[1]                             Under Canadian tort law, there is no doubt that governments may sometimes be held liable for damage caused by their negligence in the same way as private defendants. At the same time, the law of negligence must account for the unique role of public authorities in governing society in the public interest. Public bodies set priorities and balance competing interests with finite resources. They make difficult public policy choices that impact people differently and sometimes cause harm to private parties. This is an inevitable aspect of the business of governing. Accountability for that harm is found in the ballot box, not the courts. Courts are not institutionally designed to review polycentric government decisions, and public bodies must be shielded to some extent from the chilling effect of the threat of private lawsuits.

[2]                             Accordingly, courts have recognized that a sphere of government decision-making should remain free from judicial supervision based on the standard of care in negligence. Defining the scope of this immunity has challenged courts for decades. In R. v. Imperial Tobacco Canada Ltd., 2011 SCC 42, [2011] 3 S.C.R. 45, this Court explained that “core policy” government decisions — defined as “decisions as to a course or principle of action that are based on public policy considerations, such as economic, social and political factors” — must be shielded from liability in negligence (para. 90). In ascertaining whether a decision is one of core policy, the key focus is always on the nature of the decision.

[3]                             In the decade since Imperial Tobacco, there has been continued confusion on when core policy immunity applies. This appeal requires the Court to clarify how to distinguish immune policy decisions from government activities that attract liability for negligence. We conclude that the rationale for core policy immunity serves as an overarching guiding principle. Core policy decisions are immune from negligence liability because each branch of government has a core institutional role and competency that must be protected from interference by the other branches. We identify four factors from this Court’s jurisprudence that help in assessing the nature of a government’s decision: (1) the level and responsibilities of the decision-maker; (2) the process by which the decision was made; (3) the nature and extent of budgetary considerations; and (4) the extent to which the decision was based on objective criteria. The separation of powers rationale animating the immunity guides how the factors weigh in the analysis.

[4]                             The respondent, Taryn Joy Marchi, was injured while attempting to cross a snowbank created by the appellant, the City of Nelson, British Columbia. She sued the City for negligence. Dismissing her claim, the trial judge concluded that the City did not owe Ms. Marchi a duty of care because its snow removal decisions were core policy decisions. In the alternative, he also found that there was no breach of the standard of care and, if there was a breach, Ms. Marchi was the proximate cause of her own injuries. The Court of Appeal concluded that the trial judge erred on all three conclusions and ordered a new trial.

[5]                             We agree with the Court of Appeal that the trial judge erred on all three conclusions. On duty of care, the relevant City decision was not a core policy decision immune from negligence liability. The City therefore owed Ms. Marchi a duty of care. On standard of care and causation, the trial judge’s analysis was tainted by legal errors. As key factual findings are required, this Court is not well placed to determine the standard of care and causation issues. We would therefore dismiss the appeal and order a new trial in accordance with these reasons.

[6]                             The City of Nelson experienced heavy snowfall on January 4th and 5th, 2015. It started plowing and sanding the streets to respond to the snowfall. Among the tasks completed by City employees was the clearing of snow in angled parking stalls on Baker Street, located in the downtown core. Employees plowed the snow to the top of the parking spaces, creating a snowbank along the curb that separated the parking stalls from the sidewalk. Having created the snowbank, the City did not clear an access route to the sidewalk for drivers parking in the stalls.

[7]                             On the evening of January 6th, Ms. Marchi parked in one of the angled parking stalls on Baker Street. She was attempting to access a business, but the snowbank created by the City blocked her route to the sidewalk. She decided to cross the snowbank. As her right foot stepped onto it, however, she dropped through the snow, stepped directly into an area which bent her forefoot up, and seriously injured her leg. She sued the City for negligence and the parties agree that she suffered $1 million in damages.

[8]                             Since 2000, the City has relied on a written document called “Streets and Sidewalks Snow Clearing and Removal” (Policy). Broadly, the Policy states that snow removal, sanding, and plowing will be carried out “on a priority schedule to best serve the public and accommodate emergency equipment within budget guidelines” (A.R., vol. I, at p. 56). The Policy sets out the following priorities: emergency routes and the downtown core; transit routes; plowing hills; cross streets; and dead end streets. Ms. Marchi was injured in the 300 block of Baker Street, which is in the “downtown core”. The Policy also provides specific guidelines that snow plowing will occur during the early morning hours and that snow removal may be carried out as warranted by buildup levels. It does not specifically mention clearing parking stalls or creating snowbanks.

[9]                             In addition to the written Policy, the City also has several unwritten practices. For example, it plows, sands, and removes snow from the designated sidewalk route and the various stairs located in the City. It focuses on Baker Street in the downtown core for snow removal, but to ensure safety, City workers begin to remove snow from other areas, including the civic centre and around schools, when the downtown core starts to get busy (typically around 11:00 a.m.). They return to Baker Street as soon as possible. The City does not remove snow from the downtown core overnight due to noise complaints received in the past as well as the cost of overtime.

[10]                        Throughout the snowfall, the City’s public works supervisor followed the Policy and made decisions about how many employees should be on snow removal shifts. Her evidence was that all streets in the City are first cleared of snow, and snowbanks are only removed after all snow plowing is complete. The downtown core was completely cleared of snow, and all snowbanks were removed, by January 9th, 2015.

[11]                        The trial judge held that the City did not owe Ms. Marchi a duty of care because its snow removal decisions were core policy decisions. The City followed its written and unwritten policies on snow removal and its decisions were dictated by the availability of resources. Alternatively, the trial judge found that the City did not breach the standard of care because the snowbank did not pose an objectively unreasonable risk of harm — the City did what was reasonable in the circumstances. In the further alternative, the trial judge concluded that the City’s alleged negligence did not cause the accident because Ms. Marchi was the “author of her own misfortune” (para. 45).

[12]                        The Court of Appeal unanimously allowed the appeal and ordered a new trial. On duty of care, it held that the trial judge did not properly engage with the distinction between government policy and operation, simply accepting the City’s submission that all snow removal decisions were core policy decisions. On standard of care, it held that the trial judge’s analysis was improperly coloured by his view that the snow removal decisions were core policy decisions. The trial judge accepted the City’s submission that this was “the way it has always been done” (para. 35) without engaging with other municipalities’ evidence on snow removal. On causation, the Court of Appeal held that the trial judge misunderstood how to factor in Ms. Marchi’s own fault. The trial judge improperly reasoned that, if Ms. Marchi could have avoided the accident, she was the proximate cause of her injuries. This was a failure to apply the “but for” test for causation.

[13]                        There are three issues on appeal: whether the trial judge erred in concluding that the City did not owe Ms. Marchi a duty of care because its snow removal decisions were core policy decisions immune from negligence liability; whether the trial judge erred in his standard of care analysis; and whether the trial judge erred in his causation analysis.

