*634 **883 *394 This case concerns a massive, months-long leak from a natural gas storage facility located just outside Los Angeles. According to the allegations before us, the accident severely harmed the economy of a nearby suburb. We must decide if local businesses - none of which allege they suffered personal injury or property damage - may recover in negligence for income *635 lost because of the leak. Our decision turns on whether the entity that allegedly caused the leak had a tort duty to guard against what we and other courts have termed "purely economic losses."

The businesses argue that they deserve compensation for such losses, that the entity responsible must bear the full costs of its alleged negligence so tort law can play its essential role of forcing people and organizations to take sufficient account of the risks they generate, and that courts can sensibly apportion liability under these circumstances within meaningful limits. Tort law indeed lies in the heartland of our common law system. It serves society's interest in allocating risks and costs to those who can better prevent them, and it provides aggrieved parties with just compensation. But a proper assessment of competing considerations in light of our precedent suggests, *395 and the extent of consensus across other jurisdictions confirms, that claims for purely economic losses suffered from mere proximity to an industrial accident create intractable line-drawing problems for courts. So the claims before us are best not treated as compensable in negligence.

We therefore affirm the judgment of the Court of Appeal.

Because this case comes to us at the demurrer stage, we take as true all properly pleaded material facts - but not conclusions of fact or law. ( Centinela Freeman Emergency Medical Associates v. Health Net of California, Inc. (2016) 1 Cal.5th 994 , 1010, 209 Cal.Rptr.3d 280 , 382 P.3d 1116 ( Centinela ).)

Near the northwestern corner of Los Angeles lies Porter Ranch, a residential neighborhood home to some 30,000 people. Southern California Gas Company (SoCalGas) stores vast amounts of natural gas in an underground facility in the hills surrounding the community. Known today as the "Aliso Facility," that subterranean storage site was once an oil reservoir. It was repurposed about 40 years ago for its present use. SoCalGas supplies over 21 million people with natural gas from its four storage facilities, but the Aliso Facility is the company's largest. It holds up to 80 billion cubic feet of natural gas, which SoCalGas pumps underground at high pressure into more than 100 "injection wells." Because natural gas is odorless, SoCalGas adds a nausea-causing chemical to the gas so that people notice when a leak happens.

In October 2015, a leak happened - and people noticed. An uncontrolled flow of natural gas from the Aliso Facility coated nearby neighborhoods in an oily mist. At its peak, the leak released some 55 tons of natural gas every hour. Porter Ranch residents reported unpleasant odors, headaches, dizziness, and respiratory problems. In addition to those symptoms, students at local schools complained of nosebleeds and vomiting.

That November, the Los Angeles County health department directed SoCalGas to establish a relocation program available to Porter Ranch residents who lived within a five-mile radius of the leak site. The Department of Conservation's Division of Oil, Gas, and Geothermal Resources required SoCalGas to provide real-time data about the leak, **884 as well as a timeline for stopping it. A month later, with the flow of gas slowing but still significant, the Los Angeles County Board of Education decided to relocate students and *396 staff from two Porter Ranch schools for the duration of the academic year. And a month after that, SoCalGas expanded its relocation program, *636 citing complaints of poor air quality from people living outside the initial five-mile boundary. About 15,000 people were relocated in total, scattering to locations dozens - and in some cases hundreds - of miles away.

SoCalGas finally got the leak under control in February 2016 - four months after detecting it. All told, about 100,000 tons of natural gas escaped the Aliso Facility, releasing enough greenhouse gases into the atmosphere to erase several years' worth of efforts to combat climate change in California.

Plaintiffs are Porter Ranch area businesses seeking to represent a class of "[a]ll persons and entities conducting business within five miles of the [Aliso] Facility from October 23, 2015 to [the] present." They allege that SoCalGas's negligence caused the leak. The resulting relocation of many Porter Ranch residents devastated the local economy: by depriving local businesses of customers, the environmental disaster cost local businesses considerable earnings.

