Opinion of the Court by Kruger, J.

To state a cause of action for negligence, a plaintiff must establish the defendant owed a legal duty of care. Generally speaking, all persons have a duty to take reasonable care in their activities to avoid causing injury, though particular policy considerations may weigh in favor of limiting that duty in certain circumstances. (Civ. Code, § 1714; Rowland v. Christian (1968) 69 Cal.2d 108 (Rowland).) The issue before us concerns how courts should decide whether a defendant has a legal duty to take action to protect the plaintiff from injuries caused by a third party. Despite a substantial body of case law addressing the issue, the Courts of Appeal have remained uncertain about the proper legal framework to apply. Distilling the principles articulated in prior cases, we now clarify that whether to recognize a duty to protect is governed by a two-step inquiry. First, the court must determine whether there exists a special relationship between the parties or some other set of circumstances giving rise to an affirmative duty to protect. Second, if so, the court must consult the factors described in Rowland to determine whether relevant policy considerations counsel limiting that duty. Because the Court of Appeal in this case employed the correct framework for decision, we affirm its judgment and remand for further proceedings. I. This case comes to us at the demurrer stage, so for present purposes we assume the truth of the allegations in the complaint. (Southern California Gas Leak Cases (2019) 7 Cal.5th 391, 395 (Gas Leak Cases).) As teenagers, plaintiffs Yazmin Brown, Kendra Gatt, and Brianna Bordon trained in the Olympic sport of taekwondo. They traveled to compete at various events in California and throughout the country with their coach, Marc Gitelman. Gitelman took advantage of these opportunities to sexually abuse the young athletes. This went on for years until the sponsor of these competitions, USA Taekwondo (USAT), banned Gitelman from coaching. Gitelman was ultimately convicted of multiple felonies for the sexual abuse of the minor athletes he trained. Plaintiffs (to whom we will generally refer as Brown) filed this civil suit against Gitelman and several others, including respondents USAT and the United States Olympic Committee (USOC).1 USOC is a federally chartered nonprofit corporation whose central function is to coordinate amateur sports throughout the country for athletes hoping to one day compete in the Olympics. (See 36 U.S.C. § 220505(c)(1).) In this role, USOC certifies and oversees each sport’s national governing body, the entity responsible for conducting and administering the sport in the United States. USAT is the national governing body for the sport of taekwondo. If an athlete wishes to compete in taekwondo at the Olympics or in any other USAT-sponsored competition (as Brown and the other plaintiffs did), the athlete

1 In June 2019, USOC’s name was changed to the United States Olympic and Paralympic Committee. must become a member of USAT and train under a USAT- registered coach (as Gitelman was before USAT banned him). As relevant here, Brown alleged that USOC and USAT were negligent in failing to protect her from Gitelman’s abuse. 2 Brown emphasized that the sexual abuse of young athletes was a known problem; Gitelman’s abuse came on the heels of a series of similar instances of abuse of minors by their coaches dating back to the 1980’s. In the wake of these incidents, USOC mandated that national governing bodies adopt a Safe Sport Program to protect athletes from such abuse. Brown alleged that USAT failed to implement the program in a timely fashion — a fact known to USOC, which placed USAT on probation as a result. Brown further alleged that USAT took insufficient steps to protect Gitelman’s victims once it was made aware of her allegations: USAT temporarily suspended Gitelman, but nevertheless permitted him to continue coaching at USAT competitions for several months before ultimately placing him on its list of banned coaches. USOC and USAT both demurred to the complaint. As relevant here, they argued Brown had not adequately alleged they had an affirmative duty to take action to protect her and the other plaintiffs from Gitelman’s abuse. The trial court sustained both demurrers without leave to amend and entered judgments of dismissal.

2 Brown also raised various other claims against USOC and USAT, including claims that USOC and USAT were vicariously liable for Gitelman’s conduct and claims for negligent hiring and intentional infliction of emotional distress. Those claims are not before us here. Brown appealed. The Court of Appeal reversed the judgment dismissing USAT but affirmed as to USOC. (Brown v. USA Taekwondo (2019) 40 Cal.App.5th 1077, 1083 (Brown).) In determining whether Brown had adequately alleged each defendant had a legal duty to protect plaintiffs from Gitelman’s abuse, the court employed a two-part framework. As a general rule, the court explained, “ ‘ “one owes no duty to control the conduct of another, nor to warn those endangered by such conduct.” ’ ” (Id. at p. 1091, quoting Regents of University of California v. Superior Court (2018) 4 Cal.5th 607, 619 (Regents).) An exception to this no-duty-to-protect rule exists for cases in which the defendant has a special relationship with either the dangerous third party or with the victim. (Brown, at p. 1091.) But even when the so-called special relationship exception applies, the policy considerations described in Rowland, supra, 69 Cal.2d 108, may weigh against imposing a duty to protect in a given case.3 (Brown, at p. 1092.) The court thus asked, first, whether Brown had adequately alleged a special relationship between the parties that gave rise to a legal duty to protect, and second, whether the Rowland factors weighed in favor of limiting or eliminating this duty.

3 These considerations include “the foreseeability of harm to the plaintiff, the degree of certainty that the plaintiff suffered injury, the closeness of the connection between the defendant’s conduct and the injury suffered, the moral blame attached to the defendant’s conduct, the policy of preventing future harm, the extent of the burden to the defendant and consequences to the community of imposing a duty to exercise care with resulting liability for breach, and the availability, cost, and prevalence of insurance for the risk involved.” (Rowland, supra, 69 Cal.2d at p. 113.) Applying this framework, the Court of Appeal concluded that Brown had adequately alleged that USAT owed a duty to protect her from Gitelman. The court first concluded Brown had sufficiently alleged a special relationship between USAT and Gitelman that enabled USAT to control Gitelman’s actions, as demonstrated by the fact that USAT had registered him as a coach, took disciplinary action against him, and ultimately barred him from coaching. (Brown, supra, 40 Cal.App.5th at pp. 1094–1095.) The court then went on to consider whether the Rowland factors counseled against imposing a duty on USAT, and determined they did not. (Id. at pp. 1095–1101.) By contrast, the Court of Appeal concluded that USOC, unlike USAT, had no special relationship with either the plaintiffs or Gitelman, and thus no legal duty to protect the plaintiffs from Gitelman’s abuse. The court explained that Brown’s case for imposing an affirmative duty on USOC rested largely on allegations that USOC had the ability to regulate USAT’s conduct. The court considered this insufficient to establish a special relationship that would enable USOC to control Gitelman’s conduct, or that would give plaintiffs reason to look to the USOC for protection. (Brown, supra, 40 Cal.App.5th at pp. 1101–1103.) Having concluded that Brown’s allegations faltered at the first step of the analysis, the court declined to consider how the Rowland factors might apply to USOC. (Id. at p. 1103.) The Court of Appeal’s decision added to a considerable body of law addressing the connection between the special relationship doctrine and the Rowland factors in cases alleging a duty to protect the plaintiff from harms caused by third parties. The appellate courts that have addressed the issue have adopted various approaches. Several other Courts of Appeal have employed the same two-part framework as the court in this case, holding that a plaintiff must satisfy both the special relationship test and the Rowland factors before a duty to protect the plaintiff from third party harm can be imposed on the defendant. (See, e.g., Barenborg v. Sigma Alpha Epsilon Fraternity (2019) 33 Cal.App.5th 70, 77 [“Thus, plaintiffs alleging a defendant had a duty to protect them must establish: (1) that an exception to the general no-duty-to-protect rule applies and (2) that the Rowland factors support imposition of the duty.”]; Doe v. United States Youth Soccer Assn., Inc. (2017) 8 Cal.App.5th 1118, 1128 [noting that if courts find a special relationship, they go on to “balance[] the policy factors set forth in Rowland [citation] to assist in their determination of the existence and scope of a defendant’s duty in a particular case”].) Other Courts of Appeal, however, have held that a plaintiff can establish a duty to protect by satisfying either the special relationship doctrine or the Rowland factors. Under this approach, Rowland serves as an independent source of duty. (Juarez v. Boy Scouts of America, Inc. (2000) 81 Cal.App.4th 377, 401–402, 410–411 [finding duty under Rowland, but concluding in the alternative that the plaintiff satisfied the special relationship test]; see Adams v. City of Fremont (1998) 68 Cal.App.4th 243, 267–276, 282–288 (Adams) [noting that Rowland factors and the special relationship test are sometimes in conflict and finding no duty to protect under either test, while concluding that this court has generally favored applying Rowland’s multifactor test over the special relationship test]; cf. University of Southern California v. Superior Court (2018) 30 Cal.App.5th 429, 451 [finding no duty under either the special relationship test or Rowland, though recognizing that applying Rowland “may be unnecessary”].) Still other courts have taken the view that the special relationship test incorporates the Rowland factors — that is, that “[r]esolution of the issue whether a special relationship exists giving rise to a duty to protect (or warn) comprehends consideration of the same factors underlying any duty of care analysis” under Rowland. (Hansra v. Superior Court (1992) 7 Cal.App.4th 630, 646 (Hansra); accord, Titus v. Canyon Lake Property Owners Assn. (2004) 118 Cal.App.4th 906, 911–912 (Titus).) Whereas the Court of Appeal in this case employed a two-part framework to evaluate defendants’ legal duty to protect, these courts have reduced the inquiry to a single step, applying the Rowland factors to determine whether a special relationship exists. In view of the different approaches taken by the Courts of Appeal, we granted review to clarify the applicable framework for determining whether a defendant has a duty to protect a plaintiff from harm caused by a third party. We conclude the two-part framework the Court of Appeal applied in this case accurately reflects the law as stated in this court’s precedents, and we accordingly affirm the court’s judgment.4

