reasonable person would have expected, or foreseen, that negligent advice would lead to the loss in question by virtue of the plaintiff’s reliance on the advice.

The duty of care inquiry leads to the two‑part test set out in Anns v. London Borough of Merton , [1977] 2 All E.R. 492 (H.L.) . The first part of the test asks whether there is proximity, or a sufficiently close relationship, between the parties. It focuses on the connection between the defendant’s undertaking or statement and the loss claimed. The purpose for which the statement was made is pivotal, and is a matter of fact to be determined on the evidence adduced at trial.

In this case, three purposes of Livent’s audit statements are discernable: (1) to report accurately on Livent’s finances and provide it with audit opinions on which it could rely for the purpose of attracting investment; (2) to uncover errors or wrongdoing for the purpose of enabling Livent itself to correct or otherwise respond to the misfeasance; and (3) to provide audit reports on which Livent’s shareholders could rely to supervise Livent’s management. The scope of Deloitte’s duty of care is defined solely by these purposes.

Deloitte’s wrongful act did not deprive Livent of the ability to attract investment capital. In fact, Livent attracted a great deal of capital on the strength of Deloitte’s statements. Likewise, Deloitte’s wrongful act did not prevent Livent from detecting misfeasance in the company’s management, which Livent would have corrected had it known. Finally, Livent did not prove that Deloitte’s wrongdoing prevented its shareholders from exercising supervision in a manner that would have ended the company’s loss‑creating activities at an earlier date. The trial judge did not find that Livent’s shareholders relied on Deloitte’s negligent audit statements, or that had they received and relied on accurate statements, they would have acted in a way that would have prevented Livent from carrying on business and diminishing its assets in the period between the issuance of the relevant statements and Livent’s insolvency. Crucially, the trial judge did not ask whether the shareholders had in fact relied on the audits and he did not ask whether, if they had relied, this reliance prevented them from taking steps to alter course. Finally, he did not ask whether these actions, had they been taken, would have prevented the losses that Livent built up during the seven‑month period in question. If the trial judge had asked these questions, he would have been obliged to answer them in the negative, since Livent offered no proof to support affirmative answers. As a result, the factual basis for establishing loss on the basis of shareholder supervision was entirely lacking.

The majority suggests that, had Deloitte provided sound audit reports, Livent’s shareholders and management may have made decisions that would have limited the company’s losses. While this may be true, it is not enough to rely on unproven assertions to define the scope of the duty of care and to subsequently demonstrate causation. The majority’s approach suggests that an auditor will generally become the underwriter for any losses suffered by a client following a negligent audit report. This, notwithstanding subsequent decisions — reliant or capricious — made by the client’s shareholders. However, reliance cannot be presumed; it must be proved.

Because the loss at issue has not been shown to fall within the scope of Deloitte’s duty of care, the first step of the Anns test is not established. It is therefore unnecessary to go on to ask whether prima facie liability is negated by policy considerations unrelated to the relationship between the parties. However, were it necessary to do so, the policy considerations of unfair allocation of loss and indeterminacy would preclude imposing liability on Deloitte.

Cases Cited

By Gascon and Brown JJ.

Applied: Hercules Managements Ltd. v. Ernst & Young , [1997] 2 S.C.R. 165; distinguished : South Australia Asset Management Corp. v. York Montague Ltd. , [1997] A.C. 191 ; Canadian Dredge & Dock Co. v. The Queen , [1985] 1 S.C.R. 662; Hart Building Supplies Ltd. v. Deloitte & Touche , 2004 BCSC 55 , 41 C.C.L.T. (3d) 240; explained: Anns v. London Borough of Merton , [1977] 2 All E.R. 492 ; Cooper v. Hobart , 2001 SCC 79 , [2001] 3 S.C.R. 537; referred to: Bow Valley Husky (Bermuda) Ltd. v. Saint John Shipbuilding Ltd. , [1997] 3 S.C.R. 1210; Canadian National Railway Co. v. Norsk Pacific Steamship Co. , [1992] 1 S.C.R. 1021; Kamloops (City) v. Nielsen , [1984] 2 S.C.R. 2; Haig v. Bamford , [1977] 1 S.C.R. 466; Edwards v. Law Society of Upper Canada , 2001 SCC 80 , [2001] 3 S.C.R. 562; Odhavji Estate v. Woodhouse , 2003 SCC 69 , [2003] 3 S.C.R. 263; Childs v. Desormeaux , 2006 SCC 18 , [2006] 1 S.C.R. 643; Hill v. Hamilton‑Wentworth Regional Police Services Board , 2007 SCC 41 , [2007] 3 S.C.R. 129; Fullowka v. Pinkerton’s of Canada Ltd. , 2010 SCC 5 , [2010] 1 S.C.R. 132; Saadati v. Moorhead , 2017 SCC 28 , [2017] 1 S.C.R. 543; Donoghue v. Stevenson , [1932] A.C. 562; Caparo Industries plc. v. Dickman , [1990] 1 All E.R. 568 ; Glanzer v. Shepard , 135 N.E. 275 (1922); Ultramares Corp. v. Touche , 174 N.E. 441 (1931); Yuen Kun Yeu v. Attorney‑General of Hong Kong , [1988] 1 A.C. 175 ; Edgeworth Construction Ltd. v. N. D. Lea & Associates Ltd. , [1993] 3 S.C.R. 206; Gross v. Great‑West Life Assurance Co. , 2002 ABCA 37 , 299 A.R. 142; Mustapha v. Culligan of Canada Ltd. , 2008 SCC 27 , [2008] 2 S.C.R. 114; Overseas Tankship (U.K.) Ltd. v. Morts Dock & Engineering Co. , [1961] A.C. 388 ; Hughes‑Holland v. BPE Solicitors , [2017] UKSC 21, [2017] 2 W.L.R. 1029 ; Nykredit Mortgage Bank plc. v. Edward Erdman Group Ltd. (No. 2) , [1997] 1 W.L.R. 1627 ; Platform Home Loans Ltd. v. Oyston Shipways Ltd. , [2000] 2 A.C. 190 ; Clements v. Clements , 2012 SCC 32 , [2012] 2 S.C.R. 181; Rainbow Industrial Caterers Ltd. v. Canadian National Railway Co. , [1991] 3 S.C.R. 3; Hall v. Hebert , [1993] 2 S.C.R. 159; British Columbia v. Zastowny , 2008 SCC 4 , [2008] 1 S.C.R. 27; Stone & Rolls Ltd. (in liquidation) v. Moore Stephens , [2009] UKHL 39, [2009] 1 A.C. 1391 ; 373409 Alberta Ltd. (Receiver of) v. Bank of Montreal , 2002 SCC 81 , [2002] 4 S.C.R. 312; Bilta (U.K.) Ltd. (in liquidation) v. Nazir (No. 2) , [2015] UKSC 23, [2016] A.C. 1 .

By McLachlin C.J. (dissenting in part)

Caparo Industries plc. v. Dickman , [1990] 1 All E.R. 568 ; Ultramares Corp. v. Touche , 174 N.E. 441 (1931); D’Amato v. Badger , [1996] 2 S.C.R. 1071; Hercules Managements Ltd. v. Ernst & Young , [1997] 2 S.C.R. 165; Canadian National Railway Co. v. Norsk Pacific Steamship Co. , [1992] 1 S.C.R. 1021; R. v. Imperial Tobacco Canada Ltd. , 2011 SCC 42 , [2011] 3 S.C.R. 45; Cooper v. Hobart , 2001 SCC 79 , [2001] 3 S.C.R. 537; BG Checo International Ltd. v. British Columbia Hydro and Power Authority , [1993] 1 S.C.R. 12; South Australia Asset Management Corp. v. York Montague Ltd. , [1996] 3 All E.R. 365 ; Hughes‑Holland v. BPE Solicitors , [2017] UKSC 21, [2017] 2 W.L.R. 1029 ; Hogarth v. Rocky Mountain Slate Inc. , 2013 ABCA 57 , 542 A.R. 289; Wightman v. Widdrington (Succession) , 2013 QCCA 1187 ; Platform Home Loans Ltd. v. Oyston Shipways Ltd. , [1999] 1 All E.R. 833 ; Mustapha v. Culligan of Canada Ltd. , 2008 SCC 27 , [2008] 2 S.C.R. 114; Citadel General Assurance Co. v. Vytlingam , 2007 SCC 46 , [2007] 3 S.C.R. 373; Westmount (City) v. Rossy , 2012 SCC 30 , [2012] 2 S.C.R. 136; Anns v. London Borough of Merton , [1977] 2 All E.R. 492 ; Sutherland Shire Council v. Heyman (1985), 1988 ABCA 234, 60 A.L.R. 1; Overseas Tankship (U.K.) Ltd. v. Morts Dock & Engineering Co. , [1961] A.C. 388 ; Candler v. Crane Christmas & Co. , [1951] 1 All E.R. 426 ; Burns v. Homer Street Development Limited Partnership , 2016 BCCA 371 , 91 B.C.L.R. (5th) 383; Aneco Reinsurance Underwriting Ltd. (in liquidation) v. Johnson & Higgins Ltd. , [2001] UKHL 51 , [2001] 2 All E.R. (Comm.) 929; Canadian Imperial Bank of Commerce v. Deloitte & Touche , 2016 ONCA 922 , 133 O.R. (3d) 561; Temseel Holdings Ltd. v. Beaumonts Chartered Accountants , [2002] EWHC 2642 (Comm.), [2003] P.N.L.R. 27; B.D.C. Ltd. v. Hofstrand Farms Ltd. , [1986] 1 S.C.R. 228; Asamera Oil Corp. v. Sea Oil & General Corp. , [1979] 1 S.C.R. 633.

Statutes and Regulations Cited

Business Corporations Act , R.S.O. 1990, c. B.16, Part XII, ss. 153, 154.

Companies’ Creditors Arrangement Act , R.S.C. 1985, c. C‑36.

Corporations Act , R.S.M. 1987, c. C225.

Negligence Act , R.S.O. 1990, c. N.1, s. 3.

Authors Cited

Beever, Allan. Rediscovering the Law of Negligence . Oxford: Hart, 2007.

Black’s Law Dictionary , 10th ed. by Bryan A. Garner, ed. St. Paul, Minn.: Thomson Reuters, 2014, “indeterminate”.

Blom, Joost. “Do We Really Need the Anns Test for Duty of Care in Negligence?” (2016), 53 Alta. L. Rev. 895.

Hohfeld, Wesley Newcomb. “Some Fundamental Legal Conceptions as Applied in Judicial Reasoning” (1913), 23 Yale L.J. 16.

Jutras, Daniel. “Civil Law and Pure Economic Loss: What Are We Missing?” (1987), 12 Can. Bus. L.J. 295.

Klar, Lewis N., and Cameron S. G. Jefferies. Tort Law , 6th ed. Toronto: Thomson Reuters, 2017.

Linden, Allen M., and Bruce Feldthusen. Canadian Tort Law , 8th ed. Markham, Ont.: LexisNexis, 2006.

MacPherson, Darcy L. “Emaciating the Statutory Audit — A Comment on Hart Building Supplies Ltd. v. Deloitte & Touche ” (2005), 41 Can. Bus. L.J. 471.

Tushnet, Mark V. “Defending the Indeterminacy Thesis”, in Brian Bix, ed., Analyzing Law: New Essays in Legal Theory . Oxford: Clarendon Press, 1998, 223.

Waddams, S. M. The Law of Damages , 5th ed. Toronto: Canada Law Book, 2012.

Weinrib, Ernest J. “The Disintegration of Duty” (2006), 31 Adv. Q. 212.

APPEAL from a judgment of the Ontario Court of Appeal (Strathy C.J. and Blair and Lauwers JJ.A.), 2016 ONCA 11 , 128 O.R. (3d) 225, 393 D.L.R. (4th) 1, 342 O.A.C. 201, 52 B.L.R. (5th) 225, 31 C.B.R. (6th) 205, 24 C.C.L.T. (4th) 177, [2016] O.J. No. 51 (QL), 2016 CarswellOnt 122 (WL Can.), affirming a decision of Gans J., 2014 ONSC 2176 , 11 C.B.R. (6th) 12, 10 C.C.L.T. (4th) 182, 26 B.L.R. (5th) 15, [2014] O.J. No. 1635 (QL), 2014 CarswellOnt 4365 (WL Can.). Appeal allowed in part, McLachlin C.J. and Wagner and Côté JJ. dissenting in part.

Peter H. Griffin , Matthew Fleming , Scott Rollwagen and Nina Bombier , for the appellant.

Peter F. C. Howard , Patrick O’Kelly , Nicholas McHaffie and Aaron Kreaden , for the respondent.

Markus Koehnen , David Kent and Jeffrey Levine , for the intervener the Canadian Coalition for Good Governance.

Guy J. Pratte , Nadia Effendi and Duncan A. W. Ault , for the intervener Chartered Professional Accountants of Canada.

The judgment of Karakatsanis, Gascon, Brown and Rowe JJ. was delivered by

Gascon and Brown JJ. —

I. Introduction

[ 1 ] This appeal provides the Court with an opportunity to affirm the analytical framework by which liability may be imposed in cases of negligent misrepresentation or performance of a service by an auditor.

[ 2 ] There is substantial agreement between us and the Chief Justice. We agree on the general analytical framework governing negligent misrepresentation claims (Chief Justice’s reasons, at paras. 146-147). And we agree that Deloitte & Touche (now Deloitte LLP) should not be liable for its corporate client Livent Inc.’s increase in liquidation deficit which followed Deloitte’s provision of negligent services in relation to the solicitation of investment.

[ 3 ] We conclude, however, that Deloitte should be liable for the increase in Livent’s liquidation deficit which followed the statutory audit. In H ercules Managements Ltd. v. Ernst & Young , [1997] 2 S.C.R. 165, this Court recognized that a statutory audit is prepared to allow shareholders to collectively “supervise management and to take decisions with respect to matters concerning the proper overall administration of the corporatio[n]” which permits “the shareholders, acting as a group, to safeguard the interests of the corporatio[n]” (para. 56 (emphasis deleted)). This describes precisely the function which Livent’s shareholders were unable to discharge by reason of Deloitte’s negligence. As a consequence, Livent’s corporate life was artificially prolonged, resulting in the interim deterioration of its finances. There was a sufficient evidentiary basis for liability based on impaired shareholder supervision. Application of the Anns/Cooper framework, coupled with the basis for auditor liability specifically identified by this Court in Hercules , would lead us to uphold the trial judge’s finding of liability in relation to the negligently prepared statutory audit.

[ 4 ] As a result, we would allow the appeal from the decision of the Ontario Court of Appeal, 2016 ONCA 11 , 128 O.R. (3d) 225, but only in part.

II. Facts and Judicial History

[ 5 ] We generally agree with the facts and judicial history set out by the Chief Justice in her reasons. In particular, she correctly identifies the trial judge’s core finding that Deloitte’s conduct fell below the standard of care on two occasions: “. . . either when it failed to discover the fraud and act on that discovery in August 1997, or when it signed off on Livent’s 1997 financial statements in April 1998” (Chief Justice’s reasons, at para. 127; trial reasons, 2014 ONSC 2176 , 10 C.C.L.T. (4th) 182, at paras. 241-42 ). We, like the Chief Justice, do not dispute these core findings. Some further elaboration upon them is, however, helpful.