[14]                        Duty of care is the central issue. To determine whether the trial judge erred, we proceed as follows. First, we set out the duty of care framework. Second, we explain how the previously established category from Just v. British Columbia, 1989 CanLII 16 (SCC), [1989] 2 S.C.R. 1228, operates, clarifying why this case falls within the Just category. Third, we consider the law on distinguishing core policy decisions from government activities that attract liability in negligence. We then apply the law to the trial judge’s determination in this case that the City did not owe Ms. Marchi a duty of care.

[15]                        The foundation of the modern law of negligence is the neighbour principle established in Donoghue v. Stevenson, 1932 CanLII 536 (FOREP), [1932] A.C. 562 (H.L.), under which “parties owe a duty of care to those whom they ought reasonably to have in contemplation as being at risk when they act” (Rankin (Rankin’s Garage & Sales) v. J.J., 2018 SCC 19, [2018] 1 S.C.R. 587, at para. 16). The neighbour principle does not discriminate between private and public defendants — it is applicable to both alike, subject to any contrary statutory provision or common law principle (Cooper v. Hobart, 2001 SCC 79, [2001] 3 S.C.R. 537, at para. 22).

[16]                        In Canada, the Anns/Cooper test provides a unifying framework to determine when a duty of care arises under the wide rubric of negligence law, including for allegations of negligence against government officials. But as Cooper and subsequent cases make clear, the framework applies differently depending on whether the plaintiff’s claim falls within or is analogous to an established duty of care or whether the claim is novel because proximity has not been recognized before.

[17]                        In novel duty of care cases, the full two-stage Anns/Cooper framework applies. Under the first stage, the court asks whether a prima facie duty of care exists between the parties. The question at this stage is whether the harm was a reasonably foreseeable consequence of the defendant’s conduct, and whether there is “a relationship of proximity in which the failure to take reasonable care might foreseeably cause loss or harm to the plaintiff” (Rankin’s Garage, at para. 18). Proximity arises in those relationships where 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 upon the defendant” (Cooper, at paras. 32 and 34).

[18]                        If there is sufficient proximity to ground a prima facie duty of care, it is necessary to proceed to the second stage of the Anns/Cooper test, which asks whether there are residual policy concerns outside the parties’ relationship that should negate the prima facie duty of care (Cooper, at para. 30). As stated in Cooper, at para. 37, the residual policy stage of the Anns/Cooper test raises questions relating to “the effect of recognizing a duty of care on other legal obligations, the legal system and society more generally”, such as:

[19]                        When the duty of care at issue is not novel, there is generally no need to proceed through the full two-stage Anns/Cooper framework. Over the years, courts in Canada have developed a body of negligence law recognizing categories of cases in which a duty of care has previously been established (Cooper, at para. 41; Childs v. Desormeaux, 2006 SCC 18, [2006] 1 S.C.R. 643, at para. 15; Mustapha v. Culligan of Canada Ltd., 2008 SCC 27, [2008] 2 S.C.R. 114, at para. 5). In such cases, “the requisite close and direct relationship is shown” and the first stage of the Anns/Cooper framework will be complete, as long as the risk of injury was reasonably foreseeable (Deloitte & Touche v. Livent Inc. (Receiver of), 2017 SCC 63, [2017] 2 S.C.R. 855, at para. 26). The second stage of the Anns/Cooper test will rarely be necessary because residual policy concerns will have already been taken into account when the duty was first established (Cooper, at paras. 36 and 39; Livent, at paras. 26 and 28; see also Edwards v. Law Society of Upper Canada, 2001 SCC 80, [2001] 3 S.C.R. 562, at paras. 9-10).

[20]                        This Court’s majority decision in Just established a duty of care. The plaintiff sought damages for personal injury suffered when a boulder fell from a slope above a public highway onto his car. He claimed that the defendant government owed a private law duty of care to properly maintain and inspect the highway and that his loss was caused by the government’s negligent failure to do so.

[21]                        Just provided this Court with an opportunity to apply the full two-stage duty of care framework to a case involving personal injury on a public road. At the prima facie stage, the Court held that users of a highway are in a sufficiently proximate relationship to the province because in creating public highways, the province creates a physical risk to which road users are invited. The province or department in charge can also readily foresee a risk to road users if highways are not reasonably maintained (p. 1236).

[22]                        At the second stage, the Court in Just did not have residual policy concerns about indeterminate liability or the effect of recognizing a duty on other legal obligations. The Court found that the duty of care should apply to public authority defendants “unless there is a valid basis for its exclusion” (Just, at p. 1242). The Court referred to two such bases: first, statutory provisions that exempt the defendant from liability, and second, immunity for “true” policy decisions (pp. 1240-44). While such policy decisions are exempt from claims in negligence, the operational implementation of policy may be subject to the duty of care in negligence. 

[23]                        The Court thus determined that public authorities owe road users a duty to keep roads reasonably safe, but recognized that the duty was subject to a public authority’s immunity for true policy decisions. On the facts of Just, the impugned system of inspection was operational in nature, meaning it could be reviewed by a court to determine whether the government breached the standard of care (pp. 1245-46). The Court’s reasoning is worth quoting at length:

[24]                        As we shall explain, the principles and considerations set out by the Court in Just to assist in distinguishing between policy and operation are relevant to any case in which a public authority is alleged to have been negligent, whether it falls under an established or analogous duty of care or a novel duty of care. The Court recognized the continuing judicial struggle to differentiate policy from operation, but nonetheless understood the necessity of ascertaining when public authorities owe duties in negligence.

[25]                        The Just category of duty of care is firmly established in Canadian law. Over a decade later in Cooper, when this Court gave examples of categories in which proximity had previously been recognized, it specifically observed: “. . . governmental authorities who have undertaken a policy of road maintenance have been held to owe a duty of care to execute the maintenance in a non-negligent manner” (para. 36, citing Just, and Swinamer v. Nova Scotia (Attorney General), 1994 CanLII 122 (SCC), [1994] 1 S.C.R. 445).

[26]                        The case at bar raises the issue of how the decision in Just applies to Ms. Marchi’s claim. Below, we explain why this case fits within the duty of care previously recognized by this Court in Just.  

[27]                        To determine whether a previously established category of duty applies, “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” (Livent, at para. 28). In Just, the Court’s finding of proximity was based on various factors, including the nature of the loss (personal injury) and the fact that the injury occurred on a highway to which the public was invited. Users of the highway could expect that it would be reasonably maintained and there was a reasonably foreseeable risk that harm might befall highway users if it was not.

[28]                        A substantial number of cases have applied the Just category, including decisions of this Court. In Brown v. British Columbia (Minister of Transportation and Highways), 1994 CanLII 121 (SCC), [1994] 1 S.C.R. 420, the plaintiff’s car accident occurred on a sheet of black ice on the road. The Court held that the duty of care to reasonably maintain roads from Just “would extend to the prevention of injury to users of the road by icy conditions” (Brown, at p. 439). Similarly, in Swinamer, a large tree fell on the plaintiff’s truck while he was driving, causing serious injuries. The Court held that the duty of care from Just to reasonably maintain roads clearly applied (pp. 457-59). In both cases, however, the Court went on to find that the decisions at issue were core policy decisions immune from negligence liability. Lower courts have also applied Just where a pedestrian alleges that they suffered personal injury because a public authority failed to maintain a public road, sidewalk, or path in a reasonably safe condition (e.g., Tambeau v. Vancouver (City), 2001 BCSC 651, 20 M.P.L.R. (3d) 195; Talarico v. Northern Rockies (Regional District), 2008 BCSC 861, 47 M.P.L.R. (4th) 242, at paras. 57-58; Bowden v. Withrow’s Pharmacy Halifax (1999) Ltd., 2008 NSSC 252, 48 M.P.L.R. (4th) 250, at para. 113; Lichy v. City of Surrey, 2016 BCPC 55).