II.

A. To establish a cause of action for negligence, the plaintiff must show that the “defendant had a duty to use due care, that he breached that duty, and that the breach was the proximate

4 We express no view on the merits of the Court of Appeal’s application of the special relationship test to either USAT or USOC. These fact-dependent issues fall outside the scope of the only question presented for our review. or legal cause of the resulting injury.” (Nally v. Grace Community Church (1988) 47 Cal.3d 278, 292.) Recovery for negligence depends as a threshold matter on the existence of a legal duty of care. (Gas Leak Cases, supra, 7 Cal.5th at p. 397.) Duty is not universal; not every defendant owes every plaintiff a duty of care. A duty exists only if “ ‘the plaintiff’s interests are entitled to legal protection against the defendant’s conduct.’ ” (Dillon v. Legg (1968) 68 Cal.2d 728, 734, quoting Prosser, Torts (3d ed. 1964) § 53, p. 332.) Whether a duty exists is a question of law to be resolved by the court. (Bily v. Arthur Young & Co. (1992) 3 Cal.4th 370, 397.) The “general rule” governing duty is set forth in Civil Code section 1714 (section 1714). (Cabral v. Ralphs Grocery Co. (2011) 51 Cal.4th 764, 771 (Cabral).) First enacted in 1872, section 1714 provides: “Everyone is responsible . . . for an injury occasioned to another by his or her want of ordinary care or skill in the management of his or her property or person . . . .” (Id., subd. (a).) This statute establishes the default rule that each person has a duty “to exercise, in his or her activities, reasonable care for the safety of others.” (Cabral, at p. 768.) Section 1714 states a broad rule, but it has limits. We have explained that the law imposes a general duty of care on a defendant only when it is the defendant who has “ ‘created a risk’ ” of harm to the plaintiff, including when “ ‘the defendant is responsible for making the plaintiff’s position worse.’ ” (Lugtu v. California Highway Patrol (2001) 26 Cal.4th 703, 716, quoting Weirum v. RKO General, Inc. (1975) 15 Cal.3d 40, 49; see Lugtu, at p. 716 [“Under general negligence principles, . . . a person ordinarily is obligated to exercise due care in his or her own actions so as not to create an unreasonable risk of injury to others . . . .” (Citing § 1714.)].) The law does not impose the same duty on a defendant who did not contribute to the risk that the plaintiff would suffer the harm alleged. Generally, the “person who has not created a peril is not liable in tort merely for failure to take affirmative action to assist or protect another” from that peril. (Williams v. State of California (1983) 34 Cal.3d 18, 23 (Williams); accord, Weirum, at p. 49; see Rest.3d Torts, Liability for Physical and Emotional Harm (2012) § 37 [Generally, “[a]n actor whose conduct has not created a risk of physical or emotional harm to another has no duty of care to the other.”].) For example, a person who stumbles upon someone drowning generally has no legal duty to help the victim. The same rule applies to a person who stumbles upon a mugging, for “as a general matter, there is no duty to act to protect others from the conduct of third parties.” (Delgado v. Trax Bar & Grill (2005) 36 Cal.4th 224, 235 (Delgado); see also Regents, supra, 4 Cal.5th at p. 619 [Generally, “ ‘one owes no duty to control the conduct of another, nor to warn those endangered by such conduct.’ ”].)5 This general rule, we have explained, “derives from the common law’s distinction between misfeasance and nonfeasance, and its reluctance to impose liability for the latter.” (Tarasoff v. Regents of University of California (1976)

5 While these examples involve bystanders in the usual sense of the term, it bears emphasis that the relevant legal question is whether the defendant has engaged in activities that created or increased the plaintiff’s risk of harm. A defendant may have greater involvement in the plaintiff’s activities than a chance spectator yet play no meaningful part in exposing the plaintiff to harm. (Cf. conc. opn., post, at p. 7 [rejecting the idea that USOC was a bystander in this case].) 17 Cal.3d 425, 435, fn. 5.) That distinction has deep roots in the law. (See, e.g., Bohlen, The Moral Duty to Aid Others as a Basis of Tort Liability (1908) 56 U.Pa. L.Rev. 217, 219 [“There is no distinction more deeply rooted in the common law and more fundamental than that between misfeasance and non-feasance, between active misconduct working positive injury to others and passive in action, a failure to take positive steps to benefit others, or to protect them from harm not created by any wrongful act of the defendant.”].)6 And although it may sometimes produce outcomes that appear “[m]orally questionable” (Tarasoff, at p. 435, fn. 5), there are several reasons the no-duty-to-protect rule has endured. The most commonly cited reason for the rule is rooted in “the liberal tradition of individual freedom and autonomy” — the idea that a person should be able to freely choose whether to come to the aid of a stranger, without fear of incurring legal liability for the choice. (Rest.3d Torts, supra, § 37, com. e, p. 5.) But our cases have recognized other reasons as well, including “ ‘the difficulties of setting any standards of unselfish service to fellow men,’ ” and the challenge of “ ‘making any workable rule to cover possible situations where fifty people might fail to rescue.’ ” (Tarasoff, at p. 435, fn. 5, quoting Prosser, Torts (4th ed. 1971) § 56, p. 341.)