[ 6 ] The trial judge’s findings of negligence can be divided into two separate events: (1) Deloitte’s approval of a 1997 press release (“Press Release”) and provision of a comfort letter (“Comfort Letter”); and (2) Deloitte’s preparation and approval of the 1997 clean audit opinion (“1997 Audit”). We would not label all of these documents “audit statements”. Indeed, collapsing the distinctions between these documents obfuscates a proper duty of care analysis.

[ 7 ] Livent asserts that it detrimentally relied on Deloitte in each of these events, which impaired its ability to oversee its operations. Specifically, Livent says that, had Deloitte been prudent in relation to these representations, Livent’s life would not have been artificially extended and that, in turn, it would have suffered less corporate loss (calculated as the increase in the deficit between its liabilities and assets at the time of its liquidation): trial reasons, at paras. 23-25, citing Livent’s amended statement of claim, at paras. 210 and 212. A detailed recounting of the events pertaining to these two representations is, therefore, critical to the negligence analysis in this case.

A. Primary Negligence Finding: The Press Release and Comfort Letter (August to October 1997)

[ 8 ] Chronologically, the first representations found by the trial judge to be negligence causing compensable harm were the Press Release and Comfort Letter.

[ 9 ] The Comfort Letter pertains to an agreement whereby Dundee Realty Corp. sought to purchase the air rights above Livent’s Pantages Theatre and adjacent lands (“Air Rights Agreement”). Deloitte audited the accounting and reporting relating to that purchase, and identified irregularities in the accounting for the reporting of profit. Ultimately, Livent and Deloitte disagreed about the irregularities, which left Deloitte with a choice between resigning (and reporting those irregularities to regulatory authorities and the next auditor), and remaining (thereby effectively capitulating to Livent’s views on how the irregularities should be reported). Deloitte, negligently, chose the latter route. It did not resign or inform anyone of the accounting irregularities. Instead, it helped prepare, and approved, the Press Release of September 2, 1997, which misrepresented the basis for the reporting of profit arising from the Air Rights Agreement.

[ 10 ] Further, that Press Release was issued “on the eve of a public offering for which [Deloitte] was going to have to provide a comfort letter” (trial reasons, at para. 193). As a result, Deloitte — again, negligently — provided the Comfort Letter on October 10, 1997, in support of the U.S. $125 million debenture underwriting. The purpose underlying the Press Release and the Comfort Letter is critical. It was not to inform Livent of its own financial position, but rather, to inform investors of Livent’s financial position, furnishing “comfort” in respect of their investment (despite one of Deloitte’s senior partners’ express acknowledgment that Deloitte was in no position “to provide any comfort to any regulators, underwriters or audit committee members as to the interim financial statement’s conformity with GAAP”) (trial reasons, at para. 178 (emphasis added; emphasis in original deleted)). Casting “professional skepticism, if not GAAS, aside” (para. 209), Deloitte approved the Press Release and Comfort Letter — all, seemingly, to maintain its profitable relationship with Livent.

[ 11 ] Given the foregoing, the trial judge assessed Livent’s injury as of a “measurement date” of August 31, 1997, i.e., the date on which Deloitte would, acting reasonably, have resigned. The trial judge also reduced Livent’s damages by 25 percent, however, for “contingencies” said to represent the amount Livent would have lost, even without Deloitte’s negligence.

[ 12 ] Deloitte appeals the trial judge’s award of damages, which amounts to the measure of damages (75 percent of damages overall) that he estimated to have arisen after the date on which Deloitte should have resigned.

B. Alternative Negligence Finding: The 1997 Audit (April 1998)

[ 13 ] In the alternative, the trial judge held that, if Deloitte reasonably refrained from resigning in August or September of 1997, it was negligent in preparing the 1997 Audit which was finalized in April 1998. That audit, which lacked “independent thought”, essentially tracked the statutory audit for 1996 (“1996 Audit”), despite (1) Livent now presenting inordinate risk given its “more than . . . modest history of aggressive, if not questionable, accounting practices” (trial reasons, at para. 211); and (2) Deloitte discovering, before the audit was completed, that Livent had intentionally deceived it as to the nature of its contractual dealings underlying the Air Rights Agreement. Somehow, this latter discovery of deliberate deception — after which “all hell broke loose” (para. 213) — was not enough to persuade Deloitte to terminate its engagement with Livent, despite all of its testifying senior partners acknowledging that “their collective professional skepticism would have been at the highest level” at this time (para. 214), and despite Livent’s after-the-fact explanation for this deception “ma[king] no sense, whatsoever” (para. 234(5)). Deloitte’s willingness to succumb to Livent’s transparently fraudulent demands left the trial judge “breathless” (para. 238) and was “beyond [his] comprehension” (para. 239).

[ 14 ] Given the foregoing, the trial judge also assessed Livent’s injury as of an “alternative” measurement date of March 31, 1998, i.e., the date on which Deloitte would, acting reasonably, have provided a prudent audit opinion (trial reasons, at para. 306, fn. 188, and para. 369, fn. 228).

[ 15 ] We reiterate that the purpose of the representation is critical. Unlike the Press Release and Comfort Letter (which were intended to inform investors of Livent’s financial position), the 1997 Audit was intended to inform Livent of its own financial position for various purposes, including, most importantly, shareholder oversight of management.

III. Analysis

A. Duty of Care

[ 16 ] Traditionally, the test from Anns v. London Borough of Merton , [1977] 2 All E.R. 492 (H.L.) , governed the duty analysis in decisions of this Court addressing claims for pure economic loss ( Hercules ; Bow Valley Husky (Bermuda) Ltd. v. Saint John Shipbuilding Ltd. , [1997] 3 S.C.R. 1210; Canadian National Railway Co. v. Norsk Pacific Steamship Co. , [1992] 1 S.C.R. 1021). Significantly, however, the Anns test for establishing tort liability in Canada has since been refined. In Cooper v. Hobart , 2001 SCC 79 , [2001] 3 S.C.R. 537, this Court provided greater certainty to the law of tort by clarifying the factors which may be considered at each stage of the Anns test. While the resulting Anns/Cooper framework has yet to be applied by this Court in a case of auditor’s negligence, we adopt this statement of La Forest J. for the Court in Hercules : “. . . to create a ‘pocket’ of negligent misrepresentation cases . . . in which the existence of a duty of care is determined differently from other negligence cases would, in my view, be incorrect” (para. 21).

[ 17 ] We turn, therefore, to consider the test for establishing tort liability, beginning with this Court’s decision in Hercules , and the proper application of the general Anns/Cooper framework to cases of auditors’ liability.

(1) Hercules : The Anns Test

[ 18 ] In Hercules , this Court recognized a duty owed by an auditor in preparing a statutory audit of its corporate client. While the Court dismissed the plaintiff shareholders’ claim for lost personal investments, it consistently maintained that a claim by the corporation itself for its own losses resulting from a negligent statutory audit could have succeeded (paras. 58-59; see also paras. 1 and 60-64):

All the participants in this appeal . . . raised the issue of whether the appellants’ claims in respect of the losses they suffered in their existing shareholdings through their alleged inability to oversee management of the corporations ought to have been brought as a derivative action . . . .

. . . if an action is to be brought in respect of such losses, it must be brought either by the corporation itself (through management) or by way of a derivative action.

[ 19 ] The duty analysis in Hercules entailed applying the then-current test for recognizing a duty of care in Canadian negligence law: the Anns test. Comprising two stages, the Anns test asked (1) whether a prima facie duty of care exists between the parties; and (2) if so, whether there are any residual policy considerations which should negate or limit the scope of the duty, the class of persons to whom it is owed or the damages to which a breach of it may give rise ( Hercules , at para. 20; Kamloops (City) v. Nielsen , [1984] 2 S.C.R. 2, at pp. 10-11; Norsk , at p. 1155; Bow Valley , at para. 47).

[ 20 ] Under the Anns test, a prima facie duty of care is recognized where a “sufficiently close relationship between the plaintiff and the defendant” exists such that “in the reasonable contemplation of the [defendant], carelessness on its part may cause damage to the [plaintiff]” ( Hercules , at para. 22; Kamloops , at p. 10). In other words, where injury to the plaintiff is a reasonably foreseeable consequence of the defendant’s negligence, a duty of care would, prima facie , arise. This relationship, where present, was labelled one of “proximity” ( ibid. ). In Hercules , the Court provided greater particularity to the test of reasonable foreseeability which established proximity under the Anns test in the context of claims for pure economic loss arising from negligent misrepresentation or performance of a service. Specifically, it stated that proximity would inhere in a relationship where two criteria are met: (1) that the defendant should reasonably foresee that the plaintiff will rely on his or her representation; and (2) that the plaintiff’s reliance would, in the circumstances of the case, be reasonable. The Court explained that considering the plaintiff’s reliance within the test for the reasonable foreseeability of injury did not “abandon the basic tenets underlying the [ Anns ] formula” (para. 25). Rather, as the plaintiff’s injury in cases of pure economic loss arising from negligent misrepresentation or performance of a service stems from his or her detrimental reliance, the reasonableness of that reliance informs the determination of whether his or her injury is reasonably foreseeable (paras. 25-26). Where, therefore, the Anns test was applied to cases of negligent misrepresentation, reasonable foreseeability of injury alone, as arising from reasonable reliance, was sufficient to establish a proximate relationship supporting a prima facie duty of care ( Hercules , at paras. 25 and 27; Norsk , at p. 1154; Bow Valley , at para. 61).

[ 21 ] The Anns test thereby set a low threshold at the first stage, imposing duties in relation to a nearly limitless class of persons who might rely on representations for nearly limitless purposes. Indeed, as this Court stated in Hercules , “[i]n modern commercial society, the fact that audit reports will be relied on by many different people (e.g., shareholders, creditors, potential takeover bidders, investors, etc.) for a wide variety of purposes will almost always be reasonably foreseeable to auditors themselves” (para. 32). For that reason — that is, because of the low “foreseeability” threshold for establishing a prima facie duty of care at the first stage of the Anns test — the Court looked to the second stage of the Anns test to negate or narrow the duty on the basis of the “policy consideration” of indeterminacy. It was here that the Court looked to the identity of the plaintiffs and the purpose of the audit opinion to deny liability for investment and devaluation losses of individual shareholders (paras. 27-28; see also Haig v. Bamford , [1977] 1 S.C.R. 466). Specifically, the Court found that one of the purposes of a statutory audit — that is, to “allo[w] shareholders, as a group , to supervise management and to take decisions with respect to matters concerning the proper overall administration of the corporatio [ n ]” (para. 56 (emphasis in original)) — would have permitted the corporate client to recover its own losses at the time of receivership had the claim been brought in the corporation’s name. As we will explain, Livent’s injury following the 1997 Audit is precisely the type of injury described in Hercules as being compensable.

(2) Cooper : Refining the Anns Test

[ 22 ] While this Court’s holding in Hercules remains binding authority governing an auditor’s duty of care in relation to a statutory audit, the framework by which that duty is imposed has since been refined. In the companion cases of Cooper and Edwards v. Law Society of Upper Canada , 2001 SCC 80 , [2001] 3 S.C.R. 562, this Court revised the Anns test by distinguishing more clearly between foreseeability and proximity, and by placing greater emphasis on a more demanding first stage of the two-stage analysis ( Cooper , at para. 30 ). While, therefore, we rely on Hercules for the general proposition that an auditor may owe its client a duty of care in relation to a particular undertaking, it is the Anns/Cooper framework to which we must have reference in identifying a principled basis for imposing liability. And, properly applied, that framework will rarely, if ever, give rise to a prima facie duty of care that could result in indeterminate liability. Accordingly, and with great respect for contrary views, there is no reason to resort to the second stage in order to negate all liability in this case.

(a) Stage One: Prima Facie Duty of Care

[ 23 ] In Cooper , this Court recognized that “foreseeability alone” is not enough to establish a prima facie duty of care (para. 22; see also Edwards , at para. 9 ). In doing so, it signalled a shift from the Anns test, which had grounded the recognition of a prima facie duty upon mere foreseeability of injury ( Hercules , at paras. 25 and 27; Norsk , at p. 1154; Bow Valley , at para. 61). After Cooper , the first stage of the Anns/Cooper framework would require “something more” ( Cooper , at para. 29 ). That “something more” is proximity ( Odhavji Estate v. Woodhouse , 2003 SCC 69 , [2003] 3 S.C.R. 263, at paras. 47-48 ; Childs v. Desormeaux , 2006 SCC 18 , [2006] 1 S.C.R. 643, at para. 12 ; Hill v. Hamilton-Wentworth Regional Police Services Board , 2007 SCC 41 , [2007] 3 S.C.R. 129, at para. 23 ; and Fullowka v. Pinkerton’s of Canada Ltd. , 2010 SCC 5 , [2010] 1 S.C.R. 132, at para. 18 ).

[ 24 ] In Cooper , the Court did not indicate whether proximity or reasonable foreseeability should be assessed first. In cases of negligent misrepresentation or performance of a service, however, proximity will be more usefully considered before foreseeability. What the defendant reasonably foresees as flowing from his or her negligence depends upon the characteristics of his or her relationship with the plaintiff, and specifically, in such cases, the purpose of the defendant’s undertaking. That said, both proximity and foreseeability of injury merit further reflection.

(i) Proximity

[ 25 ] Assessing proximity in the prima facie duty of care analysis entails asking whether the parties are in such a “close and direct” relationship that it would be “just and fair having regard to that relationship to impose a duty of care in law” ( Cooper , at paras. 32 and 34 ).

[ 26 ] Under the Anns test, proximity did not, “in and of itself, provide a principled basis on which to make a legal determination” ( Hercules , at para. 23). Rather, proximity was a “label” which expressed nothing more than a “result, judgment or conclusion” ( ibid. ), where mere reasonable foreseeability of injury could be shown. While, under the Anns/Cooper framework, the proximity analysis has become more analytically robust, this descriptive component remains. By this, we mean that the term “proximity” is still used, in part, as a shorthand description of those categories of relationships in which proximity has already been found to exist ( Cooper , at para. 23 ). If a relationship falls within a previously established category, or is analogous to one, then the requisite close and direct relationship is shown. So long, then, as a risk of reasonably foreseeable injury can also be shown — or has already been shown through an analogous precedent — the first stage of the Anns/Cooper framework is complete and a duty of care may be identified ( ibid. , at para. 36). In such circumstances, the second stage of the Anns/Cooper framework will seldom be engaged because any residual policy considerations will have already been taken into account when the proximate relationship was first identified ( ibid. , at para. 39; Edwards , at para. 10 ).

[ 27 ] This Court has on occasion defined previously established categories of proximity in broad terms. In Hill , for example, the Court listed “[t]he duty of care of the motorist to other users of the highway; the duty of care of the doctor to his patient; the duty of care of the solicitor to her client” (para. 25). Proximate relationships will not always, however, be identified so generally. In particular, whether proximity exists between two parties at large, or whether it inheres only for particular purposes or in relation to particular actions, will depend upon the nature of the particular relationship at issue ( ibid. , at para. 27; Haig , at p. 479). Indeed, and as we explain below, factors which support recognizing “novel” proximate relationships do so based upon the characteristics of the parties’ relationship and the circumstances of each particular case ( Cooper , at paras. 34-35 ).