[29]                        As demonstrated by Just and subsequent jurisprudence, the factors uniting cases under the Just category are as follows: a public authority has undertaken to maintain a public road or sidewalk to which the public is invited, and the plaintiff alleges they suffered personal injury as a result of the public authority’s failure to maintain the road or sidewalk in a reasonably safe condition. Where these factors are present, the Just category will apply, obviating the need to establish proximity afresh.

[30]                        In this case, the plaintiff suffered significant physical injury on a municipal street in the City’s downtown core. By plowing the parking spaces on Baker Street, the City invited members of the public to use them to access businesses along the street. The plaintiff was attempting to do just that when she fell into a snowbank that had been created by the City during snow removal. The Just category covers a variety of situations, including the prevention of injuries from rocks falling onto the road (Just), the prevention of injuries from trees falling onto the road (Swinamer), and the prevention of injuries from black ice on the road (Brown). It also clearly extends to the prevention of injuries from snowbanks created by a government defendant on the road and sidewalk. In our view, Ms. Marchi has proved that her circumstances fall within the scope of the Just category. As discussed below, it remains open to the City to prove that the relevant government decision was a core policy decision immune from liability in negligence.

[31]                        We are also of the view that the relationship between the plaintiff and defendant is sufficiently close to satisfy a novel proximity analysis. This case involves foreseeable physical harm to the plaintiff and therefore engages one of the core interests protected by the law of negligence (Cooper, at para. 36). Other hallmarks of proximity are also evident: road users are physically present on a space controlled by the public authority; they are invited to the risk by the public authority; and the public authority intends and plans for people to use its roads and sidewalks. It would be reasonably foreseeable to the City that carrying out snow removal in a negligent manner could cause harm to those invited to use the streets and sidewalks in the downtown core.

[32]                        We note that the City suggested in oral argument that its duty of care is grounded in the Occupiers Liability Act, R.S.B.C. 1996, c. 337. The Act does not apply to a “public road” or “public highway” occupied by a municipality (s. 8(2)). We heard no submissions on whether Ms. Marchi’s fall occurred on a public road or highway as those terms are defined in the Act. The trial judge noted that it makes no practical difference in this case whether the Act applies, and for the purposes of these reasons, we assume it does not. Both parties agreed that core policy immunity must be addressed in any event.

[33]                        Finally, some confusion arose in oral argument about where to consider core policy immunity in the duty of care analysis when a previously established category applies. In novel duty of care cases, this Court has conceived of core policy immunity as a stage two residual policy consideration, as was done in Imperial Tobacco and Cooper. Where an established duty of care applies, on the other hand, this Court has stated that a full two-stage Anns/Cooper analysis is generally unnecessary (Cooper, at para. 39). Thus, where the Just category applies, there is no need to repeat the full two-stage analysis already done in Just.

[34]                        Nonetheless, the decision in Just did not decide for all future purposes when an impugned government decision with respect to road maintenance is core policy. The core policy analysis in one case will not necessarily apply to other cases because the factual nature of a decision will likely vary from case to case. While other stage two concerns, like indeterminate liability, need not be taken into account because they were rejected in Just, it will be open to a public authority to raise core policy immunity in each case depending on the factual circumstances.

[35]                        As the City acknowledged before this Court, the onus is always on the public authority to establish that it is immune from liability because a core policy decision is at issue. Thus, once a plaintiff proves that her case falls within the Just category, a duty of care will be imposed, unless the public authority can show that the relevant government decision is protected by core policy immunity. In addition, the government decision must neither be irrational nor taken in bad faith (Imperial Tobacco, at para. 90).

[36]                        For the purposes of this case, we need not decide whether core policy immunity is best conceived of as a rule for how the Just category operates, or whether it should be viewed as a stage two consideration under the Anns/Cooper framework even when an established category of duty applies. It makes no practical difference to the outcome of the appeal. Regardless of where core policy immunity is located in the duty of care framework, the same principles apply in determining whether an immune policy decision is at issue. Those principles apply in any case in which a public authority defendant raises core policy immunity, whether the case involves a novel duty of care, falls within the Just category, or falls within another established or analogous category. What is most important is that immunity for core policy decisions made by government defendants is well understood and fully explored where the nature of the claim calls for it. It is for this reason that we will now articulate the principles underlying the immunity.

[37]                        Before determining whether core policy immunity applies in this case, we will examine the law on how to distinguish core policy decisions from government activities that attract liability for negligence. First, we trace the development of government liability for negligence, explaining why core policy decisions are immune from liability. Second, we describe the principles and factors this Court has already developed to identify a core policy decision. Third, we provide additional guidance on this issue by offering some clarifications and a framework to structure the analysis. Fourth, we apply the law on core policy immunity to the facts of this case.

[38]                        Before the enactment of Crown proceedings legislation in the mid-twentieth century, governments in Canada could not be held directly or vicariously liable for the negligence of Crown servants (P. W. Hogg, P. J. Monahan and W. K. Wright, Liability of the Crown (4th ed. 2011), at p. 7). As government functions expanded, however, this state of affairs became untenable; governments were increasingly involved in activities “that would have led to tortious liability if they had occurred between private citizens” (Just, at p. 1239). Accordingly, Parliament and the provincial legislatures enacted legislation allowing the Crown to be held liable for the torts of officials in a manner akin to private persons. British Columbia’s Crown Proceeding Act, R.S.B.C. 1996, c. 89, s. 2, for example, provides that “the government is subject to all the liabilities to which it would be liable if it were a person”. Even before this legislation was enacted, municipal corporations were distinguished from the Crown and held liable for claims in negligence from the late eighteenth century onwards (D. G. Boghosian and J. M. Davison, The Law of Municipal Liability in Canada (loose-leaf), at § 2.4; N. v. Poole Borough Council (AIRE Centre Intervening), [2019] UKSC 25, [2020] A.C. 780, at para. 26).

[39]                        Applying private law negligence principles to public authorities presents “special problems” (Sutherland Shire Council, at p. 456, per Mason J.). While legislation makes the Crown subject to liability as though it were a person, “the Crown is not a person and must be free to govern and make true policy decisions without becoming subject to tort liability as a result of those decisions” (Just, at p. 1239). Government decision-making occurs across a wide spectrum. At one end are public policy choices that only governments make, such as decisions taken “at the highest level” of government to adopt a course of action based on health policy or other “social and economic considerations” (Imperial Tobacco, at para. 95). Courts are reluctant to impose a common law duty of care in relation to these policy choices (see Dalehite v. United States, 346 U.S. 15 (1953), at p. 57, per Jackson J., dissenting). At the other end of the spectrum, government employees who drive vehicles or public authorities who occupy buildings clearly owe private law duties of care and must act without negligence (L. N. Klar and C. S. G. Jefferies, Tort Law (6th ed. 2017), at p. 348). Tort law must ensure that liability is imposed in this latter category of cases without extending too far into the sphere of public policy decisions.