6 Although our precedents have sometimes referred to the distinction between “misfeasance” and “nonfeasance,” we now understand this terminology to be imprecise and prone to misinterpretation. “The proper question is not whether an actor’s failure to exercise reasonable care entails the commission or omission of a specific act.” (Rest.3d Torts, supra, § 37, com. c, p. 3.) Rather, it is “whether the actor’s entire conduct created a risk of harm.” (Ibid.) The no-duty-to-protect rule is not absolute, however; this court has recognized a number of exceptions. (Delgado, supra, 36 Cal.4th at p. 235.) Under some circumstances, a defendant may have an affirmative duty to protect the plaintiff from harm at the hands of a third party, even though the risk of harm is not of the defendant’s own making. (Ibid.; see also Rest.3d Torts, supra, § 37.) For example, if a person does choose to “undertake[] to come to the aid of another,” she may then have an affirmative duty to exercise reasonable care in that undertaking. (Williams, supra, 34 Cal.3d at p. 23 [describing the negligent undertaking doctrine]; see, e.g., Paz v. State of California (2000) 22 Cal.4th 550, 559; Rest.3d Torts, supra, § 42.) We here focus, along with the parties, on another basis for finding an affirmative duty: In a case involving harm caused by a third party, a person may have an affirmative duty to protect the victim of another’s harm if that person is in what the law calls a “special relationship” with either the victim or the person who created the harm. (See, e.g., Regents, supra, 4 Cal.5th at pp. 619–620; Delgado, at p. 235; Williams, at p. 23; see generally Rest.3d Torts, supra, §§ 40, 41.)7

7 This is not an exhaustive list. An affirmative duty to protect may also arise if, for example, the Legislature imposes one by statute. (See Hoff v. Vacaville Unified School Dist. (1998) 19 Cal.4th 925, 938; Rest.3d Torts, supra, § 38.) Regardless of whether there is a basis for recognizing an affirmative duty, the no-duty-to-protect rule will not relieve the defendant of an otherwise applicable duty to exercise reasonable care when, by its own conduct, the defendant has increased the risk of harm to the plaintiff. (See, e.g., Kesner v. Superior Court (2016) 1 Cal.5th 1132, 1163 [“Although we have held that the existence of a relationship between the plaintiff and the defendant is one A special relationship between the defendant and the victim is one that “gives the victim a right to expect” protection from the defendant, while a special relationship between the defendant and the dangerous third party is one that “entails an ability to control [the third party’s] conduct.” (Regents, supra, 4 Cal.5th at p. 619.) Relationships between parents and children, colleges and students, employers and employees, common carriers and passengers, and innkeepers and guests, are all examples of special relationships that give rise to an affirmative duty to protect. (Id. at pp. 619–620; see Rest.3d Torts, supra, §§ 40–41.) The existence of such a special relationship puts the defendant in a unique position to protect the plaintiff from injury. The law requires the defendant to use this position accordingly. (See, e.g., Rest.3d Torts, supra, § 40, com. h, pp. 42–43.) Where the defendant has neither performed an act that increases the risk of injury to the plaintiff nor sits in a relation to the parties that creates an affirmative duty to protect the plaintiff from harm, however, our cases have uniformly held the defendant owes no legal duty to the plaintiff. Our decision in Williams, supra, 34 Cal.3d 18, is illustrative. The question there was whether highway patrol officers who had aided an injured motorist after an accident had a duty to investigate, secure information, or preserve evidence for the motorist to use in later civil litigation against the driver who caused her injury. (Id. at p. 21.) We began our analysis by reciting the general rule that

basis for finding liability premised on the conduct of a third party [citations], we have never held that such a relationship is a prerequisite to finding that a defendant had a duty to prevent injuries due to its own conduct or possessory control.”].) one has no duty to come to the aid of another. (Id. at pp. 23–24.) We then went on to consider whether the special relationship or negligent undertaking exceptions to the rule applied. Answering that question in the negative, we concluded the officers owed no duty to assist the motorist in preserving evidence. (Id. at pp. 27–28.) Other cases are to similar effect. (See, e.g., Davidson v. City of Westminster (1982) 32 Cal.3d 197, 203–209 [officers had no duty to protect victim of assault because they had not increased the risk of harm to the victim, they had no special relationship with the assailant or the victim, and they had not invited the plaintiff to depend on their protection]; Zelig v. County of Los Angeles (2002) 27 Cal.4th 1112, 1130 [officers had no duty to protect courthouse visitors from a third party assailant where officers had not increased the risk of harm to the victim, had no special relationship with either party, and had not given the victim a false sense of security by inviting her to depend on special protection].) B. Brown argues for a different approach to the duty to protect. She argues that even if the defendant lacks any special relationship with the parties and there are no other circumstances that would give rise to an affirmative duty to protect, such a duty may nonetheless arise after considering the policy factors set out in the landmark decision in Rowland, supra, 69 Cal.2d 108. We reject the argument. The multifactor test set forth in Rowland was not designed as a freestanding means of establishing duty, but instead as a means for deciding whether to limit a duty derived from other sources. The specific question in Rowland concerned the relationship between the common law duties of landowners and the general duty of care codified in section 1714. At common law, a landowner’s duty of care to his or her visitors varied based on the type of visitor. While landowners owed invitees an ordinary duty of care to maintain the premises in a safe condition, they generally owed trespassers and licensees only a duty to refrain from willful injury. (Rowland, supra, 69 Cal.2d at p. 114.) We held, however, that these “rigid common law classifications” were incompatible with California law. (Id. at p. 118.) We explained that “the basic policy of this state set forth by the Legislature in section 1714 . . . is that everyone is responsible for an injury caused to another by his want of ordinary care or skill in the management of his property.” (Rowland, at pp. 118–119.) And while there are exceptions to section 1714’s general rule, “in the absence of [a] statutory provision declaring an exception . . . , no such exception should be made unless clearly supported by public policy.” (Rowland, at p. 112.) In the passage of Rowland that has now become a touchstone of our negligence jurisprudence, we summarized the policy considerations that guide the inquiry. To depart from the general principle that all persons owe a duty of care to avoid injuring others, we explained, “involves the balancing of a number of considerations”: “the foreseeability of harm to the plaintiff, the degree of certainty that the plaintiff suffered injury, the closeness of the connection between the defendant’s conduct and the injury suffered, the moral blame attached to the defendant’s conduct, the policy of preventing future harm, the extent of the burden to the defendant and consequences to the community of imposing a duty to exercise care with resulting liability for breach, and the availability, cost, and prevalence of insurance for the risk involved.” (Rowland, supra, 69 Cal.2d at pp. 112–113.) We reasoned that while the common law categories of landowner duties might align with some of these considerations in some cases, they did not align in every case. It followed that a victim’s status as a trespasser, licensee, or invitee cannot be determinative of a landowner’s duties. The inquiry whether a landowner owes a duty to her invitees instead begins with the “basic policy” that “everyone is responsible for an injury caused to another by his want of ordinary care or skill in the management of his property,” and then considers whether more particular considerations of policy call for departure from the basic rule. (Id. at pp. 118–119.) Rowland itself referred to this multifactor test as a guide for determining whether to recognize an “exception” to the general duty of care under section 1714. (Rowland, supra, 69 Cal.2d at p. 113.) And in numerous cases since Rowland, we have repeated that the Rowland factors serve to determine whether an exception to section 1714’s general duty of reasonable care is warranted, not to determine whether a “ ‘new duty’ ” should be created. (Kesner v. Superior Court, supra, 1 Cal.5th at p. 1143 [“Because Civil Code section 1714 establishes a general duty to exercise ordinary care in one’s activities, . . . we rely on these factors not to determine ‘whether a new duty should be created, but whether an exception to Civil Code section 1714 . . . should be created,’ ” quoting Cabral, supra, 51 Cal.4th at p. 783]; accord, e.g., Gas Leak Cases, supra, 7 Cal.5th at p. 398 [“[Under section 1714], we presume the defendant owed the plaintiff a duty of care and then ask whether the circumstances ‘justify a departure’ from that usual presumption.”]; Ballard v. Uribe (1986) 41 Cal.3d 564, 572, fn. 6 [similar].) Although Rowland itself concerned the general duty of care in section 1714, we have also routinely applied Rowland to consider whether to recognize exceptions to affirmative duties to protect or warn. For example, in our recent decision in Regents, supra, 4 Cal.5th at page 627, we held that the special relationship between a university and its students creates a “duty to use reasonable care to protect their students from foreseeable acts of violence in the classroom or during curricular activities.” We then turned to Rowland to decide whether policy considerations “justif[ied] excusing or limiting” that duty to protect, and concluded the answer was no. (Regents, at p. 628; see id. at pp. 628–634.) This is but one example of many; a long line of cases before Regents had taken the same approach. (C.A. v. William S. Hart Union High School Dist. (2012) 53 Cal.4th 861, 877 (William S. Hart) [finding special relationship between school employees and students and then analyzing Rowland factors to determine “[a]dditional limits” on the “scope of the duty implicated in this and similar cases”]; id. at pp. 869–871, 877–879; Castaneda v. Olsher (2007) 41 Cal.4th 1205, 1213 [finding affirmative duty based on special relationship between landlord and tenants and then analyzing Rowland factors to determine “duty’s existence and scope”]; Delgado, supra, 36 Cal.4th at p. 244 [finding special relationship between business proprietor and its tenants, patrons, and invitees imposed general duty on proprietor to take “ ‘reasonable steps to secure common areas against foreseeable criminal acts of third parties’ ” and then analyzing Rowland factors to determine scope of duty]; id. at pp. 235–236, 244–247; Morris v. De La Torre (2005) 36 Cal.4th 260, 269, 271–272, 276–277 [same].) The cases recognize that even when two parties may be in a special relationship, the unforeseeability of the kind of harm suffered by the plaintiff or other policy factors may counsel against establishing an affirmative duty for one party to protect the other. Notwithstanding this considerable body of case law, Brown points to our decision in Nally v. Grace Community Church, supra, 47 Cal.3d 278 as evidence that we have sometimes treated Rowland as an independent source of duty, and not merely as a guide to whether to create an exception to a duty otherwise established. In that case, parents of a suicide victim sued the victim’s nontherapist church counselors for failure to protect the victim from suicide. (Nally, at p. 292.) We held the nontherapist counselors had no duty to protect, consistent with the traditional rule that “one is ordinarily not liable for the actions of another and is under no duty to protect another from harm, in the absence of a special relationship of custody or control.” (Id. at p. 293.) We next held there was no such special relationship involved in the case. (Id. at p. 296.) But we then considered the Rowland factors, in belt-and- suspenders fashion, to “explain further why we should not impose a duty to prevent suicide on defendants and other nontherapist counselors.” (Ibid.) As Brown notes, some Courts of Appeal have understood Nally to mean that the Rowland factors and the special relationship test are both sources of duty. (See, e.g., Adams, supra, 68 Cal.App.4th at p. 267.) This understanding is mistaken. Nally never squarely addressed the proper role of Rowland in its analysis, nor did it purport to qualify or limit the considerable body of case law explaining that Rowland is a guide to determining when to create exceptions from duties otherwise established. And as Brown acknowledged at oral argument, neither Nally nor any other decision of this court has done what Brown now asks us to do: rely solely on the Rowland factors to create a duty to take action to protect the plaintiff from third party harm.8