[ 28 ] It follows that, where a party seeks to base a finding of proximity upon a previously established or analogous category, a court should be attentive to the particular factors which justified recognizing that prior category in order to determine whether the relationship at issue is, in fact, truly the same as or analogous to that which was previously recognized. And, by corollary, courts should avoid identifying established categories in an overly broad manner because, again, residual policy considerations are not considered where proximity is found on the basis of an established category ( Cooper , at para. 39 ). Analytically, this makes sense. For a court to have previously recognized a proximate relationship, second-stage residual policy considerations must already have been taken into account. When, therefore, a court relies on an established category of proximity, it follows “that there are no overriding policy considerations that would [negate] the duty of care” ( ibid. ). A consequence of this approach, however, is that a finding of proximity based upon a previously established or analogous category must be grounded not merely upon the identity of the parties, but upon examination of the particular relationship at issue in each case. Otherwise, courts risk recognizing prima facie duties of care without any examination of pertinent second-stage residual policy considerations.

[ 29 ] Where an established proximate relationship cannot be found, courts must undertake a full proximity analysis. To determine whether the “‘close and direct’ relationship which is the hallmark of the common law duty of care” exists ( Saadati v. Moorhead , 2017 SCC 28 , [2017] 1 S.C.R. 543, at para. 24 , citing Cooper , at para. 32 , and Donoghue v. Stevenson , [1932] A.C. 562 (H.L.), at pp. 580-81), courts must examine all relevant “factors arising from the relationship between the plaintiff and the defendant” ( Cooper , at para. 30 (emphasis in original); Edwards , at para. 9 ; Childs , at para. 24 ; Odhavji , at para. 50; Hill , at para. 24 ; Fullowka , at para. 26 ; Saadati , at para. 24 ). While these factors are diverse and depend on the circumstances of each case ( Cooper , at para. 35 ), this Court has maintained that they include “expectations, representations, reliance, and the property or other interests involved” ( ibid. , at para. 34; Odhavji , at para. 50; Fullowka , at para. 26 ) as well as any statutory obligations ( Cooper , at para. 38 ; Edwards , at paras. 9 and 13 ; Odhavji , at para. 56).

[ 30 ] In cases of pure economic loss arising from negligent misrepresentation or performance of a service, two factors are determinative in the proximity analysis: the defendant’s undertaking and the plaintiff’s reliance. Where the defendant undertakes to provide a representation or service in circumstances that invite the plaintiff’s reasonable reliance, the defendant becomes obligated to take reasonable care. And, the plaintiff has a right to rely on the defendant’s undertaking to do so (W. N. Hohfeld, “Some Fundamental Legal Conceptions as Applied in Judicial Reasoning” (1913), 23 Yale L.J. 16, at pp. 49-50). These corollary rights and obligations create a relationship of proximity ( Haig , at p. 477; Caparo Industries plc. v. Dickman , [1990] 1 All E.R. 568 (H.L.), at pp. 637-38 ; Glanzer v. Shepard , 135 N.E. 275 (N.Y. 1922) at pp. 275-76; Ultramares Corp. v. Touche , 174 N.E. 441 (N.Y. 1931), at pp. 445-46; E. J. Weinrib, “The Disintegration of Duty” (2006), 31 Adv. Q. 212, at p. 230).

[ 31 ] Rights, like duties, are, however, not limitless. Any reliance on the part of the plaintiff which falls outside of the scope of the defendant’s undertaking of responsibility — that is, of the purpose for which the representation was made or the service was undertaken — necessarily falls outside the scope of the proximate relationship and, therefore, of the defendant’s duty of care (Weinrib; A. Beever, Rediscovering the Law of Negligence (2007), at pp. 293-94). This principle, also referred to as the “end and aim” rule, properly limits liability on the basis that the defendant cannot be liable for a risk of injury against which he did not undertake to protect ( Glanzer , at pp. 275 and 277; Ultramares , at pp. 445-46; Haig , at p. 482). By assessing all relevant factors arising from the relationship between the parties, the proximity analysis not only determines the existence of a relationship of proximity, but also delineates the scope of the rights and duties which flow from that relationship. In short, it furnishes not only a “principled basis upon which to draw the line between those to whom the duty is owed and those to whom it is not” ( Fullowka , at para. 70 ), but also a principled delineation of the scope of such duty, based upon the purpose for which the defendant undertakes responsibility. As we will explain, these principled limits are essential to determining the type of injury that was a reasonably foreseeable consequence of the defendant’s negligence.

(ii) Reasonable Foreseeability

[ 32 ] Assessing reasonable foreseeability in the prima facie duty of care analysis entails asking whether an injury to the plaintiff was a reasonably foreseeable consequence of the defendant’s negligence ( Cooper , at para. 30 ).

[ 33 ] Broadly speaking, reasonable foreseeability concerns the likelihood of injury arising from the defendant’s negligence ( Donoghue , at p. 580). This inquiry is not amenable to, and does not require, actuarial precision. The jurisprudence gives content, however, to the foreseeability inquiry, providing courts with guidance. In the abstract, a defendant’s negligent misrepresentation or performance of a service could potentially give rise to innumerable injuries tangentially cascading from the originally contemplated service. This was so in Hercules , where the Court recognized that an auditor’s statement could be relied upon by a potentially limitless number of individuals (e.g., shareholders or takeover bidders), for a potentially limitless array of purposes (e.g., investments or takeover bids), any of which could result in various foreseeable injuries.

[ 34 ] As we have already observed, however, reasonable foreseeability of injury is no longer the sole consideration at the first stage of the Anns/Cooper framework. Since Cooper , both reasonable foreseeability and proximity — the latter expressed in Cooper as a distinct and more demanding hurdle than reasonable foreseeability — must be proven in order to establish a prima facie duty of care. And, in cases of negligent misrepresentation or performance of a service, the proximate relationship — grounded in the defendant’s undertaking and the plaintiff’s reliance — informs the foreseeability inquiry. Meaning, the purpose underlying that undertaking and that corresponding reliance limits the type of injury which could be reasonably foreseen to result from the defendant’s negligence.

[ 35 ] As a matter of first principles, it must be borne in mind that an injury to the plaintiff in this sort of case flows from the fact that he or she detrimentally relied on the defendant’s undertaking, whether it take the form of a representation or the performance of a service. It follows that an injury to the plaintiff will be reasonably foreseeable if (1) the defendant should have reasonably foreseen that the plaintiff would rely on his or her representation; and (2) such reliance would, in the particular circumstances of the case, be reasonable ( Hercules , at para. 27). Both the reasonableness and the reasonable foreseeability of the plaintiff’s reliance will be determined by the relationship of proximity between the parties; a plaintiff has a right to rely on a defendant to act with reasonable care for the particular purpose of the defendant’s undertaking, and his or her reliance on the defendant for that purpose is therefore both reasonable and reasonably foreseeable. But a plaintiff has no right to rely on a defendant for any other purpose, because such reliance would fall outside the scope of the defendant’s undertaking. As such, any consequent injury could not have been reasonably foreseeable.

[ 36 ] We add this. Under the Anns test, the Court recognized that auditors may owe a prima facie duty of care to an innumerable number of parties on the basis of reasonable foreseeability alone ( Hercules , at para. 32). We acknowledge that the Anns/Cooper framework, when applied to cases of negligent misrepresentation, will give rise to a far narrower scope of reasonably foreseeable injuries and, therefore, a narrower range of prima facie duties of care. This is no indictment of the Anns/ Cooper analysis. Rather, it was the very purpose and effect of this Court’s instruction in Cooper that “something more ” than mere foreseeability is required at the first stage of the Anns/Cooper framework. By requiring examination of the relationship between the parties as we have just discussed, Cooper gave Canadian courts a more complete array of legal tools to determine whether it is “just and fair” to impose a prima facie duty of care.

(b) Stage Two: Residual Policy Considerations

[ 37 ] Where a prima facie duty of care is recognized on the basis of proximity and reasonable foreseeability, the analysis advances to stage two of the Anns/Cooper framework. Here, the question is whether there are “residual policy considerations” outside the relationship of the parties that may negate the imposition of a duty of care ( Cooper , at para. 30 ; Edwards , at para. 10 ; Odhavji , at para. 51).

[ 38 ] By “residual”, we mean that such considerations “are not concerned with the relationship between the parties [already considered at stage one], but with the effect of recognizing a duty of care on other legal obligations, the legal system and society more generally” ( Cooper , at para. 37 ; see also Edwards , at para. 10 ). To the extent, therefore, that stage one of the prima facie duty of care is said to engage “policy” considerations arising from the relationship between the parties — i.e., the recognition that it is sound “policy” to only hold defendants liable for negligence when they are in a proximate relationship with the plaintiff and when the injury suffered was reasonably foreseeable (see Cooper , at para. 25 ) — such “policy” considerations are not revisited at stage two ( ibid. , at para. 28). Indeed such re consideration would be both redundant and analytically confusing ( ibid. , at para. 29).

[ 39 ] Cooper , and in particular, its strict delineation between “factors arising from the relationship [between the parties]” (para. 30 (emphasis in original)) and factors that “are not concerned with the relationship between the parties” (para. 37) has impacted the stage at which certain factors are considered within the Anns/Cooper framework. For example, principles that were traditionally considered at the second stage of the Anns test in cases of negligent misrepresentation, such as (1) whether the defendant knew the identity of the plaintiff or the class of plaintiffs who would rely on its representation; and (2) whether the reliance losses claimed by the plaintiff stem from the particular transaction in respect of which the statement at issue was made ( Hercules , at paras. 27 and 40; Bow Valley , at paras. 55-56), are no longer considered at the second stage. This is because, as we have explained, these factors arise from the relationship between the parties and are, therefore, properly accounted for under the first stage proximity and reasonable foreseeability analysis.

[ 40 ] What, then, remains to be considered at the second stage of the Anns/Cooper framework? In Cooper , this Court identified factors which are external to the relationship between the parties, including (1) whether the law already provides a remedy; (2) whether recognition of the duty of care creates “the spectre of unlimited liability to an unlimited class”; and (3) whether there are “other reasons of broad policy that suggest that the duty of care should not be recognized” (para. 37). In this way, the residual policy inquiry is a normative inquiry. It asks whether it would be better, for reasons relating to legal or doctrinal order, or reasons arising from other societal concerns, not to recognize a duty of care in a given case.

[ 41 ] The place within the Anns/Cooper framework of this policy inquiry is significant. It follows the proximity and foreseeability inquiries. The policy inquiry assesses whether, despite the proximate relationship between the parties, and despite the reasonably foreseeable quality of the plaintiff’s injury, the defendant should nonetheless be insulated from liability ( Cooper , at para. 30 ; Odhavji , at para. 51). That it would limit liability in the face of findings of both proximity and reasonable foreseeability makes plain how narrowly it should be relied upon ( Cooper , at para. 30 , citing Yuen Kun Yeu v. Attorney-General of Hong Kong , [1988] 1 A.C. 175 (P.C.) ; Edgeworth Construction Ltd. v. N.D. Lea & Associates Ltd. , [1993] 3 S.C.R. 206, at p. 218). Only in rare cases — such as those concerning decisions of governmental policy ( Cooper , at paras. 38 and 53 ) or quasi-judicial bodies ( ibid ., at para. 52; Edwards , at para. 19 ) — should liability be denied when a defendant’s negligence causes reasonably foreseeable injury to a plaintiff with whom he or she shares a close and direct relationship. In light of the above, the stage at which certain factors are considered in the Anns/Cooper framework is material.

[ 42 ] In this case, the Chief Justice finds that, if it were necessary to proceed to the second stage of the Anns/Cooper framework, she would insulate Deloitte from liability based on the residual policy consideration of indeterminacy (para. 166). We concede that indeterminate liability may, in some cases, be a legitimate residual policy consideration ( Cooper , at paras. 37 and 54 ; Hercules , at para. 31). In our view, however, rarely, if ever, should a concern for indeterminate liability persist after a properly applied proximity and foreseeability analysis ( Saadati , at para. 34 ; Fullowka , at para. 70 ). Robust application of stage one of the Anns/Cooper framework should almost always obviate concerns for indeterminate liability. This follows from an appreciation of what indeterminate liability, as a concept, actually means.

[ 43 ] Indeterminate liability is liability of a specific character , not of a specific amount . In particular, indeterminate liability should not be confused with significant liability ( Gross v. Great-West Life Assurance Co. , 2002 ABCA 37 , 299 A.R. 142, at para. 38 ). Certain activities — like flying commercial aircraft, manufacturing pharmaceutical drugs, or auditing a large corporation — may well give rise to significant liability. But such liability arises from the nature of the defendant’s undertakings and of the severe but reasonably foreseeable scale of injury that can result where such undertakings are negligently performed. This explains the significant compensation which these high risk undertakings typically attract. It also explains why contractual disclaimers limiting liability may often be warranted ( Edgeworth , at p. 220). In contrast, the liability arising from these “high risk” undertakings may only be characterized as “indeterminate” if the scope of such liability is impossible to ascertain ( Black’s Law Dictionary (10th ed. 2014), sub verbo “indeterminate”). In other words, liability is truly “indeterminate” if “the accepted sources of law and the accepted methods of working with those sources such as deduction and analogy — are insufficient to resolve the question” (M. V. Tushnet, “Defending the Indeterminacy Thesis”, in B. Bix, ed., Analyzing Law: New Essays in Legal Theory (1998), 223, at pp. 224-25). More specifically, there are three pertinent aspects to so-called “indeterminacy” in these cases: (1) value indeterminacy (“liability in an indeterminate amount”); (2) temporal indeterminacy (“liability . . . for an indeterminate time”); and (3) claimant indeterminacy (“liability . . . to an indeterminate class”): Hercules , at para. 31, citing Ultramares , at p. 444. Naturally, when a claim has value, temporal, and claimant indeterminacy, our legal tools are insufficient to resolve the quantum of infinite damages that will flow from such a claim.

[ 44 ] All this said, it would be very difficult for liability of an indeterminate character, so understood, to survive a robust analysis of proximity and foreseeability at the first stage of the Anns/Cooper framework. In cases of negligent misrepresentation or performance of a service, the requisite proximity analysis will address claimant indeterminacy because the class of claimants is determinate , including only those in respect of whom the defendant undertook to act. Likewise, foreseeability, which is constrained by the purpose of the undertaking in question, should address concerns about value indeterminacy, because the value of damages is limited — that is, determined — by the reasonably foreseeable quality of the injury ( Hercules , at para. 32). Finally, proximity and foreseeability should both address temporal indeterminacy since the longer the period of time over which injury is said to have occurred, the less likely the defendant undertook to protect against it and the less foreseeable the injury, taken as a whole. Hence Cardozo C.J.’s statement in the oft-cited Ultramares decision that a duty which gives rise to indeterminacy “enkindle[s] doubt whether a flaw may not exist in the implication of a duty that exposes to these consequences” (p. 444; see also Weinrib, at p. 231; Beever, at p. 275). In other words, a finding of indeterminate liability at the damages stage strongly suggests that a legal error occurred at the duty stage, since a finding of a prima facie duty of indeterminate scope underlies the resulting indeterminate liability.

[ 45 ] We would add one final point. Indeterminate liability is a residual policy consideration, nothing more. The presence of indeterminacy need not be dispositive of liability in all cases. To approach the analysis otherwise would transform indeterminate liability from a policy consideration into a policy veto . While indeterminacy may militate against liability, other policy considerations — such as the immense profit margins that “high risk” actors often benefit from, or the extent to which “high risk” actors voluntarily assume the risk of indeterminate liability — may ultimately justify maintaining that liability, despite its indeterminacy (Beever, at p. 293). Even, therefore, in the rare case where indeterminate liability survives the proximity and foreseeability inquiries, it is not automatic that such indeterminacy will necessarily govern ( Fullowka , at para. 70 ). Indeed, any so-called “indeterminate liability” which survives stage one of the Anns/Cooper framework presumably arises from the risk against which the defendant voluntarily undertook to protect the plaintiff and, therefore, may justly and fairly result in liability.