[40]                        Although there is consensus “that the law of negligence must account for the unique role of government agencies”, there is disagreement on how this should be done (Imperial Tobacco, at para. 76). Some even argue that private law principles of negligence are wholly incompatible with the role and nature of public authorities. Echoing the obiter in Paradis Honey Ltd. v. Canada (Attorney General), 2015 FCA 89, [2016] 1 F.C.R. 446, at paras. 130 and 139, for example, the City of Abbotsford intervened to propose that only public law principles should govern public authority liability. Instead of examining how core policy immunity operates within negligence law, it suggests that courts should focus on indefensibility in the administrative law sense and exercise remedial discretion where appropriate to grant monetary relief.

[41]                        Such an approach has no basis in this Court’s jurisprudence. It also runs counter to Crown proceedings legislation in Canada, which subjects the Crown to liability as if it were a private person. This Court’s approach has been to accept that, “[a]s a general rule, the traditional tort law duty of care will apply to a government agency in the same way that it will apply to an individual” (Just, at p. 1244). However, to resolve the tension arising from the application of private law negligence principles to public authorities, the Court has adopted the principle from Anns v. Merton London Borough Council, [1978] A.C. 728 (H.L.), that certain policy decisions should be shielded from liability for negligence, as long as they are not irrational or made in bad faith. This approach accounts for the unique nature of public authority defendants and is firmly grounded in both the legislation and this Court’s jurisprudence dating back to Barratt v. Corporation of North Vancouver, 1980 CanLII 219 (SCC), [1980] 2 S.C.R. 418, and Kamloops (City of) v. Nielsen, 1984 CanLII 21 (SCC), [1984] 2 S.C.R. 2.

[42]                        The primary rationale for shielding core policy decisions from liability in negligence is to maintain the separation of powers. Subjecting those decisions to private law duties of care would entangle the courts in evaluating decisions best left to the legislature or the executive. The executive, legislative, and judicial branches of government play distinct and complementary roles in Canada’s constitutional order (Ontario v. Criminal Lawyers’ Association of Ontario, 2013 SCC 43, [2013] 3 S.C.R. 3, at paras. 27-29). Each branch also has core institutional competencies: the legislative branch has the power to make new laws, the executive branch executes the laws enacted by the legislative branch and the judicial branch decides disputes arising under the laws (P. W. Hogg and W. K. Wright, Constitutional Law of Canada (5th ed. Supp.), at § 9:1 (“Definition of responsible government”)).

[43]                        It is fundamental to the constitutional order that each branch plays its proper role and that “no one of them overstep its bounds, that each show proper deference for the legitimate sphere of activity of the other” (New Brunswick Broadcasting Co. v. Nova Scotia (Speaker of the House of Assembly), 1993 CanLII 153 (SCC), [1993] 1 S.C.R. 319, at p. 389; see also Just, at p. 1239). Separation of powers thus protects the independence of the judiciary, the legislature’s ability and freedom to pass laws (Mikisew Cree First Nation v. Canada (Governor General in Council), 2018 SCC 40, [2018] 2 S.C.R. 765, at paras. 2 and 35; Chagnon v. Syndicat de la fonction publique et parapublique du Québec, 2018 SCC 39, [2018] 2 S.C.R. 687, at para. 65), and the executive’s ability to execute those laws, set priorities, and allot resources for good governance. Since municipalities hold delegated provincial powers, they enjoy the same protection for certain responsibilities.

[44]                        Core policy decisions of the legislative and executive branches involve weighing competing economic, social, and political factors and conducting contextualized analyses of information. These decisions are not based only on objective considerations but require value judgments — reasonable people can and do legitimately disagree (see B. A. Peterson and M. E. Van Der Weide, “Susceptible to Faulty Analysis: United States v. Gaubert and the Resurrection of Federal Sovereign Immunity” (1997), 72 Notre Dame L. Rev. 447, at p. 450). If courts were to weigh in, they would be second-guessing the decisions of democratically-elected government officials and simply substituting their own opinions (Andrews v. Law Society of British Columbia, 1989 CanLII 2 (SCC), [1989] 1 S.C.R. 143, at p. 194, per La Forest J. (concurring); S. M. Makuch, “Municipal Immunity From Liability in Negligence”, in F. M. Steel and S. Rodgers-Magnet, eds., Issues in Tort Law (1983), 221, at pp. 232-34).

[45]                        Relatedly, the adversarial process and the rules of civil litigation are not conducive to the kind of polycentric decision-making done through the democratic process (Hogg, Monahan and Wright, at p. 226; Home Office v. Dorset Yacht Co., [1970] A.C. 1004 (H.L.), per Lord Diplock at pp. 1066-1068). Nor is the fact that one core policy decision is better than the other amenable to proof in the sense that courts usually require (Just, at pp. 1239‑40, citing Blessing v. United States, 447 F. Supp. 1160 (E.D. Pa. 1978); see also Berkovitz v. United States, 486 U.S. 531 (1988), at p. 545).

[46]                        Moreover, if all government decisions were subject to tort liability, this could hinder good governance by creating a chilling effect (L. N. Klar, “The Supreme Court of Canada: Extending the Tort Liability of Public Authorities” (1990), 28 Alta. L. Rev. 648, at p. 650; see also United States v. Muniz, 374 U.S. 150 (1963), at p. 163). Public authorities must be allowed to “adversely affect the interests of individuals” when making core policy decisions without fear of incurring liability (Laurentide Motels Ltd. v. Beauport (City), 1989 CanLII 81 (SCC), [1989] 1 S.C.R. 705, at p. 722).

[47]                        For these reasons, although there is no question that the legislative and executive branches sometimes make core policy decisions that ultimately cause harm to private parties (Klar, at p. 650), the remedy for those decisions must be through the ballot box instead of the courts (Anns, at p. 754; George v. Newfoundland and Labrador, 2016 NLCA 24, 378 Nfld. & P.E.I.R. 46, at para. 159). Unlike public (administrative) law, where delegated government decisions are reviewed by the courts to uphold the rule of law, private law liability for core policy decisions would undermine our constitutional order. 

[48]                        Conversely, there are good reasons to hold public authorities liable for negligent activities falling outside this core policy sphere where they cause harm to private parties. “Municipalities function in many ways as private individuals or corporations do” and have the ability to “spread losses” (Makuch, at p. 239). Liability for operational activities is “a useful protection to the citizen whose ever-increasing reliance on public officials seems to be a feature of our age” (Kamloops, at p. 26).