8 Although Brown has not raised the point in her briefing, we acknowledge that certain language in other cases could be read as suggesting such an approach. For example, in Palma v. U.S. Industrial Fasteners, Inc. (1984) 36 Cal.3d 171, we said of a suit seeking to impose liability for leaving a commercial truck unsecured in a high-crime area, where it was stolen and later used to cause injury: “[O]rdinarily in the absence of a special relationship there is no duty to control the conduct of a third person so as to prevent him from harming another, but where special circumstances exist in which there is ‘a greater potentiality of foreseeable risk or more serious injury, or [which] require a lesser burden of preventative action,’ the risk is deemed unreasonable and imposes a duty to third persons.” (Id. at pp. 184–185, quoting Hergenrether v. East (1964) 61 Cal.2d 440, 444.) But as the broader context of Palma and other related cases makes clear, the focus of the duty inquiry in these cases is not on the defendant’s duty to protect the victim from the conduct of a third party, but instead on the defendant’s general duty under section 1714 to exercise due care in his or her own conduct. While a car owner ordinarily cannot be held liable simply for allowing her car to be stolen and used for harm (Cabral, supra, 51 Cal.4th at p. 779), in some cases, like Palma, the defendant’s decision to leave a vehicle unguarded does increase the risks the vehicle will be harmfully misused. (See Rest.3d Torts, supra, § 19, reporters’ note, com. c, p. 222 [citing vehicle-theft cases to illustrate the proposition that “[i]f the third party’s misconduct is among the risks making the defendant’s conduct negligent, then ordinarily plaintiff’s harm will be within the defendant’s scope of liability” (id., com. c, p. 216)].) Such cases “can be contrasted to cases in which the defendant merely takes no action to protect the plaintiff against the possibility of third-party misconduct” — which is Brown’s theory of liability in asserting the existence of a special relationship. (Rest.3d Torts, supra, § 19, com. e, p. 218; see also Precedent aside, Brown argues we should now take that step in order to establish what she terms a “more flexible and holistic approach to duty,” particularly in cases involving minor victims of sexual abuse. This approach would allow courts to make a case-by-case policy judgment under Rowland as to whether a defendant may be held liable for failing to protect the victim of harm caused by another, even if the defendant were in no special position to control the wrongdoer or to offer protection to the victim, and there were no other circumstances giving rise to an affirmative duty to take action. Without denying the gravity of the injuries these plaintiffs suffered, nor the broader problem of sexual abuse of minors in organized youth sports and other activities, we decline Brown’s invitation to take that step. The requirement of an affirmative duty to protect itself embodies a policy judgment of considerable standing: A defendant cannot be held liable in negligence for harms it did not cause unless there are special circumstances — such as a special relationship to the parties — that give the defendant a special obligation to offer protection or assistance. This rule reflects a long-standing balance between several competing interests. It avoids difficult questions about how to measure the legal liability of the stranger who fails to take affirmative steps to prevent foreseeable harm, instead leaving the stranger to make his or her own choices about what assistance to offer. (See pp. 9–10, ante.) At the same time, it extends a right of recovery to individuals in relationships involving dependence or control, and who by virtue of those relationships have reason to expect the defendant’s protection.

id., illus. 1–3, pp. 218–219 [discussing circumstances where foreseeable risk makes a defendant’s conduct negligent].) (See Regents, supra, 4 Cal.5th at p. 621 [“ ‘[A] typical setting for the recognition of a special relationship is where “the plaintiff is particularly vulnerable and dependent upon the defendant who, correspondingly, has some control over the plaintiff’s welfare.” ’ ”].) Where such a special relationship exists between the defendant and a minor, the obligation to provide such protection and assistance may include a duty to protect the minor from third party abuse. (See, e.g., William S. Hart, supra, 53 Cal.4th at pp. 869–872, 879 [imposing such a duty]; Pamela L. v. Farmer (1980) 112 Cal.App.3d 206, 211 [same].) And there may be other circumstances that give rise to a comparable affirmative duty to protect. (Kesner v. Superior Court, supra, 1 Cal.5th at p. 1163.) But where no such circumstances exist, the Rowland factors do not serve as an alternative basis for imposing duties to protect. The purpose of the Rowland factors is to determine whether the relevant circumstances warrant limiting a duty already established, not to recognize legal duties in new contexts. (See Rowland, supra, 69 Cal.2d at p. 113; see also, e.g., Regents, supra, 4 Cal.5th at p. 628; cf. Rest.3d Torts, supra, § 40, com. b, p. 40 [“Even though an affirmative duty might exist pursuant to this Section, a court may decide, based on special problems of principle or policy, that no duty or a duty other than reasonable care exists.”].) The question remains whether Rowland has any role to play at all in cases concerning affirmative duties to protect. As noted, some courts have suggested the answer is no; that the special relationship test essentially encompasses the policy considerations set out in Rowland and renders it unnecessary to give separate consideration to the Rowland factors in determining whether to recognize a legal duty to protect. (Compare Hansra, supra, 7 Cal.App.4th at p. 646 [considering both tests together as one] and Titus, supra, 118 Cal.App.4th at pp. 911–912 [same] with Adams, supra, 68 Cal.App.4th at pp. 267–276, 282–288 [recognizing differences between them].) The suggestion is incorrect. While the Rowland factors do overlap to some degree with the considerations that determine the existence of a special relationship, application of one test does not obviate the need for the other. This is because the two tests operate differently. A court considers whether the parties have a special relationship by considering the particular facts and circumstances of their association with one another. The Rowland factors, by contrast, consider, “at a relatively broad level of factual generality,” whether policy considerations justify limiting any resulting duty of protection. (Cabral, supra, 51 Cal.4th at p. 772; see Regents, supra, 4 Cal.5th at p. 629 [“In considering [the Rowland factors], we determine ‘not whether they support an exception to the general duty of reasonable care on the facts of the particular case before us, but whether carving out an entire category of cases from that general duty rule is justified by clear considerations of policy.’ ”].) Our opinion in Castaneda v. Olsher, supra, 41 Cal.4th 1205 is illustrative of the difference between the two inquiries. In that case, the plaintiff was injured as a bystander to a gang- related shooting at the mobilehome park where he lived. He sued his landlord for negligence. (Id. at pp. 1211–1212.) Although the parties were in a special relationship, we concluded that the landlord did not have a duty to “withhold rental units from those they believe to be gang members” in order to protect his other tenants. (Id. at p. 1216.) We reasoned that requiring as much would, as a general proposition, result in “arbitrary discrimination on the basis of race, ethnicity, family composition, dress and appearance, or reputation” — all in service of a tenant screening process that was unlikely to effectively prevent injuries like those the plaintiff had suffered. (Ibid.) Thus, despite the existence of a special relationship, imposing such a duty on landlords would not be “fair or workable,” nor would it be “consistent with our state’s public policy as a whole.” (Ibid.) In other cases, a court might conclude that duty should not be imposed because, for example, the type of harm the plaintiff suffered was unforeseeable, or because there was no moral blameworthiness associated with the defendant’s conduct, notwithstanding the defendant’s special relationship to the plaintiff. Put differently, even when a special relationship gives rise to an affirmative duty to protect, a court must still consider whether the policy considerations set out in Rowland warrant a departure from that duty in the relevant category of cases.