B. Application

[ 46 ] Having set out the proper legal framework for establishing liability in cases of pure economic loss arising from negligent misrepresentation or performance of a service, we turn now to apply that framework to the trial judge’s two findings of negligence in this case.

[ 47 ] In summary, at the first stage of the Anns / Cooper framework, a duty of care is established where proximity and reasonably foreseeability of injury are found. In our view, Deloitte’s undertakings in relation to soliciting investment, and the 1997 Audit, gave rise to proximate relationships. The purpose of those undertakings, in turn, determines the type of injury that was reasonably foreseeable as a result of Livent’s reliance. Livent relied on the 1997 Audit for the purpose it was provided. Thus, a resulting injury was reasonably foreseeable. The same cannot be said, however, in respect of Deloitte’s negligent assistance in soliciting investment.

[ 48 ] At the second stage of the Anns / Cooper framework, residual policy considerations may negate Deloitte’s duty of care. But none apply to the negligent provision of the 1997 Audit.

(1) Solicitation of Investment (August to October 1997)

(a) Prima Facie Duty of Care

(i) Proximity

[ 49 ] The proximity analysis first asks whether the relationship at issue falls within, or is analogous to, a previously recognized category of proximity ( Cooper , at para. 36 ; Edwards , at para. 9 ).

[ 50 ] In Hercules , this Court found that an auditor may be in a proximate relationship with its corporate client sufficient to give rise to a duty of care. That proximate relationship was not, however, between an auditor and its client at large. Rather, the recognized relationship was limited to the preparation of a statutory audit (para. 14).

[ 51 ] In this case, the asserted proximate relationship is not so narrow in scope. Livent claims that Deloitte owed it a duty of care in relation both to (1) the preparation of the 1997 Audit; and (2) the approval of the Press Release and preparation of the Comfort Letter. We see it as vital to the resolution of this case to distinguish between these two sets of documents.

[ 52 ] The mere fact that proximity has been recognized as existing between an auditor and its client for one purpose is insufficient to conclude that proximity exists between the same parties for all purposes. As discussed above, an overly broad characterization of an established category of proximity which fails to consider the scope of activity in respect of which proximity was previously recognized, risks a premature imposition of a prima facie duty of care. In our respectful view, this very error impairs the reasons of the trial judge and the Court of Appeal. This approach is fundamentally inconsistent with the framework set out by this Court in Cooper . For this reason, we do not agree that this Court has previously established a proximate relationship as between an auditor and its client for the purposes of soliciting investment. In these circumstances, we must undertake a full proximity analysis.

[ 53 ] As we have indicated above, the full proximity analysis in cases of negligent misrepresentation is focussed upon the purpose of the defendant’s undertaking and the plaintiff’s reliance. From August to October of 1997, the services which Deloitte provided to Livent — particularly its ongoing assistance in relation to the Press Release and the provision of the Comfort Letter — were undertaken for the purpose of helping Livent to solicit investment. Given this undertaking, Livent was entitled to rely upon Deloitte to carry out these services with reasonable care. From this, it follows that a relationship of proximity arose in respect of the content of Deloitte’s undertaking . Deloitte’s undertaking did not entitle Livent to rely on Deloitte’s services and representations for all possible purposes. Rather, the “close and direct” relationship which obligated Deloitte to act with reasonable care was limited to the purpose for which Deloitte undertook to act. In this regard, we agree with the Chief Justice that “[l]oss that results from [Livent’s] inability to attract investment . . . may fall within the scope of Deloitte’s duty of care”, though only in relation to the Press Release and Comfort Letter (para. 153).

(ii) Reasonable Foreseeability

[ 54 ] Having established a relationship of proximity for the purpose of soliciting investment, Livent asserts that the increase in its liquidation deficit beginning in the fall of 1997 was a reasonably foreseeable consequence of Deloitte’s negligence, because “[t]he false financial picture that ought not to have been certified by Deloitte” was relied upon by Livent to artificially extend its solvency (R.F., at para. 108). In other words, had Deloitte resigned rather than continued to assist Livent in soliciting investment, Livent would have known its actual finances and avoided their interim deterioration. In our view, however, this type of injury was not a reasonably foreseeable consequence of Deloitte’s negligent assistance in soliciting investment. This follows from our earlier observations about how the scope of the parties’ proximate relationship limits the type of injuries that are reasonably foreseeable.

[ 55 ] In cases of negligent misrepresentation or performance of a service, a plaintiff’s injury will be reasonably foreseeable where (1) the defendant should reasonably foresee that the plaintiff will rely on his or her representation; and (2) reliance by the plaintiff would, in the particular circumstances of the case, be reasonable ( Hercules , at para. 27). Whether reliance is reasonable and reasonably foreseeable will turn on whether the plaintiff had a right to rely on the defendant for that purpose . Here, Livent argues that it detrimentally relied on Deloitte’s services and representations to artificially extend the life of the corporation. This reliance is not, however, tied to the solicitation of investment, but was a matter of oversight of management. Phrased in terms of Deloitte’s undertaking, during the fall of 1997 Deloitte undertook to assist Livent in soliciting investment, not in oversight of management. Losses related to this undertaking — for example, an inability to solicit investment because of Deloitte’s negligence — may be recoverable from Deloitte. But losses outside the scope of this undertaking, including those claimed here relating to a lack of oversight of management extending Livent’s solvency, are not recoverable from Deloitte. Simply put, Deloitte never undertook, in preparing the Comfort Letter, to assist Livent’s shareholders in overseeing management; it cannot therefore be held liable for failing to take reasonable care to assist such oversight. And, given that Livent had no right to rely on Deloitte’s representations for a purpose other than that for which Deloitte undertook to act, Livent’s reliance was neither reasonable nor reasonably foreseeable. Consequently, the increase in Livent’s liquidation deficit which arose from its reliance on the Press Release and Comfort Letter was not a reasonably foreseeable injury.

[ 56 ] This is not to say that Livent had no resources for oversight at the time Deloitte assisted in soliciting investment. Indeed, for internal oversight purposes, Livent could reasonably rely on Deloitte’s 1996 Audit. Unlike the Comfort Letter, the 1996 Audit was prepared for the purpose of assisting shareholder oversight of management. As a consequence, its negligent preparation could result in reasonably foreseeable injury flowing from the shareholders’ inability to oversee management. The trial judge, however, made a finding of fact that any negligence in Deloitte’s preparation of the 1996 Audit caused no injury to Livent. As this finding has not been cross-appealed by Livent, we make no further comment on it.

(b) Residual Policy Considerations

[ 57 ] Having concluded that no prima facie duty of care arose in respect of Deloitte’s assistance in soliciting investment and the resulting increase in Livent’s liquidation deficit, there is no need to consider residual policy considerations.

(2) 1997 Clean Audit Opinion (April 1998)

(a) Prima Facie Duty of Care

(i) Proximity

[ 58 ] This Court has previously established that an auditor owes its corporate client a duty of care in the preparation of a statutory audit. It follows that the established proximate relationship in Hercules will be dispositive of the existence of a duty of care in this case, unless the purpose of Deloitte’s undertaking to prepare such an audit in this case can be distinguished from the undertaking in Hercules . As we will show, it cannot.

[ 59 ] In Hercules , at para. 48, this Court cited Lord Oliver’s statement in Caparo , at p. 583, identifying the purposes of a statutory audit:

It is the auditors’ function to ensure, so far as possible, that the financial information as to the company’s affairs prepared by the directors accurately reflects the company’s position in order first, to protect the company itself from the consequences of undetected errors or, possibly, wrongdoing . . . and, second, to provide shareholders with reliable intelligence for the purpose of enabling them to scrutinise the conduct of the company’s affairs and to exercise their collective powers to reward or control or remove those to whom that conduct has been confided. [Emphasis added; emphasis in original deleted.]

[ 60 ] These purposes, according to La Forest J., were no different under the statutory audit provisions in Manitoba’s Corporations Act , R.S.M. 1987, c. C225, which were at issue in Hercules . Regarding the second purpose, this Court stated that a statutory audit was necessary to “permit the shareholders, as a body, to make decisions as to the manner in which they want the corporation to be managed, to assess the performance of the directors and officers, and to decide whether or not they wish to retain the existing management or to have them replaced” ( Hercules , at para. 49). The purpose of the audited reports then “was, precisely, to assist the collectivity of shareholders of the audited companies in their task of overseeing management” ( ibid. ).

[ 61 ] No party before us has suggested that the purposes for which a statutory audit is prepared, and which have been recognized in Canadian law for 20 years, have changed. These purposes are consistent with the governing statute in this case: Ontario’s Business Corporations Act , R.S.O. 1990, c. B.16 (“ OBCA ”). In particular, ss. 153 and 154 of the OBCA require Deloitte, as Livent’s auditor, to examine Livent’s financial statements in order for Livent’s directors to fulfill their obligation to place a yearly statutory audit before its shareholders at the annual general meeting. And, while the engagement letters between Deloitte and Livent indicated that the detection of fraud was not guaranteed even where Deloitte acted with all reasonable care, they did not disclaim liability for negligently failing to uncover fraud. Thus, in our view, Deloitte did not alter the purpose for which it undertook to provide the 1997 Audit or disclaim liability in relation to that purpose.

[ 62 ] Given the foregoing, no basis exists for distinguishing the purpose of the statutory audit in this case from the purpose which underlay the statutory audit in Hercules . It follows that proximity is established between Livent and Deloitte in relation to the statutory audit, on the basis of the previously recognized proximate relationship identified by this Court.

(ii) Reasonable Foreseeability

[ 63 ] Livent says that the increase in its liquidation deficit was a reasonably foreseeable consequence of Deloitte’s negligent audit, because the audit preserved a false financial picture upon which Livent relied to artificially extend its solvency and delay filing for bankruptcy. In other words, if Deloitte had taken reasonable care in auditing Livent, then Livent would have discovered the fraud and avoided the interim deterioration of its assets.

[ 64 ] In our view, this type of injury was a reasonably foreseeable consequence of Deloitte’s negligent audit. The purpose of the 1997 Audit was, as this Court described in Hercules , two-fold: (1) to protect the company from the consequences of undetected errors and wrongdoing; and (2) to provide shareholders with reliable intelligence enabling oversight (para. 48, citing Caparo , at p. 583). Those purposes, as we have already described in our discussion of proximity generally, inform the scope of reasonably foreseeable injury. Specifically, at the time Deloitte undertook to provide the 1997 Audit, Livent was entitled to rely on Deloitte to take reasonable care in doing so for these recognized purposes. Livent’s reliance on Deloitte for the purpose of overseeing the conduct of management was therefore both reasonable and reasonably foreseeable. And, as Livent’s injury arises from its detrimental reliance, the injury linked to that reliance is itself reasonably foreseeable.

[ 65 ] It follows that the type of injury Livent suffered here was a reasonably foreseeable consequence of Deloitte’s negligence. Through the 1997 Audit, Deloitte undertook to assist Livent’s shareholders in scrutinizing management conduct. By negligently conducting the audit, and impairing Livent’s shareholders’ ability to oversee management, Deloitte exposed Livent to reasonably foreseeable risks, including “business losses” that would have been avoided with a proper audit. Indeed, the risk of injury flowing from undetected fraud is precisely the type of injury statutory audits seek to avoid.

[ 66 ] We add one final point in this regard. In Hercules (at para. 48), this Court cited Caparo for the proposition that statutory audits are conducted, in part, “to provide shareholders with reliable intelligence for the purpose of enabling them to scrutinise the conduct of the company’s affairs”. If subsequent business decisions that would not have survived such scrutiny do not fall within the scope of an auditor’s duty of care, one wonders what injury, if any, could result in liability for a negligent audit with respect to this recognized auditing purpose. Corporate scrutiny connotes both knowledge of problems within the corporation, and decisions reflecting an appreciation of those problems. Indeed, it is only by acting on the knowledge contained in an audit that is the product of reasonable care that corporation’s avoid losses that would have otherwise occurred without that audit.

(b) Residual Policy Considerations

[ 67 ] Having found a proximate relationship based on a previously recognized category, we need not consider residual policy considerations to negate or limit the scope of the duty of care ( Cooper , at para. 39 ). Nonetheless, as the Chief Justice finds, in the alternative, that the policy consideration of indeterminate liability would deny recovery in this case (paras. 165-166), it is useful to examine how the established proximate relationship engaged in this case precludes indeterminate liability.

[ 68 ] As discussed, the character of indeterminacy in these cases has three pertinent aspects: (1) temporal; (2) claimant; and (3) value ( Hercules , at para. 31, citing Ultramares , at p. 444). None of them arise here, consistent with our earlier observation that a robust application of the Anns/Cooper framework will usually, if not always, preclude the imposition of liability that is in any way indeterminate ( Saadati , at para. 34 ; Fullowka , at para. 70 ).

[ 69 ] Here, as to temporal indeterminacy, any suggestion that Livent could recover indefinitely from the negligent preparation of the 1997 Audit fundamentally mischaracterizes the scope of annual statutory audits. The injury flowing from the 1997 Audit could not be assessed over an indeterminate time window. Rather, statutory audits must occur annually ( OBCA , s. 154). As a result, the liability that could attach to one year’s negligent audit could not extend beyond the following year’s audit, which would effectively supersede the prior year’s audit as the factual and legal cause of the injury alleged. Put simply, the time window during which liability might flow from a single negligent statutory audit is not indeterminate. It is one year.

[ 70 ] Regarding claimant indeterminacy, the class of claimants here could not be further from indeterminate : it represents one single claimant — Livent. In Hercules , this Court noted that “audit reports will be relied on by many different people (e.g., shareholders, creditors, potential take-over bidders, investors, etc.)” (para. 32). That claim gave rise to indeterminate liability because the class of claimants (the “many different people”) was indeterminate. For example, any number of investors could rely on an audit to inform their investment decisions. This case, in contrast, is entirely distinguishable. The fact of a single potential claimant raises no concern of claimant indeterminacy.

[ 71 ] We note, parenthetically, that Deloitte characterizes Livent’s claim as, in reality (that is, in light of its insolvency), a claim by its various stakeholders. But this submission conflates the plaintiff, Livent, with the stakeholders who may benefit from the success of Livent’s claim, thereby disregarding Livent’s separate corporate personality. More importantly, it directly contradicts this Court’s holding in Hercules that a derivative action — which, too, could benefit various stakeholders — is the appropriate vehicle for a claim regarding a negligent statutory audit (paras. 1 and 58-64).

[ 72 ] The absence of temporal and claimant indeterminacy in turn explains the absence of value indeterminacy in this case. Here, Livent’s improvident use of investment funds could not result in liability of an indeterminate value. Rather, the liability in this case could not exceed the losses of a single corporation. When undertaking to audit Livent, Deloitte must have known that Livent was a substantial corporation, and in turn, that it could suffer large financial losses if misinformed by its auditor. But significant liability is distinct from indeterminate liability ( Gross , at para. 38 ). Put differently, Deloitte was, indeed, “able to gauge the scale of its potential liability” before undertaking the 1997 Audit (Chief Justice’s reasons, at para. 176). This is a far cry from the limitless potential quantum of lost investments by innumerable third parties relying on audit statements for their own investment decisions (see Hercules , at para. 32). The concern that Deloitte did not know “the scope of [its] liability at the time [it took] on [its] engagement” with Livent (Chief Justice’s reasons, at para. 176) conflates indeterminate liability with undetermined liability.