[49]                        As we will explain, the rationale for core policy immunity — protecting the legislative and executive branch’s core institutional roles and competencies necessary for the separation of powers — should serve as an overarching guiding principle in the analysis. Ultimately, whether a public authority ought to be immune from negligence liability depends on whether and the extent to which the underlying separation of powers rationale is engaged (see, e.g., United States v. Varig Airlines, 467 U.S. 797 (1984), at p. 814; Berkovitz, at pp. 536-37).

[50]                        This Court explained in Just that “[t]rue policy decisions” must be distinguished from “operational implementation” which is subject to private law principles of negligence (p. 1240). As demonstrated by lower court decisions and academic literature — as well as the submissions before this Court — the question of what constitutes a “true” or core policy decision is a “vexed one, upon which much judicial ink has been spilled” (Imperial Tobacco, at para. 72). There can be no magic formula or litmus test producing an obvious answer for every government decision (para. 90).

[51]                        Nevertheless, our jurisprudence provides helpful guidance. Core policy decisions, shielded from negligence liability, are “decisions as to a course or principle of action that are based on public policy considerations, such as economic, social and political factors, provided they are neither irrational nor taken in bad faith” (Imperial Tobacco, at para. 90). They are a “narrow subset of discretionary decisions” because discretion “can imbue even routine tasks” and protecting all discretionary government decisions would therefore cast “the net of immunity too broadly” (paras. 84 and 88).

[52]                        Activities falling outside this protected sphere of core policy — that is, activities that open up a public authority to liability for negligence — have been defined as “the practical implementation of the formulated policies” or “the performance or carrying out of a policy” (Brown, at p. 441; see also Laurentide Motels, at p. 718). Such “operational” decisions are generally “made on the basis of administrative direction, expert or professional opinion, technical standards or general standards of reasonableness” (Brown, at p. 441).

[53]                        In Imperial Tobacco, McLachlin C.J. suggested that the policy/operational distinction may not work well as a legal test, as many decisions can be characterized as one or the other when abstractly pitting policy against operation (para.  78). The Court in Imperial Tobacco therefore chose to focus on the positive features of core policy decisions and move away from the policy-operational distinction (para. 87). While we agree, the distinction nevertheless remains useful. In some cases, the juxtaposition of core policy and operational implementation may clearly identify the decisions that should not be subject to court oversight as opposed to those which attract liability in negligence.

[54]                        However, the key focus must remain on the nature of the decision (Just, at p. 1245; see also Imperial Tobacco, at para. 87), and this focus is supported by the identification of additional hallmarks of core policy decisions. In Just, this Court explained that core policy decisions will usually (but not always) be made “by persons of a high level of authority” (p. 1245). This was later echoed by McLachlin C.J. in Imperial Tobacco when she stated that, generally, core policy decisions will be made “by legislators or officers whose official responsibility requires them to assess and balance public policy considerations” (para. 87). In Brown, the Court explained that core policy decisions involve “planning and predetermining the boundaries of [a government’s] undertakin[g]” (p. 441). In addition, “decisions concerning budgetary allotments for departments or government agencies will be classified as policy decisions” (Just, at pp. 1242 and 1245 (emphasis added)).

[55]                        The characteristics of “planning”, “predetermining the boundaries” or “budgetary allotments” accord with the underlying notion that core policy decisions will usually have a sustained period of deliberation, will be intended to have broad application, and will be prospective in nature. For example, core policy decisions will often be formulated after debate — sometimes in a public forum — and input from different levels of authority. Government activities that attract liability in negligence, on the other hand, are generally left to the discretion of individual employees or groups of employees. They do not have a sustained period of deliberation, but reflect the exercise of an agent or group of agents’ judgment or reaction to a particular event (see H. J. Krent, “Preserving Discretion Without Sacrificing Deterrence: Federal Governmental Liability in Tort” (1991), 38 U.C.L.A. L. Rev. 871, at pp. 898-99).

[56]                        Thus, four factors emerge from this Court’s jurisprudence that help in assessing the nature of a government’s decision: (1) the level and responsibilities of the decision-maker; (2) the process by which the decision was made; (3) the nature and extent of budgetary considerations; and (4) the extent to which the decision was based on objective criteria.

[57]                        Below, we offer two clarifications and provide a framework to structure the analysis.

[58]                        The first clarification is that a public servant’s choice on how to approach government services frequently involves financial implications. For this reason, the mere presence of budgetary, financial, or resource implications does not determine whether a decision is core policy — too many government decisions, even the most operational decisions, involve some consideration of a department’s budget or the scarcity of its resources (see, e.g., Peterson and Van Der Weide, at pp. 498-501; A. Deegan, “The Public/Private Law Dichotomy And Its Relationship With The Policy/Operational Factors Distinction in Tort Law” (2001), 1 Q.U.T.L.J.J. 241, at p. 253). In Just, Cory J. was referring to budgetary allotments for departments, which are far removed from the budgetary decisions of individual employees on a day-to-day basis (p. 1245). While a department’s budget “has the feel of a large and complex issue that would be miserable for a judge to attempt to unravel” (Peterson and Van Der Weide, at pp. 500-501), an individual employee’s budgetary or resource allocation decisions are not necessarily difficult to assess. Whether a government decision involved budgetary considerations cannot be a test for whether it constituted core policy; it is but one consideration among many.

[59]                        The second clarification is that the word “policy” has a wide range of meaning, from broad directions to a set of ideas or a specific plan (see, e.g., Cambridge Dictionary (online): “policy” means “a set of ideas or a plan of what to do in particular situations that has been agreed to officially by a group of people, a business organization, a government, or a political party”; Merriam-Webster’s Collegiate Dictionary (11th ed. 2003), at p. 960: “policy” means “a definite course or method of action selected from among alternatives and in light of given conditions to guide and determine present and future decisions” or “a high-level overall plan embracing the general goals and acceptable procedures especially of a governmental body”). This is why our jurisprudence has so often qualified the word policy to focus on “true” or “core” policy, pointing towards the type of policy question that requires immunity. Accordingly, the fact that the word “policy” is found in a written document, or that a plan is labelled as “policy” may be misleading and is certainly not determinative of the question. Similarly, that a certain course of conduct is mandated by written government documents is of limited assistance. While core policy might be expected to be reduced to writing, this may depend on the public authority and the circumstances; implementation procedures may also be documented. The focus must remain on the nature of the decision itself rather than the format or the government’s label for the decision.

[60]                        In addition to these cautions, the principles and factors set out in our jurisprudence, viewed from the perspective of the underlying rationale for core policy immunity, provide a helpful contextual framework for determining whether a government decision is a core policy decision. As noted, the key focus is always on the nature of the decision. Public policy choices are clearly within the role and competence of the legislative and executive branches of government. A court must consider the extent to which a government decision was based on public policy considerations and the extent to which the considerations impact the underlying purpose of the immunity — protecting the legislative and executive branch’s core institutional roles and competencies necessary for the separation of powers.

[61]                        The rationale for core policy immunity should also serve as an overarching guiding principle for how to assess and weigh the factors this Court has developed for identifying core policy decisions. We will elaborate.