III.

Brown alleged that USAT and USOC acted negligently by failing to take steps to protect her from her coach’s abuse. To evaluate her claim, the Court of Appeal first asked whether a duty existed based on a special relationship. Concluding that USOC had no such relationship with Brown, the court ended its analysis with respect to that defendant. This approach was sound. And after concluding that USAT did have a special relationship with plaintiffs, the court went on to apply Rowland to determine whether to limit that potential duty — deciding the answer to that question was no. This, too, was the correct approach.9 The Court of Appeal’s judgment does not mark the end of the case. It affirms the trial court’s decision to dismiss one of several named defendants, USOC, for failure to adequately allege a special relationship giving rise to an affirmative duty to protect. Having concluded the Court of Appeal did not err by declining to apply the Rowland factors as an alternative source of duty, we now affirm the court’s judgment. On remand, Brown may continue to pursue her suit against USAT and the other remaining defendants.

KRUGER, J.

We Concur:

CANTIL-SAKAUYE, C. J.

CORRIGAN, J.

LIU, J.

CUÉLLAR, J.

GROBAN, J.

JENKINS, J.

9 We disapprove the following decisions to the extent they applied the Rowland factors as an alternative source of duty where defendant did not create the risk that resulted in plaintiff’s injuries: University of Southern California v. Superior Court, supra, 30 Cal.App.5th 429; Juarez v. Boy Scouts of America, Inc., supra, 81 Cal.App.4th 377; Adams v. City of Fremont, supra, 68 Cal.App.4th 243; Titus v. Canyon Lake Property Owners Assn., supra, 118 Cal.App.4th 906; Doe 1 v. City of Murrieta (2002) 102 Cal.App.4th 899. Concurring Opinion by Justice Cuéllar

The young women who are plaintiffs in this case achieved, in taekwondo, uncommon excellence. What these young women nonetheless experienced in the process is all too common: someone they knew, trusted and relied on — their credentialed taekwondo coach — betrayed their trust and sexually assaulted them. The majority opinion specifies how a court must consider certain presumptions and exceptions when resolving whether such plaintiffs have any right to recover from entities at whose events and facilities the wrongs occurred — in this case, the United States Olympic Committee (USOC). I write separately to explain how those presumptions and exceptions realize a fundamental substantive principle: In California, “[t]ort law” — the law of when and how individuals who have suffered harm may seek compensation for their injuries through private actions — “serves society’s interest in allocating risks and costs to those who can better prevent them, and it provides aggrieved parties with just compensation.” (Southern California Gas Leak Cases (2019) 7 Cal.5th 391, 394 (SoCal Gas).) The majority specifically holds that when a plaintiff argues defendant owes a duty based on a “special relationship,” the policy analysis we first discussed in Rowland v. Christian is only relevant to decide whether to limit that duty. (See maj. opn., ante, at pp. 15–19; Rowland v. Christian (1968) 69 Cal.2d 108, 112–113, 119 (Rowland).) As the majority opinion explains, that procedure — first decide whether policy considerations BROWN v. USA TAEKWONDO Cuéllar, J., concurring

support a duty based on a “special relationship” and then consider whether public policy “clearly require[s] an exception” (Regents of University of California v. Superior Court (2018) 4 Cal.5th 607, 628 (Regents)) — reflects our precedents and provides a consistent way for courts to answer the threshold duty question. I write separately to clarify how that procedure fits with and furthers our principles and priorities in tort law. First, we generally start by presuming everyone has a duty of reasonable care “ ‘in the management of his or her property or person’ ”; we limit it based on policy considerations like those in Rowland only in “a particular category of cases” and only if “ ‘ “clearly supported by public policy.” ’ ” (Kesner v. Superior Court (2016) 1 Cal.5th 1132, 1143, italics added (Kesner).) Second, in some cases, a defendant may argue that he or she was nothing more than a disengaged bystander — someone society recognizes as categorically outside the scope of any responsibility, having no material role in creating the risk of plaintiff’s harm and so no duty of reasonable care. Such cases of a named defendant whom society considers a true bystander are presumably rare. But the distinction between the putative bystander and the risk creator can be as subtle in principle as it is challenging to apply in many cases. So, we have developed a doctrinal mechanism to sidestep the distinction and make arguable edge cases more tractable: we allow a plaintiff to establish defendant owed a duty of reasonable care in virtue of a “special relationship” regardless of whether defendant contributed to the risk of plaintiff’s harm. As a matter of tradition, “recognizing” or “identifying” a “special relationship” is the label for weighing up those policy considerations our legal system treats as most relevant in such arguably ambiguous contexts. As the majority explains, once a court recognizes a duty based on a “special relationship,” it should also confirm that public policy doesn’t clearly support limiting the duty in a clearly defined category of cases. Third, ensuring that the duty inquiry remains focused at a relatively high level of generality on public policy preserves the proper balance between the court and the jury. And so, fourth, the procedure we reaffirm today — presuming a general duty of reasonable care or “recognizing” a “special relationship” before deciding whether public policy clearly supports limiting that duty in a category of cases — flexibly serves society’s interest in providing just compensation to aggrieved parties and allocating risks and costs to those who can better prevent them. I. A. At the core of California tort law is a rule born of common law judgments and reaffirmed in statute: “Everyone is responsible, not only for the result of his or her willful acts, but also for an injury occasioned to another by his or her want of ordinary care or skill in the management of his or her property or person . . . .” (Civ. Code, § 1714, subd. (a).)1 This is the Legislature’s “conclusory expression[]” that, as “legal duties are not discoverable facts of nature,” generally speaking, “liability should be imposed for damage done.” (Tarasoff v. Regents of University of California (1976) 17 Cal.3d 425, 434 (Tarasoff).) For “injur[ies] occasioned” to others, this principle cuts — other things being equal — in favor of widespread liability. (§ 1714, subd. (a).)