[ 73 ] The Chief Justice describes the liability sought to be imposed here as “indeterminate” because Livent’s reliance purportedly fell outside the scope of Deloitte’s undertaking (para. 170). We disagree. To the contrary, value indeterminacy is limited by the purposes for which the audit was prepared, and Livent’s reliance fell squarely within that purpose. In Hercules , this Court rejected a claim by investors because they might use audit reports for a “collateral or unintended purpose” (para. 38), thereby giving rise to indeterminate liability (since the variety of purposes to which an audit may be put is potentially limitless). But that is not the case here. The 1997 Audit was prepared for the express purpose of oversight of management by Livent’s shareholders, and the loss at issue flowed from those shareholders’ inability to conduct that oversight. It follows that the purposes underlying the 1997 Audit — of which, as we have explained, there are only two — do not give rise to potential indeterminacy, and by corollary, relate to potential losses that, too, are not indeterminate. This is not a case where, for example, an unknown third-party relied on an audit to launch a takeover bid — a purpose outside the scope of the audit ( Hercules , at para. 32). Rather, this is a case in which an established purpose of the audit was undermined, and where losses predictably flowed from that failed purpose ( Haig , at pp. 478-79).

[ 74 ] Here, one claimant (Livent) is suing Deloitte for failing to satisfy one of two auditing purposes (oversight of management) which would have been superseded by an audit one year later. No indeterminate liability arises in this context. In this regard, La Forest J.’s remarks for this Court in Hercules (at para. 37) are apposite:

. . . in cases where the defendant knows the identity of the plaintiff (or of a class of plaintiffs) and where the defendant’s statements are used for the specific purpose or transaction for which they were made, policy considerations surrounding indeterminate liability will not be of any concern since the scope of liability can readily be circumscribed.

[ 75 ] The lack of indeterminacy here between Deloitte (an auditor) and Livent (its corporate client) is unsurprising given (1) this Court’s recognition in Hercules that a duty of care exists between an auditor and its corporate client in relation to a statutory audit; and (2) this Court’s direction in Cooper that the second stage of the Anns/Cooper framework need not be considered where a previously recognized proximate relationship exists.

(c) Remoteness

[ 76 ] The Chief Justice says that Deloitte’s complete immunity from liability would similarly flow from a remoteness analysis (para. 173). We disagree.

[ 77 ] In a successful negligence action, a plaintiff must demonstrate that (1) the defendant owed him or her a duty of care; (2) the defendant’s behaviour breached the standard of care; (3) the plaintiff sustained damage; and (4) the damage was caused, in fact and in law, by the defendant’s breach ( Mustapha v. Culligan of Canada Ltd. , 2008 SCC 27 , [2008] 2 S.C.R. 114, at para. 3 ; Saadati , at para. 13 ). The principle of remoteness, or legal causation, examines whether “the harm [is] too unrelated to the wrongful conduct to hold the defendant fairly liable” ( Mustapha , at para. 12 , citing A.M. Linden and B. Feldthusen, Canadian Tort Law (8th ed. 2006), at p. 360; see also Saadati , at para. 34 ). It is trite law that “it is the foresight of the reasonable man which alone can determine responsibility” ( Mustapha , at paras. 11-13 , citing Overseas Tankship (U.K.) Ltd. v. Morts Dock & Engineering Co. , [1961] A.C. 388 (P.C.), at p. 424) . Therefore, injury will be sufficiently related to the wrongful conduct if it is a reasonably foreseeable consequence of that conduct.

[ 78 ] We acknowledge that remoteness, so understood, overlaps conceptually with the reasonable foreseeability analysis conducted in the prima facie duty of care analysis ( Mustapha , at para. 15 ). But the two are distinct: the duty analysis is concerned with the type of injury that is reasonably foreseeable as flowing from the defendant’s conduct, whereas the remoteness analysis is concerned with the reasonable foreseeability of the actual injury suffered by the plaintiff (L. N. Klar and C. S. G. Jefferies, Tort Law , (6th ed. 2017), at p. 565: “Remoteness questions deal with how far liability should extend in reference to injuries caused to the plaintiff , once a duty relationship . . .[has] been established” (emphasis added)).

[ 79 ] Remoteness, at its core, turns on the reasonable foreseeability of the actual injury suffered by the plaintiff. But, and as we have explained, the loss here — stemming from Deloitte’s failure to fulfill the specific undertaking it made to Livent — was reasonably foreseeable. It follows that remoteness is not a bar to Livent’s recovery.

[ 80 ] Nonetheless, the Chief Justice holds that Livent’s loss is too remote because it cannot be attributed to its shareholders’ reliance on the 1997 Audit for the purpose of overseeing management. Specifically, she says that “Livent did not prove and the trial judge did not find that Livent’s shareholders relied on Deloitte’s negligent audit statements, or that had they received and relied on accurate statements, they would have acted in a way that would have prevented Livent from carrying on business and diminishing its assets” (para. 159). With respect, we see the matter differently. In its amended statement of claim, Livent advanced its theory of impaired shareholder reliance (A.R., vol. III, at p. 112):

As a consequence of the Auditors’ breaches of duty, they missed repeated opportunities to uncover and reveal the accounting irregularities and errors being orchestrated by Drabinsky and Gottlieb. Consequently, the Livent Stakeholders were deprived of the opportunity to exercise their collective will by, inter alia, ousting Drabinsky and Gottlieb thereby avoiding further losses, damages and liabilities incurred by Livent and the Livent Stakeholders. [Emphasis added.]

[ 81 ] Similarly, when the trial judge summarized Livent’s position at trial, he wrote (at para. 23):

[Deloitte’s] alleged negligent issuance of unqualified opinions, in turn, deprived the honest directors and shareholders of the opportunity to put a stop to the fraud, and the losses eventually caused to the company by the fraud, at an earlier date. [Emphasis added.]

[ 82 ] The trial judge accepted this theory: “I believe that the honest directors and innocent shareholders in this case were entitled to rely on Deloitte’s audits to discharge their supervisory task” (para. 341). Nonetheless, the Chief Justice would deny liability because, in her view, “Livent offered no proof to support” the assertion that its shareholders would have called management to account had they received a non-negligent audit in March of 1998 (para. 161). But this is precisely what the record shows. On November 18, 1998, Livent received a prudently prepared audit of its restated 1997 financial statements. This prudent audit disclosed a “significant, if not staggering” difference in reported income (trial reasons, at para. 15). Specifically, the prudent audit uncovered an additional loss of over $50 million during the 1997 fiscal year.

[ 83 ] Livent’s response upon receipt of this audit report is telling and, in our view, belies any suggestion that informed shareholder scrutiny would have permitted Livent to act in any manner other than expected. That same day , “Drabinsky and Gottlieb were dismissed for cause . . . and Livent voluntarily made a petition for bankruptcy protection” in the United States (trial reasons, at para. 16). The next day , “Livent filed for protection under the Companies’ Creditors Arrangement Act, R.S.C. 1985, c. C-36 in Canada” ( ibid. ). It is difficult to conceive of a clearer demonstration of when and how Livent’s shareholders would have “prevented Livent from carrying on business and diminishing its assets” had Deloitte prepared a prudent audit in March of 1998 (Chief Justice’s reasons, at para. 159).

[ 84 ] On this record, any speculation that Livent’s shareholders might have done nothing in response to rampant fraud is simply unsustainable. Indeed, the similarly speculative claim that Livent’s shareholders might have done nothing when confronted with a notional liquidation deficit of approximately $365 million in early April 1998 (when we know that the actual liquidation deficit of $413.83 million in mid-November 1998 prompted Livent’s immediate petition into bankruptcy) is — also in light of this record — equally improbable.

[ 85 ] The Chief Justice, however, advances one additional basis upon which to deny Deloitte’s liability: that Deloitte should be liable “only for the information” it provided (the audit opinion), not “for the decision[s] to be informed by it” (para. 171). This proposition does not preclude Deloitte’s liability here.

(d) Additional Basis for Limiting Liability: Information, Advice and the “SAAMCO Principle”

[ 86 ] The Chief Justice seeks to limit Deloitte’s liability because it merely provided “information” to Livent, not “advice”, and, as a consequence, did not “assume responsibility for what the shareholders decide[d] to do with that information” (para. 170). In this regard, she cites (at para. 149) the following passage from Hughes-Holland v. BPE Solicitors , [2017] UKSC 21, [2017] 2 W.L.R. 1029, at para. 44 :

A valuer or a conveyancer, for example, will rarely supply more than a specific part of the material on which his client’s decision is based. He is generally no more than a provider of what Lord Hoffmann [in SAAMCO ] called “information”. At the opposite end of the spectrum, an investment adviser advising a client whether to buy a particular stock, or a financial adviser advising whether to invest self-invested pension fund in an annuity are likely, in Lord Hoffmann’s terminology, to be regarded as giving “advice”. Between these extremes, every case is likely to depend on the range of matters for which the defendant assumed responsibility and no more exact rule can be stated.

[ 87 ] It is true that Deloitte, as auditor, did not advise Livent on its business decisions. But it nevertheless “assumed responsibility” over providing accurate information upon which the shareholders could rely to scrutinize management conduct. Deloitte does not escape liability simply because a negligent audit, in itself, cannot cause financial harm. Audits never, in themselves, cause harm. It is only when they are detrimentally relied upon that tangible consequences ensue.

[ 88 ] The consequences of this line of reasoning should not be understated. By effectively limiting an auditor’s liability to the harms inherent in the negligent audit, while excluding those harms which are a reasonably foreseeable consequence of that negligent audit (i.e., an inability to oversee management), a central purpose for which a statutory audit is prepared is undercut, thereby immunizing auditors from liability for any act of negligence impairing oversight. This holding is inconsistent with the Court’s jurisprudence (which prescribes recovery for reasonably foreseeable injury, even if that injury is not technically immediate) and is in tension with Hercules (which sought to limit auditors’ liability for statutory audits to determinate corporate claims rather than indeterminate stakeholder claims, not to insulate auditors from virtually all liability).

[ 89 ] The Chief Justice’s reliance on the so-called “ SAAMCO principle” — derived from the House of Lords’ decision in South Australia Asset Management Corp. v. York Montague Ltd. , [1997] A.C. 191 (“ SAAMCO ”) ; see also Nykredit Mortgage Bank plc. v. Edward Erdman Group Ltd. (No. 2) , [1997] 1 W.L.R. 1627 (H.L.) ; Platform Home Loans Ltd. v. Oyston Shipways Ltd. , [2000] 2 A.C. 190 ) (H.L.); and BPE Solicitors — is in our respectful view mistaken for the same reason as her reliance on the dichotomy between information and advice as described in BPE Solicitors . Her reliance on both U.K. decisions is premised on how Deloitte never assumed responsibility for injuries resulting from Livent’s operations — either because those injuries were not caused by the mere information contained in the 1997 Audit ( BPE Solicitors ), or because those injuries did not flow from what was incorrect in the 1997 Audit ( SAAMCO ). But, in assuming responsibility for informing shareholder scrutiny of management, Deloitte did assume responsibility for injuries flowing from that impaired scrutiny.

[ 90 ] In simple terms, the SAAMCO principle denies recovery for pure economic loss where the plaintiff’s injury would still have occurred even if the defendant’s negligent misrepresentation were factually true. Rephrased as a test, the principle denies liability where an alternate cause that is unrelated to the defendant’s negligence is the true source of the plaintiff’s injury. This alternate and unrelated cause explains why the truth of the negligent misstatement has no bearing on the plaintiff’s ultimate injury (i.e., because, even with that truth, the injury would have flowed as a result of the alternate cause). Or, framed from the perspective of the duty of care, the defendant could not have undertaken to protect against injuries that would have been caused by alternate and unrelated sources. In SAAMCO , the House of Lords explained the principle with the commendably Albertan example of a mountaineer:

A mountaineer about to undertake a difficult climb is concerned about the fitness of his knee. He goes to a doctor who negligently makes a superficial examination and pronounces the knee fit. The climber goes on the expedition, which he would not have undertaken if the doctor had told him the true state of his knee. He suffers an injury which is an entirely foreseeable consequence of mountaineering but has nothing to do with his knee. [p. 213]

[ 91 ] In this example, the doctor’s negligent misrepresentation (the positive knee diagnosis) is a cause that is alternate and unrelated to the cause of the mountaineer’s injury (a mountaineering accident unrelated to the knee, for example, an avalanche). As a result, even had the doctor’s negligent misrepresentation been true (i.e., even if the mountaineer’s knee had been fit), the injury would still have occurred, since the fitness of his knee would not have prevented the injury caused by the avalanche. In other words, the doctor could not have undertaken to protect against an avalanche, which is unrelated to his or her diagnosis.

[ 92 ] Deloitte is unlike the doctor. Deloitte’s negligence related to a statutory audit, a purpose of which is management oversight by shareholders. That oversight, in turn, informs (or is related to) subsequent business decisions by the corporation. It follows that Livent’s trading losses were not an alternate and unrelated cause of Livent’s injury. To the contrary, the shareholders’ capacity to oversee the conduct of Livent’s business was entirely dependent upon the statutory audit preceding that oversight. In particular, the shareholders’ reliance on that audit and the audit’s portrayal of the directors and their business ventures was a critical component of their oversight of management — which, we reiterate, was the very purpose in respect of which Deloitte undertook to act with reasonable care.

[ 93 ] It therefore follows from a proper understanding of the SAAMCO principle that it does not limit Deloitte’s liability in respect of the 1997 Audit.

[ 94 ] We add, however, that a full consideration of SAAMCO ’s application in Canadian law by this Court should await future cases, with greater consideration of the principle by lower courts, more comprehensive submissions by counsel, and critically, with facts more analogous to those in the SAAMCO jurisprudence. SAAMCO addressed whether a negligent valuator should be liable, not only for the difference between their valuation and a prudent valuation, but also for a subsequent drop in the market which exacerbated the lenders’ losses. In this specific context, the loss attributable to the negligent valuation is easily extricable; it is the difference between the negligent valuation and a prudent valuation. But the same cannot be said in the context of statutory audits, where the “cause” of future losses flowing from future commercial activity is far more complex to isolate. Indeed, as Lord Sumption noted in BPE Solicitors (at para. 46):

Where the loss arises from a variety of commercial factors which it was for the claimant to identify and assess, it will commonly be difficult or impossible as well as unnecessary to quantify and strip out the financial impact of each one of them.

[ 95 ] In any event, the SAAMCO principle, at least in the manner the Chief Justice applies it here, conflicts with Canadian jurisprudence. Under established Canadian tort law, a defendant is liable if the plaintiff proves — in respect of causation — that the defendant caused the plaintiff’s injury in fact ( Clements v. Clements , 2012 SCC 32 , [2012] 2 S.C.R. 181, at para. 8 ) and in law ( Mustapha , at paras. 12-13 ). As we have already explained, Livent proved both in respect of its injuries after the 1997 Audit. It follows that Deloitte is liable for Livent’s injuries following that audit. Admittedly, a defendant may limit its liability, even if the plaintiff proves legal and factual causation, where the defendant proves that some of the plaintiff’s injury would have still occurred without the defendant’s negligence because of alternate hypothetical causes ( Rainbow Industrial Caterers Ltd. v. Canadian National Railway Co. , [1991] 3 S.C.R. 3, at pp. 15-16). But those alternate hypothetical causes were already accounted for in the trial judge’s 25 percent reduction in damages (trial reasons, at paras. 324-26). And, in so far as the Chief Justice holds that Livent failed to prove that alternate hypothetical causes did not cause the remaining 75 percent of damages, she disregards that it is Deloitte, not Livent, who bears the burden of proving that liability should be limited in this fashion ( Rainbow Industrial Caterers , at pp. 15-16).