[62]                        First: the level and responsibilities of the decision-maker. With this factor, what is relevant is how closely related the decision-maker is to a democratically-accountable official who bears responsibility for public policy decisions. The higher the level of the decision-maker within the executive hierarchy, or the closer the decision-maker is to an elected official, the higher the possibility that judicial review for negligence will raise separation of powers concerns or have a chilling effect on good governance. Similarly, the more the job responsibilities of the decision-maker include the assessment and balancing of public policy considerations, the more likely this factor will lean toward core policy immunity. Conversely, decisions made by employees who are far-removed from democratically accountable officials or who are charged with implementation are less likely to be core policy and more likely to attract liability under regular private law negligence principles (Just, at pp. 1242 and 1245; Imperial Tobacco, at para. 87).

[63]                        Second: the process by which the decision was made. The more the process for reaching the government decision was deliberative, required debate (possibly in a public forum), involved input from different levels of authority, and was intended to have broad application and be prospective in nature, the more it will engage the separation of powers rationale and point to a core policy decision. On the other hand, the more a decision can be characterized as a reaction of an employee or groups of employees to a particular event, reflecting their discretion and with no sustained period of deliberation, the more likely it will be reviewable for negligence.

[64]                        Third: the nature and extent of budgetary considerations. A budgetary decision may be core policy depending on the type of budgetary decision it is. Government decisions “concerning budgetary allotments for departments or government agencies will be classified as policy decisions” because they are more likely to fall within the core competencies of the legislative and executive branches (see, e.g., Criminal Lawyers’ Association, at para. 28). On the other hand, the day‑to‑day budgetary decisions of individual employees will likely not raise separation of powers concerns.

[65]                        Fourth: the extent to which the decision was based on objective criteria. The more a government decision weighs competing interests and requires making value judgments, the more likely separation of powers will be engaged because the court would be substituting its own value judgment (Makuch, at pp. 234-36 and 238). Conversely, the more a decision is based on “technical standards or general standards of reasonableness”, the more likely it can be reviewed for negligence. Those decisions might also have analogues in the private sphere that courts are already used to assessing because they are based on objective criteria.

[66]                        Thus, in the course of weighing these factors, the key focus must always be on the underlying purpose of the immunity and the nature of the decision. None of the factors is necessarily determinative alone and more factors and hallmarks of core policy decisions may be developed; courts must assess all the circumstances.

[67]                        To summarize, core policy decisions are “decisions as to a course or principle of action that are based on public policy considerations, such as economic, social and political factors, provided they are neither irrational nor taken in bad faith” (Imperial Tobacco, at para. 90). They are a “narrow subset of discretionary decisions” — meaning, the presence of choice is not a marker of core policy (ibid., at paras. 84 and 88). Core policy decisions are immune from negligence liability because the legislative and executive branches have core institutional roles and competencies that must be protected from interference by the judiciary’s private law oversight. A court must consider the extent to which a government decision was based on public policy considerations and the extent to which the considerations impact the rationale for core policy immunity.

[68]                        In addition, four factors emerge that help in assessing the nature of a government’s decision: (1) the level and responsibilities of the decision-maker; (2) the process by which the decision was made; (3) the nature and extent of budgetary considerations; and (4) the extent to which the decision was based on objective criteria. The underlying rationale — protecting the legislative and executive branch’s core institutional roles and competencies necessary for the separation of powers — serves as an overarching guiding principle for how to weigh the factors in the analysis. Thus, the nature of the decision along with the hallmarks and factors that inform its nature must be assessed in light of the purpose animating core policy immunity.

[69]                        The City does not claim any statutory exemption from the duty of care in Just and there is no allegation that it was acting irrationally or in bad faith. As such, having determined that this case falls under the Just category, the only remaining issue at the duty of care stage is whether the City is immune from liability in negligence because the plaintiff has challenged a core policy decision. If the City’s impugned actions fall outside the scope of core policy immunity, the City may be held liable for any negligence just as a private defendant would be.

[70]                        The City observes that core policy immunity is a case-specific inquiry, dependent on evidence to do with “the decision-maker’s role, the decision-maker’s level of discretion and the various social, economic and political interests at play” (A.F., at para. 88). Accordingly, the City submits that the characterization of a government decision as core policy is a question of mixed fact and law, reviewable on a standard of palpable and overriding error.

[71]                        We disagree. In 1688782 Ontario Inc. v. Maple Leaf Foods Inc., 2020 SCC 35, this Court reiterated that whether the defendant owed the plaintiff a duty of care is a question of law (para. 24; see also Rankin’s Garage, at para. 19). While the underlying facts as found by the trial judge deserve deference, whether those facts meet the legal test of core policy immunity and, by extension, whether a duty of care arises, is a question of law and must be correct.

[72]                        The trial judge concluded that the City did not owe Ms. Marchi a duty of care because its actions were the result of core policy decisions about the sequence for clearing snow from roadways. He focused on the City’s snow removal in a general way and concluded that the City “followed its policy”. In particular, it allocated several employees to work on snow removal shifts and waited to remove snowbanks from the downtown core until after all City streets were clear to “best serve the public and accommodate emergency equipment” (paras. 4 and 13). More employees could not be called on shifts due to the limited availability of equipment, and the City:

[73]                        The trial judge concluded that, as the City’s decisions were dictated by the availability of resources, the written Policy’s mandate, and several “unwritten policies”, the “City’s actions were the result of policy decisions” (paras. 5 and 14).

[74]                        The City asks this Court to endorse the trial judge’s reasoning. It submits that its written and unwritten “snow clearance and removal policies” are core policy decisions because they involve allocating scarce resources in circumstances where not all stakeholders can be satisfied at once. The allocation of “scarce resources based on a good faith exercise of discretion” is the “core of a political decision” (A.F., at para. 59). The written Policy, for example, requires snow removal to be carried out “within budget guidelines” (para. 60). The written Policy is supplemented by unwritten practices that “further attemp[t] to prioritize certain areas of the City for snow clearance and removal, allocate City resources in an efficient and effective manner, and reduce disruption to residents caused by these operations” (para. 61). At trial, many of the City’s arguments similarly focused on the public works supervisor’s budgetary decisions. The City emphasized that the January snowfall was the City’s first snowfall of the year and the public works supervisor chose not to use more than 20 percent of the year’s snow removal budget.

[75]                        Ms. Marchi submits that the trial judge erred in crucial ways, improperly focusing on snow removal in general without narrowing in on the impugned decision. This case is not about the written Policy’s priority schedule for plowing and sanding or its “snow clearance and removal policies” generally, which are unchallenged. At issue is the clearing of snow from the parking stalls in the 300 block of Baker Street and the creation of a snowbank along the curb without ensuring safe access to sidewalks. Ms. Marchi submits that even assuming that the written Policy is core policy, the clearing of parking stalls and the creation of snowbanks was not mandated by any of the City’s documents; it was the operationalization or implementation of snow removal.