1 All further statutory references are to the Civil Code. Of course, other things aren’t always equal. In Rowland, we recognized that the simple statutory presumption of a duty of reasonable care, rather than rigid common law categories, should generally guide our analysis of whether a defendant could be responsible at all. (See Rowland, supra, 69 Cal.2d 108, 118 [common law rules for landowner liability “obscure rather than illuminate the proper considerations which should govern determination of the question of duty”].) So, Rowland rejected a common law system that placed great weight on subtle, perhaps vanishing doctrinal distinctions without obvious practical or moral significance. (See id. at p. 119 [“we are satisfied that continued adherence to the common law distinctions can only lead to injustice or, if we are to avoid injustice, further fictions with resulting complexity and confusion”].) It replaced that system with a focus on the relevant consequences. (See id. at pp. 112–113, 117–119.) As we’d conveyed earlier that year, duty is just “ ‘an expression of the sum total of those considerations of policy which lead the law to say that the particular plaintiff is entitled to protection.’ ” (Dillon v. Legg (1968) 68 Cal.2d 728, 734.) That focus is how courts should determine when other things aren’t equal: presume a general duty of reasonable care, as described in section 1714, and create an exception only if “clearly supported by public policy.”2 (Rowland, at p. 112.)

2 Rowland identified a list of such considerations that would often be relevant. (Rowland, supra, 69 Cal.2d at pp. 112–113.) We have recognized, though, that the “inquiry hinges not on mere rote application of the[] separate so-called Rowland factors, but instead on a comprehensive look at the ‘ “sum total” ’ of the policy considerations at play in the context before us.” (SoCal Gas, supra, 7 Cal.5th 391, 399.) When we decide on such exceptions, we endeavor to take account of reasonable inferences about social burdens and benefits — or “policy considerations.” We do so in relatively general terms to ensure that public policy as it applies to a certain broad class of situations with sufficiently common features, rather than the bespoke details of any particular case, supports a clearly defined departure from the general principle that “a person is liable for injuries caused by his failure to exercise reasonable care in the circumstances.” (Rowland, supra, 69 Cal.2d at p. 112; see also T.H. v. Novartis Pharmaceuticals Corp. (2017) 4 Cal.5th 145, 165 (Novartis); Kesner, supra, 1 Cal.5th 1132, 1143–1144; Cabral v. Ralphs Grocery Co. (2011) 51 Cal.4th 764, 772 (Cabral); Rest.3d Torts, Liability for Physical and Emotional Harm (2012) § 7, com. a (Rest.3d) [“when liability depends on factors applicable to categories of actors or patterns of conduct, the appropriate rubric is duty” (italics added)].) Against this backdrop, a court-imposed limitation on a duty of care is appropriate “only when a court can promulgate relatively clear, categorical, bright-line rules of law applicable to a general class of cases.” (Rest.3d, supra, § 7, com. a; see also Kesner, supra, 1 Cal.5th 1132, 1143 [“The conclusion that a defendant did not have a duty constitutes a determination by the court that public policy concerns outweigh, for a particular category of cases, the broad principle enacted by the Legislature . . . ” (italics added)].) And likewise under Rowland and our subsequent decisions, the relevant policy considerations primarily relate to the social cost (or benefit) of recognizing a duty in a category of cases, whether society would be worse off for having a particular class of defendants potentially liable. (See Novartis, supra, 4 Cal.5th at p. 168 [we limit duty “ ‘where the social utility of the activity concerned is so great, and avoidance of the injuries so burdensome to society, as to outweigh the compensatory and cost-internalization values of negligence liability’ ” (italics added)]; see also, e.g., Vasilenko v. Grace Family Church (2017) 3 Cal.5th 1077, 1092, 1096; Castaneda v. Olsher (2007) 41 Cal.4th 1205, 1216–1218 (Castaneda).) In deciding whether to limit the general duty, our cases, starting with Rowland, have repeatedly emphasized defendants’ reasonable ability to anticipate a particular kind of harm. A court might limit the duty of some category of defendants who have no way to anticipate or avoid a category of harm, ensuring responsibility falls on those who can. (See, e.g., Novartis, supra, 4 Cal.5th 145, 166–167; see also Tarasoff, supra, 17 Cal.3d at pp. 434–435.) As a matter of public policy, we seek to shift losses to those most able to spread the loss or prevent the kind of harm in question — doing so reduces the number of injuries and the costs of reducing the number of injuries. (See, e.g., Kesner, supra, 1 Cal.5th at p. 1153.) And of course, such considerations may cut in the other direction and simply support the usual duty of reasonable care in a certain category of cases. (See, e.g., Peterson v. San Francisco Community College Dist. (1984) 36 Cal.3d 799, 807–809 (Peterson); Palma v. U.S. Industrial Fasteners, Inc. (1984) 36 Cal.3d 171, 184–185 (Palma).) B. Sometimes defendants may assert that they were mere bystanders to the risk of plaintiff’s harm — that the risk had nothing to do with defendants’ “activities” and so defendants had no duty to exercise “reasonable care for the safety of others.” (Cabral, supra, 51 Cal.4th at p. 768.) Arguments that sound in this key are particularly relevant in this case. In some cases, defendant may, of course, just be wrong. Whether someone is a “bystander” to a particular risk is shorthand for whether we as a society are willing to say, based on reasonable inferences about the benefits and burdens of potential liability, that a particular defendant should bear no responsibility for the risk. Though the question isn’t before us in this case, it bears emphasis, given the all-too-common fact pattern, that USOC is hardly a bystander to plaintiffs’ harm. USOC is the organizer of the activity where the harm occurred. Between the organizer of an activity where someone is wronged and a mere bystander there is generally a world of difference: one at least sets the stage for what ends up becoming a tragedy; the other at most stumbles into the theater in the last act, when the story has unfolded and its casualties are known. The person who sets the stage owes the players a general duty to exercise reasonable care. (See Lugtu v. Cal. Highway Patrol (2001) 26 Cal.4th 703, 716 [“one’s general duty to exercise due care includes the duty not to place another person in a situation in which the other person is exposed to an unreasonable risk of harm through the reasonably foreseeable conduct (including the reasonably foreseeable negligent conduct) of a third person”].) The organizer, by bringing people together, may “creat[e] the risk” even if less directly than a criminal or intentional tortfeasor. (Rest.3d, supra, § 40, com. c.; cf. Williams v. State of California (1983) 34 Cal.3d 18, 23.) Considering the benefits and burdens of imposing some responsibility on that person, a court may decide that an organizer — even one lacking either awareness of a material risk or full control of every administrative nuance — is no bystander at all. But in difficult cases, plaintiffs may argue that defendants owed them a duty of care in virtue of a “special relationship.” (See maj. opn., ante, at pp. 11–12.) In deciding whether defendant made plaintiff worse off, courts need not envision a world where defendant never existed. Instead they can conclude that: “[r]egardless of whether the actor played any role in the creation of the risk, a special relationship with others imposes a duty of reasonable care.” 3 (Rest.3d, supra, § 40, com. c, italics added; see also Regents, supra, 4 Cal.5th at pp. 619–620; Tarasoff, supra, 17 Cal.3d at p. 435.) “Special relationship” is merely a label for those policy considerations that our shared experience has taught us to treat as especially relevant in such contexts. As the Restatement explains, “The term ‘special relationship’ has no independent significance. . . . Whether a relationship is deemed special is a conclusion based on reasons of principle or policy.” (Rest.3d, supra, § 40, com. h.) Among reasons of principle or policy, our precedents place special emphasis on two reasons in particular: defendant’s ability to control the environment, to predict and prevent the risk, and plaintiff’s reasonable dependency. A college or university may be well suited to foresee and control risks to students in the campus environment, and students