C. Defences

[ 96 ] Finally, having concluded that we would uphold the trial judge’s finding that Deloitte is liable for its negligence in relation to the statutory audit, we must consider the two defences Deloitte advanced before this Court. First, Deloitte submits that both lower courts erred in failing to find that Livent’s action was not barred by the defence of illegality. Secondly, Deloitte submits that, even had Livent’s action not been barred for illegality, Deloitte should only be liable for part of the injury due to Livent’s contributory fault.

[ 97 ] Both defences rely on the applicability of the doctrine of corporate identification. As the wrongdoing at issue was not that of Livent’s but of its directors Drabinsky and Gottlieb, Deloitte cannot succeed on either defence unless it can show that the fraudulent acts of Livent’s employees should be attributed to the corporation. The corporate identification doctrine is not, however, a standalone principle; rather, it is a means by which acts may be attributed to a corporation for the particular purpose or defence at issue. It follows that corporate identification must be analyzed independently for each defence.

(1) Illegality

[ 98 ] The defence of illegality bars an otherwise valid action in tort on the basis that the plaintiff has engaged in illegal or immoral conduct and, therefore, should not recover ( Hall v. Hebert , [1993] 2 S.C.R. 159, at p. 169; British Columbia v. Zastowny , 2008 SCC 4 , [2008] 1 S.C.R. 27, at para. 20 ). Grounded in public policy, it is available in very “limited” circumstances, only where it is necessary to preserve the “integrity of the justice system” ( Hall , at pp. 179-80). And, the integrity of the justice system will only be compromised where a “damage award in a civil suit would, in effect, allow a person to profit from illegal or wrongful conduct, or would permit an evasion or rebate of a penalty prescribed by the criminal law” ( Hall , at p. 169; Zastowny , at para. 3 ).

[ 99 ] Here, the only illegal or wrongful conduct was committed by Livent’s directors, Drabinsky and Gottlieb, and portions of management. It follows that, for Deloitte to rely on the defence of illegality, it must be able to attribute the “illegal or wrongful conduct” of certain directors and managers to Livent itself, the plaintiff in this case.

[ 100 ] The test for corporate attribution was set out by this Court in Canadian Dredge & Dock Co. v. The Queen , [1985] 1 S.C.R. 662. To attribute the fraudulent acts of an employee to its corporate employer, two conditions must be met: (1) the wrongdoer must be the directing mind of the corporation; and (2) the wrongful actions of the directing mind must have been done within the scope of his or her authority; that is, his or her actions must be performed within the sector of corporate operation assigned to him. For the purposes of this analysis, an individual will cease to be a directing mind unless the action (1) was not totally in fraud of the corporation; and (2) was by design or result partly for the benefit of the corporation (pp. 681-82 and 712-13).

[ 101 ] At first glance, these conditions might seem to be satisfied in this case. Drabinsky and Gottlieb were directing minds, acting within the sector of corporate operations assigned to them, whose fraud was genuinely designed and executed in an attempt to assist Livent through the artificial extension of its life. Indeed, the application of the doctrine in this case would be consistent with the factually analogous decision in Hart Building Supplies Ltd. v. Deloitte & Touche , 2004 BCSC 55 , 41 C.C.L.T. (3d) 240. There, the test laid down by this Court in Canadian Dredge was strictly applied in the context of a civil claim for auditor’s negligence and was satisfied. The court attributed the fraudulent acts of the “alter ego and directing mind” to the auditor’s corporate client and barred recovery for the auditor’s negligent preparation of a statutory audit.

[ 102 ] In our view, however, a strict application of this Court’s decision in Canadian Dredge was not warranted in Hart , and is not warranted here. It must be recalled that Canadian Dredge was decided in the context of criminal liability. Accordingly, the underlying question there was who should bear the responsibility for the criminal actions of a corporation’s directing mind. Consequently, the policy factors identified therein which weigh in favour of imputing a corporation with the illegality or wrongdoing of its directing mind flow from the “social purpose” of holding a corporation responsible for the criminal acts of its employees where those acts are designed and carried out, at least in part, to benefit the corporation ( Canadian Dredge , at p. 704).

[ 103 ] However, as Estey J. himself recognized, the doctrine is only one of “judicial necessity” and where its application “would not provide protection of any interest in the community” or “would not advantage society by advancing law and order”, the rationale for its application “fades away” ( Canadian Dredge , at pp. 707-8 and 718-19). While public policy and judicial necessity may favour imputing the corporation with the actions of its directing minds in certain criminal prosecutions, the same cannot be said of attributing the actions of a directing mind for the purposes of a civil suit in the context of an auditor’s negligent preparation of a statutory audit. As indicated above, the very purpose of a statutory audit is to provide a means by which fraud and wrongdoing may be discovered. It follows that denying liability on the basis that an individual within the corporation has engaged in the very action that the auditor was enlisted to protect against would render the statutory audit meaningless (D. L. MacPherson, “Emaciating the Statutory Audit — A Comment on Hart Building Supplies Ltd. v. Deloitte & Touche ” (2005), 41 Can. Bus. L.J. 471). As Livent submitted, it would be perverse to deny auditor’s liability for negligently failing to detect fraud “where the harm [to the corporation] is likely to occur and likely to be most serious” (R.F., at para. 94).

[ 104 ] While, therefore, this Court’s decision in Canadian Dredge remains the authoritative test for the application of the corporate identification doctrine, we would reaffirm one qualification. The principles set out in Canadian Dredge provide a sufficient basis to find that the actions of a directing mind be attributed to a corporation, not a necessary one (pp. 681-82). As a principle that is grounded in policy, and which only serves as a means to hold a corporation criminally responsible or to deny civil liability, courts retain the discretion to refrain from applying it where, in the circumstances of the case, it would not be in the public interest to do so. And where, as here, its application would render meaningless the very purpose for which a duty of care was recognized, such application will rarely be in the public interest. If a professional undertakes to provide a service to detect wrongdoing, the existence of that wrongdoing will not normally weigh in favour of barring civil liability for negligence through the corporate identification doctrine. (That said, we leave for another day the question of whether the same is true in the context of a “one man” corporation, where the sole director and shareholder hires the auditor to uncover the wrongful action which he, himself, carries out: Stone & Rolls Ltd. (in liquidation) v. Moore Stephens , [2009] UKHL 39 , [2009] 1 A.C. 1391 ; see also 373409 Alberta Ltd. (Receiver of) v. Bank of Montreal , 2002 SCC 81 , [2002] 4 S.C.R. 312, at para. 22 ; Bilta (U.K.) Ltd. (in liquidation) v. Nazim (No. 2) , [2015] UKSC 23 , [2016] A.C. 1 , at para. 30 . )

[ 105 ] Finally, given the limited application of the defence of illegality, as recognized by this Court in Hall and Zastowny , we find no further compelling reason to justify the use of the corporate identification doctrine in these circumstances.

(2) Contributory Fault

[ 106 ] In the alternative, Deloitte submits that the Court of Appeal erred in holding Deloitte liable for the entirety of the proven loss, and specifically that Livent should have been found contributorily at fault in accordance with s. 3 of the Negligence Act , R.S.O. 1990, c. N.1:

In any action for damages that is founded upon the fault or negligence of the defendant if fault or negligence is found on the part of the plaintiff that contributed to the damages, the court shall apportion the damages in proportion to the degree of fault or negligence found against the parties respectively.

[ 107 ] Again, the only conduct implicating Livent in this case was committed by Livent’s directors, Drabinsky and Gottlieb, and portions of management. It follows that, for Deloitte to rely on the defence of contributory fault, it must be able to attribute the conduct of certain directors and managers to Livent itself.

[ 108 ] Unlike the discretionary illegality doctrine, Deloitte notes that s. 3 of the Negligence Act is mandatory (“. . . the court shall apportion the damages . . .”), and argues that corporate identification must be permitted because “the application of the corporate identification doctrine must be tailored to the terms of the particular substantive rule it serves” (A.F., at para. 132, citing C.A. reasons, at para. 157). This argument, however, misunderstands what the Negligence Act actually makes mandatory. The Negligence Act requires that a plaintiff’s fault be factored into the apportionment of damages. But corporate identification is a prerequisite to the plaintiff, Livent, being at fault. In other words, Deloitte’s claim that s. 3 of the Negligence Act requires that Livent bear its share of the fault presupposes corporate identification, in effect, putting the cart before the horse. The Negligence Act only makes contribution by a negligent plaintiff mandatory; it does not make attribution of negligence to a plaintiff mandatory.

[ 109 ] In any event, we repeat our earlier conclusion that where, as here, the use of the corporate identification doctrine would undermine the very purpose of establishing a duty of care, it will rarely be in the public interest to apply it. A negligent auditor cannot limit liability for its own negligence by attributing to the corporation the wrongful acts of its employees, such acts being the very conduct that the auditor undertook to uncover. Additionally, had Deloitte sought to limit its liability through apportionment, it need not have relied on the doctrine of corporate identification at all. Specifically, Deloitte could have brought third party claims against the guilty parties, Drabinsky and Gottlieb, for their wrongful actions. For whatever reason, it chose not to do so. Nonetheless, the availability of a third party claim against a fraudulent director weighs against the application of the doctrine. In this case, it is not in the public interest to undermine separate legal personality where the wrongdoer could have been properly named as a third party.

D. Conclusion

[ 110 ] It follows from the foregoing that we would allow the appeal, but only in part.

[ 111 ] In our view, the trial judge and Court of Appeal erred in finding that Deloitte’s negligence in relation to the Press Release and Comfort Letter resulted in injuries that were reasonably foreseeable in light of the proximate relationship between the parties. At that time, Deloitte’s services were engaged for the purpose of soliciting investment, not management oversight. As Livent’s losses did not flow from a failure to solicit investment, we would deny recovery for the increase in Livent’s liquidation deficit beginning in the fall of 1997.

[ 112 ] We would, however, allow recovery for the increase in Livent’s liquidation deficit which followed the 1997 Audit. We agree with the trial judge that “Deloitte should not have signed off on the 1997 Audit in early April 1998” (para. 242) and that the increase in Livent’s liquidation deficit which followed fell within the duty of care owed by Deloitte to Livent in relation to the preparation of a statutory audit, the express purpose of which was to assist Livent in management oversight.

[ 113 ] The trial judge assessed Livent’s damages following the 1997 Audit at $53.9 million (para. 306, fn. 188, and para. 369, fn. 228). Applying the trial judge’s 25 percent contingency reduction to this amount results in a final damages assessment of $40,425,000. This is the amount for which Deloitte is liable.

[ 114 ] Throughout these proceedings, the parties have primarily framed this dispute as one in negligence. Indeed, this was noted by the trial judge (at para. 47). At trial, Livent conceded that its losses for negligent performance of a service or breach of contract would be identical ( ibid. ). The trial judge agreed, finding that Livent’s claim in contract “succeed[ed] . . . for the [same] reasons” as its claim of negligent performance of a service and that the elements of its claim in contract were “incorporated by reference to the finding of ‘negligence’” (para. 243). Given the above, we would impose the same quantum of liability on Deloitte for the concurrent claim in breach of contract.

[ 115 ] Accordingly, the appeal is allowed in part. The amount of the trial award is reduced from $84,750,000 to $40,425,000. Livent shall have its costs throughout.

The reasons of McLachlin C.J. and Wagner and Côté JJ. were delivered by

The Chief Justice —

[ 116 ] Garth Drabinsky and Myron Gottlieb built a North American theatre empire that came to be known as Livent Inc. Seeking ever more spectacular success, they resorted to manipulating the company’s financial records. When the scheme came to light, Livent collapsed. Its assets were liquidated. Drabinsky and Gottlieb went to jail.

[ 117 ] This is among the many lawsuits that followed. The courts below held that Deloitte & Touche, a firm of accountants and Livent’s auditor, breached the duty of care it owed to Livent in failing to detect and expose Livent’s fraud, as a result of which Livent was able to continue operations and continue losing money — money it now claims from Deloitte.

[ 118 ] I agree with the courts below that Deloitte owed a duty of care to Livent, which it breached when it failed to discover and expose Livent’s fraud in the audited statements it prepared. However, I do not agree that Deloitte is liable for virtually all the loss that befell Livent as it pursued its precipitous decline into insolvency through doomed investments. I would allow the appeal.

I. Facts

[ 119 ] The saga that led to these proceedings began in 1989 when two would-be entertainment moguls, Drabinsky and Gottlieb, launched a takeover bid of their employer, Cineplex Odeon Corporation. When the bid failed, the pair formed MyGar Partnership, which purchased all of the assets and some of the liabilities of Cineplex’s live entertainment division. These included Toronto’s Pantages Theatre and the rights to a wildly successful show, The Phantom of the Opera . MyGar carried on business through its nominee corporation, Live Entertainment Corporation of Canada Inc., and both entities were rolled into Live Entertainment of Canada Inc., or Livent Inc., in 1993. Livent made its debut in Canada’s equity market with an initial public offering that year. By the end of 1995, Livent’s shares were also listed on New York’s NASDAQ exchange.

[ 120 ] Deloitte & Touche (now Deloitte LLP) became MyGar’s auditor in 1989 and continued as auditors for MyGar and its successor Livent until 1998.

[ 121 ] Livent’s strategy was vertical integration. Unlike other players in the live entertainment industry, it brought the entire enterprise, from production to performance, under one roof — a roof that Livent, as a proprietor of theatre properties, also owned. This was an immensely costly and risky undertaking. Livent invested in real estate in Toronto, Vancouver, New York, and Chicago. It produced and presented a string of spectacular (and therefore expensive) musicals. When its shows succeeded, Livent reaped all the rewards. When they did not, it bore the losses alone.

[ 122 ] Drabinsky and Gottlieb were determined to prove that their business model worked. To be sure it did, they and their associates cooked the books. They did so in four ways:

i. Drabinsky and Gottlieb pocketed some $7.5 million in kickbacks in the two years before Livent’s initial public offering. These they secured by causing MyGar to pay false or inflated invoices to various contractors, who in turn directed funds to Drabinsky and Gottlieb personally or to another company they controlled. A substantial portion of the falsified expenses were booked as assets. Drabinsky and Gottlieb’s balance sheets were thus infused with fantasy as early as 1991.

ii. Livent officers and employees distorted the company’s financial records by shifting expenses between accounting periods and from one activity or production to another. They doctored Livent’s accounting software to permit and conceal these manipulations.

iii. Livent overstated its bottom line by extending the time over which it amortized the costs of putting on its productions and, in some periods, by avoiding amortization altogether. This it did by transferring millions of dollars of “pre-production costs” from shows to fixed assets or from one show to another, and particularly to shows that had not yet opened.

iv. The company recorded imaginary revenue by entering into loan or financing agreements that it camouflaged as asset sales. Livent purported to sell various rights related to its productions and properties, but simultaneously entered into secret side agreements permitting purchasers to recover what they had paid. The revenue booked from such transactions ran well into the millions of dollars.