[76]                        We agree with Ms. Marchi and the Court of Appeal that the trial judge erred. First, he described the decision or conduct at issue too broadly, focusing on the entire process of snow removal. At issue is the City’s clearing of snow from the parking stalls in the 300 block of Baker Street by creating snowbanks along the sidewalks — thereby inviting members of the public to park in those stalls — without creating direct access to sidewalks. Even if the written Policy was core policy, this does not mean that the creation of snowbanks without clearing pathways for direct sidewalk access was a matter of core policy. In a duty of care analysis, the decision or conduct at issue must be described with precision to ensure that immunity only attaches to core policy decisions (see, e.g., Imperial Tobacco, at para. 67). The duty asserted must be tied to the negligent conduct alleged. In this case, the plaintiff claims that the City was negligent in how they actually plowed the parking spaces. The trial judge’s conclusion that the “City’s actions were the result of policy decisions” was overbroad, merging together all of the City’s snow removal decisions and activities. The City’s submissions before this Court do the same.

[77]                        Second, we agree with Ms. Marchi that the trial judge placed too much weight on the label of “policy”, appearing to accept without question the City’s description of its unwritten snow removal practices as “unwritten policies”. As we have stated above, however, the word “policy” cannot be determinative of whether government conduct should be immune from negligence liability. The trial judge’s conclusion that all of the City’s unwritten snow removal practices were “unwritten policies” improperly coloured his conclusion that all of the City’s snow removal practices were core policy.

[78]                        Third, the trial judge improperly treated budgetary implications as determinative of the core policy question. The City also goes as far as saying that all of its decisions about extending snow removal hours were core policy decisions because they struck a balance between preserving the budget for future snow events and responding appropriately to the early January snowfall. Again, as noted above, whether a government decision involved budgetary considerations cannot be a test for whether it constituted core policy: even the most routine ones involve some consideration of budget or the scarcity of resources.

[79]                        The City also suggests, relying on White J.’s comments in United States v. Gaubert, 499 U.S. 315 (1991), at p. 324, that if a discretionary government decision could have been made based on policy factors, it will be deemed a policy decision immune from liability. We would reject this submission; whether a decision was in fact core policy must be proven by the public authority.

[80]                        For these reasons the trial judge erred in his legal analysis of whether core policy immunity applied. In our view, applying the correct legal principles to the trial judge’s findings on this record, the City’s clearing of snow from the parking stalls in the 300 block of Baker Street by creating snowbanks along the sidewalks — thereby inviting members of the public to park in those stalls — without ensuring direct access to sidewalks was not the result of a core policy decision immune from negligence liability. Instead, as we shall explain, it was a routine part of the City’s snow removal process, to which little thought was given.

[81]                        The City reacted to the early January snowfall in the usual course: it followed the priority routes for plowing and sanding in the written Policy (unchallenged by Ms. Marchi); it waited to remove snowbanks from the downtown core until after all City streets were plowed; and it followed several unwritten practices, including with respect to snow removal from stairs around the City. Although clearing parking stalls was not covered in the written Policy, the City cleared the angled parking stalls in the 300 block of Baker Street and created a continuous snowbank blocking the stalls from the sidewalk. Throughout this process, the public works supervisor made decisions about how many employees to deploy. She also completed “road patrol throughout the day to ensure the streets [were] safe, and crews [were] working in a timely and efficient manner” (trial reasons, at para. 5(h)).

[82]                        The trial judge found that it “did not occur” to the supervisor that this process could be done in a different manner (para. 35). When the supervisor was asked at trial whether she had ever considered the potential dangers caused by clearing the parking stalls, she responded that her job was simply to follow “[the] normal protocol” and “follow direction from above me” (trial transcript, A.R., vol. IV, at p. 75). She also testified that changing the way the City plowed the streets would have required some “planning ahead” and she would not have had the authority to change the plowing method but would have had to ask her director (p. 79). The City chose not to call any other employees of the City as witnesses.

[83]                        On this record, the City’s decision bore none of the hallmarks of core policy. Although the extent to which the supervisor was closely connected to a democratically-elected official is unclear from the record, she disclosed that she did not have the authority to make a different decision with respect to the clearing of parking stalls (the first factor). In addition, there is no suggestion that the method of plowing the parking stalls on Baker Street resulted from a deliberative decision involving any prospective balancing of competing objectives and policy goals by the supervisor or her superiors. Indeed, there was no evidence suggesting an assessment was ever made about the feasibility of clearing pathways in the snowbanks; the City’s evidence is that this was a matter of custom (the second factor). Although it is clear that budgetary considerations were involved, these were not high-level budgetary considerations but rather the day-to-day budgetary considerations of individual employees (the third factor).

[84]                        Finally, the City’s chosen method of plowing the parking stalls can easily be assessed based on objective criteria (the fourth factor). Cases involving the Just category will not generally raise institutional competence concerns because courts routinely consider road and sidewalk maintenance issues in occupiers’ liability cases. The Just category engages conduct that is similar in kind to what courts routinely assess. The same was true in Hendry v. United States, 418 F.2d 774 (2d Cir. 1969), a public authority negligence case about finding a merchant marine officer mentally unfit for service. The court concluded that the decision was reviewable because it was “not different in kind or complexity from those which courts are accustomed to entertain when tort suits are brought against private physicians” (p. 783). In this case, the court would be well-equipped to determine whether the snowbanks posed an objectively unreasonable risk of harm (the standard of care question). The safety of a road or sidewalk can be measured based on objective or commonly accepted standards as it is in the private sector (see also Peterson and Van Der Weide, at pp. 451-52, fn. 13; J. W. Bagby and G. L. Gittings, “The Elusive Discretionary Function Exception From Government Tort Liability: The Narrowing Scope of Federal Liability” (1992), 30 Am. Bus. L.J. 223, at p. 239).

[85]                        Thus, the City has not shown that the way it plowed the parking stalls was the result of a proactive, deliberative decision, based on value judgments to do with economic, social or political considerations. In these circumstances, a court’s review of the City’s chosen means of clearing the parking stalls in the 300 block of Baker Street does not engage the underlying purpose for core policy immunity. Insulating these kinds of decisions from negligence liability does not undermine the ability to make important public interest policy choices. The public interest is not served when ad hoc decisions that fail to balance competing interests or that fail to consider how best to mitigate harms are insulated from liability in negligence. Oversight of such decisions respects the respective roles of each branch of government under the separation of powers doctrine.

[86]                        Therefore, the City has not met its burden of proving that Ms. Marchi seeks to challenge a core policy decision immune from negligence liability. While there is no suggestion that the City made an irrational or bad faith decision, the City’s “core policy defence” fails and it owed Ms. Marchi a duty of care. The regular principles of negligence law apply in determining whether the City breached the duty of care and, if so, whether it should be liable for Ms. Marchi’s damages.

[87]                        We now turn to the remaining two issues: standard of care and causation.