3 Courts may have sometimes suggested that a “special relationship” is relevant when defendant is engaged in “nonfeasance.” (See maj. opn., ante, at pp. 9–10 & fn. 6.) But our reference today to the confused and confusing “misfeasance”/“nonfeasance” distinction is just an acknowledgement of a now outmoded oddity. (See, e.g., Kesner, supra, 1 Cal.5th at p. 1163; see also Sommer v. Federal Signal Corp. (1992) 583 N.Y.S.2d 957, 961–962; Rest.3d, supra, § 37, com. c.; Abraham & Kendrick, There’s No Such Thing as Affirmative Duty (2019) 104 Iowa L.Rev. 1649, 1682–1685.) reasonably expect such protection and are especially vulnerable without it. (See, e.g., Regents, supra, 4 Cal.5th at p. 625; see also Peterson, supra, 36 Cal.3d 799, 807–809, 813–814.) Or a proprietor should be aware of a danger of assault; customers are at the mercy of the proprietor; and so the proprietor has a duty of reasonable care to reduce such risks. (See, e.g., Morris v. De La Torre (2005) 36 Cal.4th 260, 270; Delgado v. Trax Bar & Grill (2005) 36 Cal.4th 224, 241–242.) In some cases, we have also emphasized the burden on defendant of avoiding certain risks. (See Castaneda, supra, 41 Cal.4th at p. 1213.) This focus on foreseeability and defendant’s burden tracks defendant’s particular ability to reduce risk; the focus on reasonable dependence tracks the value society places on reducing that risk. As the majority explains, there is a second step when considering a duty based on a “special relationship” — whether public policy supports limiting the duty. (Maj. opn., ante, at pp. 20, 22.) As when we consider the general duty of reasonable care, the analysis of public policy rationales occurs at a relatively high level of generality. (See id. at p. 21.) But identifying a “special relationship” already means that certain policy reasons — especially defendant’s ability to reduce the risk in question and blameworthiness for failing to do so — favor requiring reasonable care. Those reasons are not likely to justify excusing from liability a category of defendants that includes the particular defendant. The “special relationship” analysis determines that any such category, whatever its precise parameters, would exclude at least the particular defendant. Instead, primarily the undesirable consequences, the social cost, of holding a category of defendants liable for a category of risk would support limiting defendant’s duty if anything would. (See, e.g., Castaneda, supra, 41 Cal.4th at p. 1216 [no duty because requiring landlords to screen tenants for gang affiliations would lead to discrimination].) C. While many of our previous decisions focus on duty, they readily convey that analyzing duty is just one part of the negligence inquiry. A duty’s existence does not determine whether defendant is liable and to what extent. (See Regents, supra, 4 Cal.5th at p. 634.) Most liability questions are case- specific and so not amenable to analysis in terms of duty — they do not allow a categorical determination whether defendant had to exercise reasonable care at all. (See Cabral, supra, 51 Cal.4th at pp. 772–774.) Unlike duty, the remaining liability questions — breach as well as factual and legal causation — are usually questions for the jury. What counts as reasonable care in a specific case, for instance, is characteristically a question of breach. (See Cabral, supra, 51 Cal.4th at p. 773.) As a policy matter, we tend to leave questions of breach to the jury as the institutional actor best situated to express, in a particular case, society’s judgment of whether the particular cost of avoiding a particular injury outweighs the particular cost of the injury. (See, e.g., Dobbs et al., The Law of Torts (2d ed. 2020) §§ 21, 161; cf. Calabresi, Concerning Cause and the Law of Torts: An Essay for Harry Kalven Jr. (1975) 43 U. Chi. L.Rev. 69, 75–76.) As we and other courts have long recognized, that judgment is the heart of what courts ask in assessing negligence. (See, e.g., United States v. Carroll Towing Co. (2d Cir. 1947) 159 F.2d 169, 173; Crane v. Smith (1944) 23 Cal.2d 288, 298 [“ ‘Where an act is one which a reasonable man would recognize as involving a risk of harm to another, the risk is unreasonable and the act is negligent if the risk is of such magnitude as to outweigh what the law regards as the utility of the act or of the particular manner in which it is done’ ”].)

II.

A. We granted review for one purpose: to clarify what steps a court should take when deciding whether a duty based on a “special relationship” exists. To provide that clarification, the majority restates the two-step process we endorsed in Rowland and have regularly followed since — start from a general duty of reasonable care based on section 1714 or a “special relationship” and then decide whether public policy requires limiting it in a clearly defined category of cases. (See maj. opn., ante, at pp. 13–17; see also, e.g., Kesner, supra, 1 Cal.5th at pp. 1143–1144.) Through that process, California tort law structures a court’s threshold decision about how potential liability affects society’s interests even as it also embodies and preserves a degree of flexibility. Maintaining the balance between structure and flexibility, guided by important, longstanding values — allocating risks and costs to those who can avoid them, and ensuring just compensation — is critical to making tort law both relevant and useful. The following scenario shows how. A plaintiff alleges that a youth organization did not exercise reasonable care leading to a program leader molesting him. (See Juarez v. Boy Scouts of America, Inc. (2000) 81 Cal.App.4th 377, 385–386 (Juarez).) The defendant organization makes six arguments that it should not have a duty of reasonable care. First, it argues that it was merely a bystander. Second, it argues that organizations like it are not generally well-situated to predict and prevent molestation. (See id. at pp. 408–409.) Third, it is specifically not well-situated to predict and prevent molestation. (See ibid.) Fourth, requiring organizations like it to prevent molestation would divert funds from youth programs and charitable enterprises thus harming society. (See id. at p. 409.) Fifth, it had procedures in place to prevent molestation. (See id. at pp. 405–406.) And sixth, the likelihood of abuse was low, so plaintiff’s molestation was not foreseeable. (See id. at pp. 403– 404.) Of course, the court may reject defendant’s “bystander” argument out of hand. If it does, the court should start with defendant’s general duty of reasonable care. And then the court considers whether large organizations’ purported inability to prevent molestation and the risk of charities diverting funds “ ‘clearly support[]’ ” limiting or eliminating the duty in an “entire category of cases.” (Cabral, supra, 51 Cal.4th at pp. 771, 772.) To assess those arguments, the court will need to determine whether “large organizations” or “charities” form a clearly defined category. But those arguments are the right kind of reason, at the right level of generality, to consider as a basis for limiting a duty. And so it is a question for the court whether, in light of any and all other policy considerations, those reasons are sufficiently substantial to support limiting a duty for charitable organizations, if that category can be clearly defined. On the other hand, the court could worry that defendant may be no more than a bystander. Perhaps the organization has an educational mission but primarily licenses its name to local chapters. Nonetheless, the court may decide defendant is in fact well-situated to prevent molestation, notwithstanding defendant’s argument to the contrary. For example, the court might reason that children in the local chapters are especially vulnerable because they are children, and the organization can control local chapters’ activities through the licensing process. So, defendant should exercise reasonable care to limit the risk of molestation regardless of whether it “created that risk.” In drawing that inference, the court “recognizes” defendant has a “special relationship” with plaintiff. The court then proceeds to a second step: deciding whether public policy “clearly supports” limiting the duty of reasonable care in a category of cases. At this step in the analysis, large organizations’ purported inability to avoid molestation can have little relevance — whatever may be true of other large organizations, the court has already rejected the argument as to this organization. But it remains for the court, in light of all other relevant public policy considerations, to decide whether concerns about charitable organizations diverting funds “clearly require an exception.” (Regents, supra, 4 Cal.5th at p. 628.) In both these versions of the scenario, the court trains attention on reasons of public policy when deciding whether to limit a duty. It does so by considering possible consequences at a high level of generality and with an eye to the loss to society from imposing liability. A court can more effectively focus at the appropriate level of generality if it considers policy reasons to limit a duty after presuming a duty or recognizing the duty of a specific defendant based on a “special relationship.” (See maj. opn., ante, at pp. 21–22.) And the requirement that such reasons clearly support limiting a duty of reasonable care ensures that courts act for good reason and not based on idle speculation. By contrast, defendant’s other arguments — about specific measures it has already taken or its ability to predict that a specific harm would arise — bear simply on the details of the present case, and not on duty. (See Cabral, supra, 51 Cal.4th at p. 773.) The argument about having done enough concerns whether defendant in fact took reasonable care, a question of breach usually for the jury. (Id. at p. 772.) And the argument about specific foreseeability would be relevant to whether plaintiff had established proximate cause, also usually a question for the jury. (Id. at pp. 772–773.) Here too, the two-step process we endorse today serves an important function: it helps courts guard against inappropriately taking questions from the jury. At the stage of deciding whether to limit a duty, courts should not look to features of the specific case but to considerations of public policy, clearly defined. By first presuming a duty or recognizing a duty based on a “special relationship” and then deciding whether public policy clearly requires limiting that duty, courts focus at the right level of generality at the right time. And that focus serves as a check on courts wading into fact-specific questions of breach or causation. (Cabral, supra, 51 Cal.4th at pp. 772–773.) In addition to these clarifying functions, the procedure we approve today remains flexible. We don’t disapprove our prior precedents that quickly or silently presume or recognize a duty before focusing primarily or entirely on whether policy considerations support it. (See maj. opn., ante, at pp. 18–19, fn. 8 [discussing Palma, supra, 36 Cal.3d at pp. 184–185].) And while we disapprove several Court of Appeal cases, we don’t disapprove them to the extent they presume a general duty of reasonable care and find the policy considerations in Rowland support holding defendants to that duty. (See, e.g., Juarez, supra, 81 Cal.App.4th 377, 401–410.) This flexibility is rooted in tort law’s decidedly limber structure: a duty of reasonable care based on a “special relationship” may be functionally indistinguishable from the general duty of reasonable care. As the Restatement acknowledges, some special-relationship-type duties “overlap with the general duty of reasonable care”; they are “a specialized application” of that general duty. (Rest.3d, supra, § 40, com. h.) In cases where “the actor’s conduct might have played a role in creating the risk to the injured party” there is a general “duty of reasonable care” even without any sort of “special relationship.” (Id., § 40, com. c, italics added.) What we must discern is if society has defensible reasons to restrict liability in certain situations; otherwise, a person’s duty is to exercise reasonable care. Whether there’s a “special relationship” is a question that structures, without ever supplanting, the ultimate inquiry — guiding deliberation just enough to avoid turning every duty question into a fact-specific monologue about the defendant’s role in creating the risk. (See, e.g., Tarasoff, supra, 17 Cal.3d at pp. 434–435; Rest.3d, supra, § 40, com. h.) B. Properly understood, the special relationship question plays a limited but important role under our law. If plaintiffs don’t want to wade into whether defendants “created the risk” of the harm plaintiffs experienced, they can argue that defendant owed a duty of reasonable care, based on a “special relationship.” To do so, plaintiffs have to show why specific policy factors under the “special relationship” rubric supported such a duty. We don’t address whether plaintiffs in this case did in fact make such a showing. And we don’t address the possibility, because it was not presented to us, that USOC had a general duty to exercise reasonable care in the management of its property or person. (See § 1714, subd. (a).) It presumptively would if it were not a “bystander,” and we certainly don’t decide public policy clearly supports exempting a category of organizations including USOC from that general duty. (See generally House Energy & Commerce Committee, Nassar and Beyond: A Review of the Olympic Community’s Efforts to Protect Athletes from Sexual Abuse (Dec. 20, 2018).) While we don’t address those claims today, we reaffirm that courts should assess such claims with the ultimate aim of deciding whether requiring reasonable care serves the goals tort law embodies: to achieve appropriate deterrence and compensate victims. (See SoCal Gas, supra, 7 Cal.5th 391, 394; see also Rest.3d, supra, § 40, com. h.)