[ 123 ] Livent not only misled the markets, it also fooled its auditor. Deloitte never uncovered the company’s schemes. Livent continued to raise investment capital and reinvest it in unprofitable theatre enterprises. Deloitte’s auditors report for Livent’s Fiscal Year 1997 did not disclose the fraud and, although Deloitte objected when Gottlieb presented a misleading quarterly financial statement to the Audit Committee in August 1997, it did not resign.

[ 124 ] The truth came to light in 1998. New equity investors appointed new management, who discovered “irregularities”. Deloitte retracted its audit opinions for 1996 and 1997. A subsequent investigation and re-audit resulted in restated financial reports. Drabinsky and Gottlieb were suspended, fired, and convicted of fraud.

[ 125 ] Livent filed for insolvency protection in both Canada and the United States in November 1998 and sold its assets in August 1999. It went into receivership the following month. The trial judge found that Livent lost $113,000,000 between the time of Deloitte’s negligent failure to end its relationship with Livent and Livent’s insolvency. He reduced that amount by 25 percent for contingencies and awarded $84,750,000 in damages. Livent seeks to recover this loss from Deloitte.

II. Judicial History

[ 126 ] Livent sued Deloitte for $450,000,000 and other relief. It advanced concurrent claims in tort and contract; the parties agreed that the damages under either head would be the same. Livent claimed that Deloitte was responsible for every dollar Livent lost from the date of Deloitte’s breach of duty. Deloitte’s defence was that most of the losses claimed were beyond the scope of Deloitte’s legal responsibility. The courts below largely sided with Livent.

A. Ontario Superior Court of Justice (Commercial List) (Gans J.), 2014 ONSC 2176 , 11 C.B.R. (6th) 12

[ 127 ] The trial judge held that Deloitte owed a duty of care to provide accurate information to Livent’s shareholders. He held that the standard of care under this duty required Deloitte to take steps that would have effectively cut off Livent’s access to the capital markets and forced it into formal insolvency as early as 14 months before it ultimately filed for insolvency protection. The trial judge concluded that Deloitte failed to meet this standard of care, either when it failed to discover the fraud and act on that discovery in August 1997, or when it signed off on Livent’s 1997 financial statements in April 1998.

[ 128 ] The trial judge held that the measure of damages was the difference between Livent’s value when the breach occurred, and Livent’s value at the time of insolvency ($113,000,000). He reduced this by 25 percent to account for “contingencies” or “trading losses” sustained as a result of Livent’s “unprofitable but legitimate theatre business” (paras. 324-26), which he held were too remote to make Deloitte liable.

[ 129 ] The trial judge also rejected Deloitte’s submission that Livent’s recovery should be either barred by the defence of illegality, or reduced on account of contributory fault pursuant to s. 3 of the Negligence Act , R.S.O. 1990, c. N.1.

[ 130 ] The trial judge consequently awarded damages to Livent for breach of its duty of care, and alternatively for breach of contract, in the amount of $84,750,000.

B. Court of Appeal for Ontario (Strathy C.J. and Blair and Lauwers JJ.A.), 2016 ONCA 11 , 128 O.R. (3d) 225

[ 131 ] The Court of Appeal upheld the trial judge’s award and dismissed both the appeal and cross-appeal.

III. Analysis

[ 132 ] The only issue on this appeal is whether Deloitte is liable for the lion’s share of the loss Livent sustained after it failed to detect the fraud of the company’s principals and report it in its audit statements (the total loss less a 25% discount for contingencies and Livent’s improvident investments). Livent claims that if Deloitte had reported the fraud when it should have, it would have been put out of business, which would have saved it from the loss that it suffered thereafter. Instead, it was able to continue to raise more capital and spend it in ways that decreased Livent’s net worth. Deloitte argues that it is not liable for the full extent of Livent’s pre-liquidation loss because this loss falls outside the scope of Deloitte’s duty of care, or is too “remote” from the legal cause.

[ 133 ] The law of negligent misstatement has limited recovery of pure economic (financial) loss for two reasons. The first reason is that it may be unfair to hold a person who makes a negligent misstatement liable for all loss incurred thereafter, where other decisions and acts contributed to that loss. This is referred to as the fair allocation of loss principle. The second reason is to avoid the spectre of indeterminate liability, which the law of negligence has never countenanced.

[ 134 ] These two reasons — one of principle and the other of policy — are complementary. They work together to ensure fair outcomes and promote predictability in the law.

[ 135 ] As Lord Bridge observed in Caparo Industries plc. v. Dickman , [1990] 1 All E.R. 568 (H.L.), at p. 576 :

To hold the maker of the statement to be under a duty of care in respect of the accuracy of the statement to all and sundry for any purpose for which they may choose to rely on it is not only to subject him, in the classic words of Cardozo C.J., to “liability in an indeterminate amount for an indeterminate time to an indeterminate class” ([ Ultramares Corp. v. Touche , 174 N.E. 441 (N.Y. 1931), at p. 444]), it is also to confer on the world at large a quite unwarranted entitlement to appropriate for their own purposes the benefit of the expert knowledge or professional expertise attributed to the maker of the statement.

See also D’Amato v. Badger , [1996] 2 S.C.R. 1071, at para. 18 ; Hercules Managements Ltd. v. Ernst & Young , [1997] 2 S.C.R. 165, at para. 31 ; Canadian National Railway Co. v. Norsk Pacific Steamship Co. , [1992] 1 S.C.R. 1021, at p. 1137, per McLachlin J.; R. v. Imperial Tobacco Canada Ltd. , 2011 SCC 42 , [2011] 3 S.C.R. 45, at para. 99 , quoting Cooper v. Hobart , 2001 SCC 79 , [2001] 3 S.C.R. 537 , at para. 54 .

[ 136 ] For these reasons, common law courts have rejected a simple “but for” test for recovery of economic loss for negligent misstatement: BG Checo International Ltd. v. British Columbia Hydro and Power Authority , [1993] 1 S.C.R. 12, at p. 44; South Australia Asset Management Corp. v. York Montague Ltd. , [1996] 3 All E.R. 365 (H.L.) (“ SAAMCO ”), at pp. 369-72 , per Lord Hoffmann; Hughes-Holland v. BPE Solicitors , [2017] UKSC 21, [2017] 2 W.L.R. 1029, at para. 38 , per Lord Sumption; Hogarth v. Rocky Mountain Slate Inc. , 2013 ABCA 57 , 542 A.R. 289, at paras. 37-38 , per Slatter J.A.

[ 137 ] A “but for” test, which asks only whether the loss would not have been incurred if the wrongful act had not been committed, casts the net of liability too widely. It would make auditors or advisors retained for limited purposes the insurers of the entire venture and all that flows from it. It would not matter that the loss would not have occurred but for other decisions (like Livent’s ill-judged investment decisions in this case). Nor would it matter that the loss was the result of a complex web of decision making, months and years after the negligent misstatement was produced. All that would be required to recover all subsequent loss from the auditors would be to show that their misstatement played a role in launching the saga of subsequent loss. This is not fair; a person giving advice for one purpose should not be held liable for how other people use that information for other purposes, or be made to carry the entire loss. And, it would lead to indeterminate liability. An auditor giving advice would never know what its exposure might be, or whether its fee is adequate to cover the risk to which it exposes itself.

[ 138 ] Rejection of a “but for” test for assessing recovery of economic loss is cemented in the common law. But it is worth noting that Quebec civil law also recognizes the need to limit recovery of economic loss to losses that bear a sufficient connection to the wrongful feature of the defendant’s conduct: see Wightman v. Widdrington (Succession de) , 2013 QCCA 1187 , at paras. 229-31 and 243-46; see also D. Jutras, “Civil Law and Pure Economic Loss: What Are We Missing?” (1987), 12 Can. Bus. L.J. 295, at pp. 308-9.

[ 139 ] Courts have given two doctrinal explanations for limiting recovery of economic loss following from a negligent misstatement. The first is to say that the scope of the duty of care of the advice-giver does not cover the loss claimed. The second is to say that the loss is too remote from the negligent act and thus was not legally caused by that act.

[ 140 ] While lawyers may argue over which approach is preferable, the fact once again is that they are complementary — two sides of the same coin. In fact, the inquiry into the duty of care and the inquiry into remoteness invoke similar considerations.

[ 141 ] The “scope of the duty of care” inquiry looks to the relationship between the defendant’s advice and the plaintiff’s loss. It asks if that relationship was “proximate”. In cases of economic loss, it inquires into the purpose for which the advice was given and asks whether a reasonable person would have expected, or “foreseen”, that negligent advice would lead to the loss in question by virtue of the plaintiff’s reliance on the advice: Hercules , at paras. 24 and 41. Put simply: Did the loss flow from the advice, or from subsequent decisions and circumstances? Thus in Caparo , at p. 581, Lord Bridge confirmed that the scope of the tort determines the extent of the remedy to which the injured party is entitled. See also Platform Home Loans Ltd. v. Oyston Shipways Ltd. , [1999] 1 All E.R. 833 (H.L.), at p. 847, per Lord Hobhouse .

[ 142 ] The “remoteness” approach looks at similar factors to the “scope of the duty of care” approach — all pointing to the wrongdoing and its proximity to the loss claimed. The factors to be considered are not closed. The advice-giver’s knowledge of the claimant’s circumstances, the reasonable expectations arising from the relationship, and the presence of intervening factors that led to the loss may figure in the analysis: Mustapha v. Culligan of Canada Ltd. , 2008 SCC 27 , [2008] 2 S.C.R. 114, at paras. 12 and 14-16 ; Citadel General Assurance Co. v. Vytlingam , 2007 SCC 46 , [2007] 3 S.C.R. 373, at para. 31 ; Westmount (City) v. Rossy , 2012 SCC 30 , [2012] 2 S.C.R. 136, at para. 48 .

[ 143 ] I agree with Lord Sumption’s observation in BPE Solicitors , a recent decision from the United Kingdom Supreme Court, that however one looks at the matter — from the perspective of the scope of the duty of care or from the perspective of remoteness — one arrives at the same point. In his Lordship’s words, “[w]hether one describes the principle . . . as turning on the scope of the duty or the extent of the liability for breach of it does not alter the way in which the principle applies”: para. 38; see also L. N. Klar and C.S.G. Jefferies, Tort Law (6th ed. 2017), at pp. 565-66.

[ 144 ] For the purposes of these reasons, I will first consider whether the losses at issue fall within the scope of Deloitte’s duty of care. This inquiry takes us to the two-part analysis for determining the existence of a duty of care and its scope prescribed in Anns v. London Borough of Merton , [1977] 2 All E.R. 492 (H.L.) — does the relationship between the parties give rise to a prima facie duty of care to avoid the type of loss claimed, and, if so, is that duty negatived by policy considerations?

[ 145 ] Courts in the United Kingdom have resiled from a two-step Anns test, but continue to insist that the scope of the duty of care must be carefully limited having regard to the relationship between the defendant’s conduct and the plaintiff’s injuries, context and policy. They have insisted that there is no such thing as a duty of care in the abstract; the duty is always defined by its scope. As Lord Bridge stated in Caparo , at p. 581, “[i]t is never sufficient to ask simply whether A owes B a duty of care. It is always necessary to determine the scope of the duty by reference to the kind of damage from which A must take care to save B harmless”: see also BPE Solicitors , at paras. 21-23; Sutherland Shire Council v. Heyman , (1985), 1988 ABCA 234, 60 A.L.R. 1 (H.C.) , at p. 40, per Brennan J.; Platform Home Loans , at p. 847, per Lord Hobhouse, quoting Overseas Tankship , (U.K.) Ltd. v. Morts Dock & Engineering Co. Ltd. [1961] A.C. 388 (P.C.) (“The Wagon Mound No. 1”) at p. 425; Candler v. Crane Christmas & Co. , [1951] 1 All E.R. 426 (C.A.) , at p. 436, per Denning L.J., dissenting.

[ 146 ] The first step of the Anns test asks whether there is proximity, or a sufficiently close relationship, between the parties. It focuses on the connection between the defendant’s undertaking (the breach of which is the wrongful act) and the loss claimed. Did the defendant owe the plaintiff a prima facie duty of care to prevent the loss, having regard to, on one hand, the reasonably foreseeable consequences of the defendant’s conduct given the proximity of the parties and, on the other hand, factors concerning the relationship between the parties that negate tort liability? ( see Cooper , at paras. 30 and 34 ). Questions of policy relating to the relationship between the parties should be considered at this step of the Anns analysis: Cooper , at para. 37 .

[ 147 ] The purpose for which the statement was made (the undertaking) is pivotal in determining whether a particular type of economic loss was the reasonably foreseeable consequence of the negligence: Hercules , at paras. 37-40 . Was it made to enable the company to raise capital? If so, a loss due to failure to raise capital may be recoverable. Was it made for the purpose of permitting shareholders to review the management of the company? If so, shareholders may recover for loss due to their inability to hold the company to account: Hercules , at paras. 51-57. In each case, one must determine the purpose for which the statement was made, and ask whether the loss in question is proximate, or closely connected to the failure of the defendant to fulfill that purpose.

[ 148 ] Where an auditor falsely or wrongfully represents that the audit statements are sound and can therefore be used for their intended purpose, this constitutes negligence, or wrongdoing. Economic loss tied to that particular wrongdoing is recoverable; other loss is not. Whether in the United Kingdom, Quebec or the common law provinces of Canada, courts focus on the precise nature of the wrongdoing to determine what loss is recoverable. As Lord Sumption explained in BPE Solicitors , to be a reasonably foreseeable consequence of the defendant’s wrongdoing, the economic loss claimed must have “flowed from the right thing, i.e. from the particular feature of the defendant’s conduct which made it wrongful”: para. 38. Or as Lord Hoffmann said in SAAMCO , at p. 371:

Normally the law limits liability to those consequences which are attributable to that which made the act wrongful. In the case of liability in negligence for providing inaccurate information, this would mean liability for the consequences of the information being inaccurate.

See also: Burns v. Homer Street Development Limited Partnership , 2016 BCCA 371 , 91 B.C.L.R. (5th) 383, at para. 106 ; Hogarth , at paras. 37-38 , per Slatter J.A.

[ 149 ] Lord Sumption pointed out in BPE Solicitors , at para. 44, that the scope of the duty of care and the extent of the defendant’s responsibility turns on the specific circumstances that inform the purpose for which the statement was prepared:

A valuer or a conveyancer, for example, will rarely supply more than a specific part of the material on which his client’s decision is based. He is generally no more than a provider of what Lord Hoffmann [in SAAMCO ] called “information”. At the opposite end of the spectrum, an investment adviser advising a client whether to buy a particular stock, or a financial adviser advising whether to invest self-invested pension fund in an annuity are likely, in Lord Hoffmann’s terminology, to be regarded as giving “advice”. Between these extremes, every case is likely to depend on the range of matters for which the defendant assumed responsibility and no more exact rule can be stated.

See also Aneco Reinsurance Underwriting Ltd. (in liquidation) v. Johnson & Higgins Ltd. , [2001] UKHL 51 , [2001] 2 All E.R. (Comm.) 929, at paras. 40-41, per Lord Steyn, and para. 66, per Lord Millett; Canadian Imperial Bank of Commerce v. Deloitte & Touche , 2016 ONCA 922 , 133 O.R. (3d) 561, at paras. 46-47 and 69-71 ; Temseel Holdings Ltd. v. Beaumonts Chartered Accountants , [2002] EWHC 2642 (Comm.), [2003] P.N.L.R. 27, at paras. 22-29 and 57-62.