[88]                        The trial judge also concluded in the alternative that, even if a duty of care was owed, the City did not breach the standard of care because the snowbank did not pose an objectively unreasonable risk of harm. After a reference to the evidence of neighbouring municipalities, he concluded:

[89]                        The City submits that when assessing the standard of care for a public body, courts must consider all of the surrounding circumstances, including budgetary pressures and the availability of personnel and equipment. In this case, the City acted reasonably within the context. Ms. Marchi submits that the trial judge’s standard of care analysis was flawed because he conflated considerations relevant to policy immunity with the standard of care analysis and made errors in his treatment of the evidence.

[90]                        We agree with Ms. Marchi and the Court of Appeal that the trial judge’s treatment of the standard of care was flawed. While he acknowledged that the standard of care applicable to the City is reasonableness, he imported considerations relating to core policy immunity into standard of care and failed to engage with the practices of the neighbouring municipalities.

[91]                        To avoid liability, a defendant must “exercise the standard of care expected that would be of an ordinary, reasonable and prudent person in the same circumstances” (Ryan v. Victoria (City), 1999 CanLII 706 (SCC), [1999] 1 S.C.R. 201, at para. 28). Relevant factors in this assessment include whether the risk of injury was reasonably foreseeable, the likelihood of damage and the availability and cost of preventative measures (P. H. Osborne, The Law of Torts (6th ed. 2020), at pp. 29-30; Bolton v. Stone, [1951] A.C. 850 (H.L.)). A reasonable person “takes precautions against risks which are reasonably likely to happen” (Bolton, at p. 863).

[92]                        The reasonableness standard applies regardless of whether the defendant is a government or a private actor (Just, at p. 1243). In Just, Cory J. recognized that the “standard of care imposed upon the Crown may not be the same as that owed by an individual” (at p. 1244). However, this is not because public policy concerns applicable to governments displace the reasonableness standard. In fact, Cory J. was clear that the analysis under duty of care must be “kept separate and distinct” from the analysis of the standard of care (at p. 1243). It is important that the standard of care analysis not be used as another opportunity to immunize governments from liability, especially when a determination has already been made that the impugned government conduct was not core policy.

[93]                        Thus, the trial judge erred in principle in reaching his conclusion on the standard of care. However, we would reject Ms. Marchi’s invitation to decide this issue without a new trial. This Court is not well-placed to make factual findings regarding the impact of the evidence from other municipalities on the obligations imposed on the City.

[94]                        In the further alternative, the trial judge found that the City was not the cause of Ms. Marchi’s injuries because Ms. Marchi assumed the risk in crossing the snowbank: she was the “author of her own misfortune”. The City submits that the trial judge’s causation analysis is owed deference and, even if a correctness standard applies, the City is not the proximate cause of Ms. Marchi’s injuries because it was not reasonably foreseeable that an individual living in a city prone to snow events would behave the way that Ms. Marchi did. Ms. Marchi submits that the trial judge erred by conflating the analysis of voluntary assumption of risk with causation — Ms. Marchi’s lack of due care is relevant to contributory negligence, not causation.

[95]                        Again, we agree with Ms. Marchi and the Court of Appeal that the trial judge also erred in his causation analysis.

[96]                        It is well established that a defendant is not liable in negligence unless their breach caused the plaintiff’s loss. The causation analysis involves two distinct inquiries (Mustapha, at para. 11; Saadati v. Moorhead, 2017 SCC 28, [2017] 1 S.C.R. 543, at para. 13; Livent, at para. 77; A.M. Linden et al., Canadian Tort Law (11th ed. 2018), at p. 309-10). First, the defendant’s breach must be the factual cause of the plaintiff’s loss. Factual causation is generally assessed using the “but for” test (Clements v. Clements, 2012 SCC 32, [2012] 2 S.C.R. 181, at paras. 8 and 13; Resurfice Corp. v. Hanke, 2007 SCC 7, [2007] 1 S.C.R. 333, at paras. 21-22). The plaintiff must show on a balance of probabilities that the harm would not have occurred but for the defendant’s negligent act.

[97]                        Second, the breach must be the legal cause of the loss, meaning that the harm must not be too far remote (Mustapha, at para. 11; Saadati, at para. 20; Livent, at para. 77). The remoteness inquiry asks whether the actual injury was the reasonably foreseeable result of the defendant’s negligent conduct (Mustapha, at paras. 14-16; Livent, at para. 79). Remoteness is distinct from the reasonable foreseeability analysis within duty of care because it focuses on the actual injury suffered by the plaintiff, whereas the duty of care analysis focuses on the type of injury (Livent, at para. 78; Klar and Jefferies, at p. 565).

[98]                        It is unclear whether the trial judge made any finding on factual causation. He concluded that factual causation could not be proven unless the City had met a standard of care that was impossible to meet. However, this conclusion appears to improperly conflate the standard of care and causation analyses. The trial judge never asked whether, but for the City’s breach of the standard of care, Ms. Marchi would have been injured.

[99]                        The trial judge then turned to remoteness, or legal causation, and concluded that “[t]he plaintiff assumed the risk of crossing the snowbank”; she was the “author of her own misfortune” (at para. 45). The trial judge misapplied a number of distinct concepts in this part of his analysis. Remoteness asks whether the specific injury was reasonably foreseeable. The relevant question is whether it was reasonably foreseeable that the City’s failure to remove the snowbanks could cause Ms. Marchi’s injury. The trial judge never addressed this remoteness question.

[100]                     Instead, the trial judge focused on Ms. Marchi’s conduct. The plaintiff’s conduct is generally relevant to a defence, such as contributory negligence or assumption of risk. Defences are distinct from the causation analysis and the onus is on the defendant to plead and prove defences (Linden et al., at p. 463; British Columbia Electric Railway Co. v. Dunphy, 1919 CanLII 34 (SCC), [1919] 59 S.C.R. 263, at p. 268).The trial judge appears to have misapplied the defence of contributory negligence. Under provincial statutes such as British Columbia’s Negligence Act, R.S.B.C. 1996, c. 333, contributory negligence is no longer a complete bar to recovery. Instead, damages are apportioned on the basis of comparative fault (s. 1(1); Resurfice Corp., at para. 21). Therefore, even if the trial judge found that Ms. Marchi was also negligent, that would not justify his conclusion that the City cannot be blamed for the accident.

[101]                     The trial judge also erred in law in relying on the plaintiff’s assumption of risk, or volenti non fit injuria, which is a complete bar to recovery. This narrowly applied defence requires the defendant to prove that the plaintiff accepted both the physical and legal risks of the activity (Linden et al., at p. 483; Dube v. Labar, 1986 CanLII 67 (SCC), [1986] 1 S.C.R. 649, at pp. 658-59). However, the plaintiff must have “understood that she bargained away her right to sue” (Linden et al., at p. 483). There was no evidentiary basis to conclude that Ms. Marchi either explicitly or implicitly bargained away her right to sue for her injuries.

[102]                     Therefore, the causation analysis will have to be considered in accordance with these reasons.

[103]                     For these reasons, the trial judgment must be set aside. On duty of care, we would conclude that the impugned City decision was not a core policy decision and the City therefore owed Ms. Marchi a duty of care. The standard of care and causation assessments require a new trial. We would therefore dismiss the appeal with costs throughout.