III.

Tort law relies heavily on the concept of duty to render tractable a reality where lives are at risk in the very world that sustains them, and people are bound by intricate and far- reaching ties of responsibility and norms of reciprocity. But “ ‘ “duty” is not sacrosanct in itself’ ”; it is a means to an end, “ ‘only an expression of the sum total of those considerations of policy which lead the law to say that the particular plaintiff is entitled to protection.’ ” (Dillon v. Legg, supra, 68 Cal.2d at p. 734.) The majority opinion today sensibly clarifies the procedure for recognizing a duty where plaintiff attempts, by arguing for a “special relationship,” to cut through the knot of whether defendant did or didn’t create a risk. Specifically, it holds that the Rowland factors as such just feature in deciding whether to limit a duty. In so doing, it reaffirms that under California law everyone presumptively owes a duty of reasonable care in the management of his or her property or person to avoid injuring others. And it doesn’t suggest that a corporate person’s duty of reasonable care in the management of its person or property extends any less than to the limits of foreseeable harm without substantial, concrete policy reasons to the contrary. The two-step procedure we endorse is grounded in long- established principles, emphasizing not only the importance of offering civil recourse and compensation to those harmed but also the value of allocating responsibility for losses to minimize future harm. We start from the premise that a duty of reasonable care ordinarily exists, whether arising generally or based on a “special relationship.” We then consider whether public policy requires limiting that duty in a clearly defined category of cases, assessed at the right level of generality. Implicit in this arrangement is the latitude for courts to minimize harms through the proper allocation of losses, and to compensate victims for their uncommon injuries — including in cases where the facts are not only tragic, but tragically all-too-common. Which is why I concur.

CUÉLLAR, J.

17 See next page for addresses and telephone numbers for counsel who argued in Supreme Court.

Name of Opinion Brown v. USA Taekwondo __________________________________________________________________________________

Unpublished Opinion Original Appeal Original Proceeding Review Granted XX 40 Cal.App.5th 1077 Rehearing Granted

__________________________________________________________________________________

Opinion No. S259216 Date Filed: April 1, 2021 __________________________________________________________________________________

Court: Superior County: Los Angeles Judge: Michael P. Vicencia

__________________________________________________________________________________

Counsel:

Estey & Bomberger, Stephen J. Estey; Corsiglia McMahon & Allard, B. Robert Allard; Turek Law, Kenneth C. Turek; Williams Iagmin and Jon R. Williams for Plaintiffs and Appellants.

Arbogast Law, David M. Arbogast; Siminou Appeals, Benjamin I. Siminou; Herzog, Yuhas, Ehrlich & Ardell, Ian I. Herzog; The Bronson Firm and Steven M. Bronson for Consumer Attorneys of California as Amicus Curiae on behalf of Plaintiffs and Appellants.

Esner, Chang & Boyer, Holly N. Boyer and Shea S. Murphy for National Crime Victim Bar Association and Manly, Stewart & Finaldi as Amici Curiae on behalf of Plaintiffs and Appellants.

Kjar, McKenna, Stockalper, Patrick E. Stockalper, Matthew A. Schiller and Mina M. Morkos; Horvitz & Levy, Mitchell C. Tilner, Steven S. Fleischman and Yen-Shyang Tseng for Defendant and Respondent USA Taekwondo.

Clyde & Co. US, Douglas J. Collodel, Margaret M. Holm, M. Christopher Hall; Covington & Burling, Beth Brinkmann, Mitch A. Kamin and Carolyn J. Kubota for Defendant and Respondent United States Olympic Committee.

Munger, Tolles & Olson, Donald B. Verrilli, Jr., Hailyn J. Chen and John B. Major for National Collegiate Athletic Association as Amicus Curiae on behalf of Defendant and Respondent United States Olympic Committee. Counsel who argued in Supreme Court (not intended for publication with opinion):

Jon R. Williams Williams Iagmin LLP 666 State Street San Diego, CA 92101 (619) 238-0370

Yen-Shyang Tseng Horvitz & Levy LLP 3601 W. Olive Ave., 8th Floor Burbank, CA 91505-4681 (818) 995-0800

Beth S. Brinkmann Covington & Burling LLP 1 CityCenter, 850 10th St. NW Washington, DC 20001 (202) 662-5312