[ 150 ] The purpose for which the audit report is provided is a matter of fact based on the evidence adduced at trial: BPE Solicitors , at para. 44; Aneco , at paras. 40-41, per Lord Steyn, and para. 66, per Lord Millett; Canadian Imperial Bank of Commerce , at para. 47 ; Temseel , at paras. 57-62. Legal obligations imposed by statute may be relevant: Hercules , at para. 49.

[ 151 ] Against this background, I turn to the question at hand: What was the scope of the duty of care owed by Deloitte to Livent? The key to answering this question is the purpose for which Deloitte prepared the statements, which in turn defines the wrongful act — the negligent failure to provide the correct information for the intended purpose.

[ 152 ] In this case, three purposes of Livent’s audit statements are discernable: (1) to report accurately on Livent’s finances and provide it with audit opinions on which it could rely for the purpose of attracting investment; (2) to uncover heretofore undetected errors or wrongdoing by Livent or its personnel for the purpose of enabling Livent itself to correct or otherwise respond to the misfeasance; and (3) to provide audit reports on which Livent’s shareholders could rely to supervise the company’s management (see e.g. Part XII, Business Corporations Act , R.S.O. 1990, c. B.16; trial reasons, at paras. 89-96 and 280). Livent was entitled to recover for losses occasioned by its own or its shareholders’ reliance on Deloitte’s audit work for these purposes.

[ 153 ] The scope of Deloitte’s duty of care is defined solely by these purposes. Did its negligence prevent Livent from attracting investment? Did its negligence prevent Livent from uncovering undetected wrongdoing for the purpose of allowing Livent to correct the misfeasance? Finally, did its negligence prevent shareholders from supervising the company’s management? Loss that results from inability to attract investment, from inability of Livent to correct undetected wrongdoing, or from inability of the shareholders to exercise their supervisory authority, may fall within the scope of Deloitte’s duty of care.

[ 154 ] The first possibility is that Deloitte’s wrongful act deprived Livent of the ability to attract investment capital. Livent does not rely on this; in fact, Livent attracted a great deal of capital on the strength of Deloitte’s statements. Indeed, that is the essence of its complaint — if it had not been able to attract this money, it would not have been able to spend it on new theatre ventures that failed and decreased Livent’s worth. The company’s assets were not diminished by an inability to attract investment, but by Livent’s improvident management of those investments, revealed only on insolvency.

[ 155 ] The second possibility is that the wrongful act prevented Livent — the company itself — from detecting misfeasance in the company’s management which Livent’s officials would have corrected if they had known the true state of affairs. This possibility envisions the situation where a company’s management, acting honestly and diligently, is unable to deal with internal misfeasance because the auditors negligently failed to reveal it.

[ 156 ] This is not Livent’s situation. Livent led no evidence that its management did not know of Drabinsky’s and Gottlieb’s misfeasance; indeed, it likely could not have done so. Drabinsky and Gottlieb, the fraudsters, were themselves the management. Far from relying on the audited statements as assurance that everything was well with the company, Drabinsky and Gottlieb knew the audit reports were inaccurate. There is no evidence that anyone at a lower level of Livent management would have blown the whistle if Livent’s statements had revealed the fraud at an earlier date.

[ 157 ] The third possibility is that Deloitte’s wrongdoing prevented Livent’s shareholders from exercising shareholder supervision in a manner that would have ended the corporation’s loss-creating activities at an earlier date. Livent argues (and my colleagues Gascon and Brown JJ. accept) that it relied on Deloitte to produce a report on the basis of which its shareholders could discharge their supervisory function, which all agree was one of the purposes for which the audited statements were prepared. Livent argues, and my colleagues conclude, that all loss that shareholder supervision might have avoided is recoverable — including the decline in the value of the company.

[ 158 ] This proposition faces two difficulties. The first is that Livent never proved, nor did the trial judge find, the elements necessary to establish it. The second is a related policy concern: to allow recovery in the absence of the required proof would be to open the door to indeterminate recovery. I will consider each of these difficulties in turn.

[ 159 ] First, although the trial judge’s reasons refer to Deloitte’s duty to Livent’s shareholders as established in Hercules , the factual basis for liability based on impaired shareholder supervision was lacking. Livent’s theory of the case was simply that all loss as a result of improvident investments after the negligent audits was compensable, on a “but for” basis. Livent did not prove and the trial judge did not find that Livent’s shareholders relied on Deloitte’s negligent audit statements, or that had they received and relied on accurate statements, they would have acted in a way that would have prevented Livent from carrying on business and diminishing its assets in the period between the issuance of the relevant statements and Livent’s insolvency.

[ 160 ] The trial judge, accepting Livent’s theory of the case, held that the duty of care must be broad enough to catch all losses that would be captured by a “but for” test. He stated the following:

In my view, the ultimate issuance or withholding of a clean opinion is but one aspect of conducting an audit in accordance with [generally accepted auditing standards]. It is not close to being the complete embodiment of the duty of care. Indeed, if [the plaintiff’s] argument were accepted, it would preclude the applicability of the “but for” test. Instead of asking whether damages would have been sustained but for the negligent failure to detect certain errors or fraud, one would be restricted to asking whether or not damages would have been sustained but for the provision of a clean opinion. [Emphasis added; para. 285.]

[ 161 ] The broad view of the duty of care taken by Livent, and accepted by the trial judge meant that the trial judge failed to consider the parameters of the shareholders’ reasonable and foreseeable reliance as required in Hercules when defining the scope of the duty of care with respect to losses stemming from impaired shareholder supervision. My colleagues conclude otherwise, noting that the trial judge believed that the shareholders were entitled to rely on Deloitte’s negligent audits as an indication of Livent’s health. Crucially, however, the trial judge did not ask whether the shareholders had in fact relied on the audits – a critical element to the cause of action. He did not ask whether, if they had relied, this reliance prevented them from taking steps to replace directors or officers or otherwise alter course. He did not ask whether this would have included shutting down Livent on March 31st, 1998 (or at least earlier than when it was shut down in November 1998). Finally, he did not ask whether these actions, had they been taken, would have prevented the losses that Livent built up during the seven-month period in question. If the trial judge had asked these questions, he would have been obliged to answer them in the negative, since Livent offered no proof to support affirmative answers.

[ 162 ] I leave aside for the purposes of this case the difficulty that the shareholders, unlike in Hercules , are not parties to the claim or the action. Assuming this would not pose a problem, it is possible to speculate that the necessary facts could have been proven (although given the short time period, it seems unlikely). But the more fundamental point is that Livent did not prove these matters, and that as a result, the factual basis for establishing loss on the basis of shareholder supervision was entirely lacking. The hypothetical chain of events required to establish liability on this ground completely bypasses the complexity of shareholder decision making.

[ 163 ] Livent’s position and the trial judge’s approach collapse the distinction between shareholder decision making, for which an auditor provides information for one purpose — holding management accountable with a view to the best interest of the company — with management decision making, for which the auditor provides information for a different purpose — responding to error and wrongdoing. Conceiving liability in this way creates a misalignment between the scope of the duty of care, the type of loss that is therein contemplated, and the actual elements that must be proven in order to make a successful claim.

[ 164 ] The second and related difficulty the shareholder supervision argument faces in this case is that it would lead to unfair allocation of loss and indeterminate liability for auditors’ statements, negating the duty of care: Hercules , at paras. 36-37.

[ 165 ] Livent’s position — that Deloitte is liable for all loss without proof of the elements required to advance a case based on impaired shareholder supervision — would result in an unfair allocation of loss as well as indeterminacy of damages. On the shareholder supervision theory advanced by Livent, breach of a duty owed primarily to the collectivity of shareholders (who do not advance a claim) and only derivatively to the corporation, would result in liability for every dollar that Livent spent after the point in time the shareholders became entitled to rely on the statements. The auditor would be the virtual guarantor of everything Livent — not the collectivity of shareholders to which the duty was owed — did thereafter. This would not be a fair allocation of responsibility. The same scenario would raise the spectre of indeterminate liability. Auditors would be unable to reasonably predict when they are providing services to clients what their ultimate liability would be. It would be out of their control. No matter how bad the decisions made by the client thereafter, no matter how complex the web of dealings that led to the ultimate loss — things that cannot be foreseen in advance — the auditor would be liable for the total loss, on the basis that it would not have occurred “but for” the negligent act.

[ 166 ] For the foregoing reasons, I conclude that the losses at issue have not been shown to fall within the scope of Deloitte’s duty of care. The first step of the Anns test is not established. It is unnecessary to go on to ask whether prima facie liability is negated by policy considerations unrelated to the relationship between the parties. However, were it necessary to do so, the policy considerations of unfair allocation of loss and indeterminacy would preclude imposing liability on Deloitte.

[ 167 ] I add a doctrinal side-note at this point. In Hercules , this Court, per La Forest J., held that the spectre of indeterminate loss might be a policy consideration negating liability of certain loss at the second step of the Anns test. However, since Cooper it has been clear that policy considerations relating to the relationship between the parties fall to be considered at the first step of Anns. It may not therefore be necessary to resort to the second step to consider the implications of indeterminate liability: see J. Blom, “Do We Really Need the Anns Test for Duty of Care in Negligence?” (2016), 53 Alta. L. Rev. 895, at pp. 906 and 908.

[ 168 ] It remains for future cases to explore the limits of an auditor’s liability for impaired shareholder supervision. Short of finding that an auditor who provides information on which shareholders will exercise their oversight powers assumes responsibility for all of the potential consequences of carrying on business, the facts as found by the trial judge do not enable us to do so here.

[ 169 ] Livent’s failure to prove that any of the loss it suffered can be attributed to its shareholders’ reliance on the negligent 1997 year-end audit report for the purpose of corporate oversight is a sufficient basis on which to allow the appeal in its entirety. But there is further problematic aspect of the shareholder supervision theory on which Livent now seeks to rely. It is one of principle.

[ 170 ] As explained above, an auditor that provides a year-end report for the purpose of enabling a company’s shareholders to supervise management does not, absent proof, assume responsibility for what the shareholders decide to do with that information. The purpose of an annual audit report is to inform shareholder decision making, not to govern it. The auditor does not underwrite the entire risk associated with the shareholders’ exercise of business judgment; it is liable only for exposing shareholders to the risk of the information it has provided being wrong. Even if the whole loss would have been avoided if the auditor had met the standard of care, the company may recover in damages only that part of the loss that may be attributed to the auditor’s breach of its duty of care, which is restricted by the purpose or purposes for which it provided its opinion. The auditor is not liable for the indeterminate quantum of loss that the shareholders’ course of action (or inaction) may trigger, since determining that course of action is beyond the auditor’s undertaking of responsibility, and thus outside the scope of its duty of care. Similarly, loss that cannot be attributed to the auditor’s breach will be too remote to recover.

[ 171 ] For this reason, the language of lost opportunity is unavailing. In any case concerned with the tortious provision of information, the plaintiff may claim that, had it only known the truth, it would have had the chance to make different choices than it ultimately did. Unless the provider of information has assumed responsibility not only for the information but also for the decision to be informed by it, what the plaintiff would have done differently if it had been provided with different information is immaterial. Rather, the question is the extent to which the loss that in fact resulted may be attributed to the wrongness (i.e., the tortious quality) of the information provided.

[ 172 ] On the trial judge’s findings, Deloitte never assumed responsibility for any of the decisions — whether of Livent’s management or of its shareholders collectively — that may be said to have occasioned Livent’s loss. What Livent proved is that it relied on Deloitte’s clean opinions to raise funds from third parties, and that it was successful in doing so. Deloitte is not liable to Livent for loss arising from Livent’s use of these funds, even if certain of Deloitte’s opinions were prepared for the purpose of attracting them, because Livent did not prove that Deloitte undertook responsibility for how Livent spent them.

[ 173 ] The foregoing analysis is based on the scope of Deloitte’s duty of care. The same result would follow on a remoteness approach to the question of liability for economic loss as a result of negligent misstatement. The question on this approach is whether the loss claimed is too remote from the wrongful act for the act to be the “legal cause” of the loss. The basic question is whether the loss is reasonably foreseeable, having regard to a variety of factors, including the relationship between the parties and the expectations that flow from it, the circumstances of the case, and other factors bearing on the connection between the wrongful act and the loss claimed including external or intervening influences. In the end, a close and proximate connection between wrongful act and the loss claimed must be established, having regard to all these things and to the purpose for which the information was furnished.

[ 174 ] It follows that Livent cannot recover the losses it claims against Deloitte. The claim in tort must be dismissed. The result is the same with respect to Livent’s action in contract; here too the losses would be too remote: see B.D.C. Ltd. v. Hofstrand Farms Ltd ., [1986] 1 S.C.R. 228, at pp. 243-44, citing Asamera Oil Corp. Ltd. v. Sea Oil & General Corp. , [1979] 1 S.C.R. 633; S. M. Waddams, The Law of Damages (5th ed. 2012), at §14.720. Given the trial judge’s determination, at para. 243, that the elements of the action in contract were identical to the elements of the action in tort, it follows that the trial judge’s conclusion that there was a breach in contract is erroneous.

[ 175 ] Although in many respects I agree with Gascon and Brown JJ.’s articulation of the general framework that governs this matter, I part company with my colleagues in two key respects. First, I take the view that, for Livent to make out its claim, it must prove on the evidence the elements required to establish Deloitte’s liability on the basis of impaired shareholder supervision. My colleagues suggest that, had Deloitte provided sound audit reports, Livent’s shareholders and management may have made decisions that would have limited the company’s losses. While this may be true, it is not enough to rely on unproven assertions to define the scope of the duty of care and to subsequently demonstrate causation. My colleagues’ approach suggests that an auditor will generally become the underwriter for any losses suffered by a client following a negligent audit report. This, notwithstanding subsequent decisions — reliant or capricious — made by the client’s shareholders. I conclude, in contrast, that reliance cannot be presumed, but must be proved.

[ 176 ] Second, I take a different view of indeterminate liability than my colleagues. They assert that liability is not indeterminate where a reviewing court can set a value or time frame for a plaintiff’s claimed loss ex post facto . However, the common law’s policy against indeterminacy is directed at ensuring that auditors and other advisors can determine the scope of their liability at the time they take on an engagement and render their services. The question is whether an auditor or other advisor was able to gauge the scale of its potential liability — in terms of the types of losses for which it undertook responsibility — before embarking on a course of conduct. Although Deloitte might have been in a position to identify the total net value of Livent, Livent has not proved that Deloitte bore responsibility for the myriad ways that Livent could have gone about depleting its value after receiving the auditor’s statements. This is what makes the liability identified by my colleagues indeterminate and therefore outside the scope of the duty of care.

[ 177 ] Having concluded that Deloitte is not liable for the losses claimed, it is not necessary to consider the apportionment of damages under s. 3 of the Negligence Act , nor is it necessary to consider whether Deloitte can raise the defence of illegality against Livent.

IV. Disposition

[ 178 ] I would allow the appeal, with costs to Deloitte.

Appeal allowed in part with costs to Livent Inc. throughout, McLachlin C.J. and Wagner and Côté JJ. dissenting in part.

Solicitors for the appellant: Lenczner Slaght Royce Smith Griffin, Toronto; Dentons Canada, Toronto.

Solicitors for the respondent: Stikeman Elliott, Toronto and Ottawa.

Solicitors for the intervener the Canadian Coalition for Good Governance: McMillan, Toronto.

Solicitors for the intervener Chartered Professional Accountants of Canada: Borden Ladner Gervais, Toronto and Ottawa.