That harm, Plaintiffs maintain, is ongoing. Sales at businesses of all stripes declined sharply, and in many cases, stayed down. Enrollment at a local martial arts center, Plaintiff King Taekwondo, nosedived during the leak and has not recovered. The same was true of a neighborhood day care, Plaintiff Polonsky Family Day Care. Restaurants, gas stations, and pharmacies were affected, too. So were beauty salons, doctor's offices, party suppliers, and a photography store.

With the en masse relocation of Porter Ranch residents and the diminution in property values caused by the leak, home mortgage lenders and home improvement businesses suffered economically as well. Plaintiff First American Realty saw clients get cold feet, loans fall out of escrow, and sales tumble. A local contractor's business dropped by 25 percent, as customers moved away or decided against home improvements for the time being. "Since the onset of the gas leak," in other words, business operations throughout Porter Ranch "have either halted or slowed substantially" - and Plaintiffs "have not yet recovered from the blow to their bottom lines."

Yet no named plaintiff in this action alleges personal injury or property damage. Accordingly, Plaintiffs acknowledge they are suing SoCalGas to recover solely for the income they lost because of the leak.

*397 C.

SoCalGas demurred, arguing that Plaintiffs' negligence claims failed as a matter of law because Plaintiffs were seeking to recover for purely economic losses. Overruling the demurrer, the trial court explained that companies "must face the full cost of accidents" they create, or else "they will underinvest in precautions." The trial court acknowledged that economic losses not flowing from conventional injury to person or property, such as physical damage, are ordinarily not recoverable in tort - and that the Court of Appeal had so held in Adams v. Southern Pacific Transportation Co. (1975) 50 Cal.App.3d 37 , 123 Cal.Rptr. 216 ( Adams ) on facts with some similarities to those here. But *637 the trial court questioned the wisdom of that rule and reasoned that Adams was no longer good law after our later decision in J'Aire Corp. v. Gregory (1979) 24 Cal.3d 799 , 157 Cal.Rptr. 407 , 598 P.2d 60 ( J'Aire ).

After SoCalGas petitioned for a writ of mandate, the Court of Appeal granted the petition and reversed the trial court. ( Southern California Gas Leak Cases (2017) 18 Cal.App.5th 581 , 583-584, 227 Cal.Rptr.3d 117 .) The Court of Appeal explained that, under California law, it is "not presumed"

**885 that a defendant owes a duty of care to guard against economic losses unaccompanied by injury to person or property. ( Id. at p. 591, 227 Cal.Rptr.3d 117 .) And without a "special relationship" between the plaintiff and the defendant stemming in this context from a "transaction," the Court of Appeal reasoned, California law did not permit recovery for the purely economic losses sought by Plaintiffs in this case. ( Id. at p. 591, 227 Cal.Rptr.3d 117 .) The Court of Appeal also took the view that our decision in J'Aire had not rejected Adams in its entirety but instead disapproved Adams only "insofar as [it] held a plaintiff can never recover purely economic losses based on a defendant's negligent conduct." ( Id. at p. 592, 227 Cal.Rptr.3d 117 , italics added.) Because Plaintiffs disclaimed any desire to further amend their complaint, the Court of Appeal directed the trial court to sustain SoCalGas's demurrer without leave to amend. ( Id . at p. 595, 227 Cal.Rptr.3d 117 .)

II.

Recovery in a negligence action depends as a threshold matter on whether the defendant had " 'a duty to use due care toward an interest of [the plaintiff's] that enjoys legal protection against unintentional invasion.' " ( Centinela , supra , 1 Cal.5th at p. 1012 , 209 Cal.Rptr.3d 280 , 382 P.3d 1116 , quoting *398 Bily v. Arthur Young & Co. (1992) 3 Cal.4th 370 , 397, 11 Cal.Rptr.2d 51 , 834 P.2d 745 ( Bily ).) We review de novo whether this " 'essential prerequisite' " to recovery is satisfied. ( Centinela , at pp. 1010, 1012, 209 Cal.Rptr.3d 280 , 382 P.3d 1116 .)

The issue here is whether SoCalGas - separate from other legal and practical reasons it had to prevent injury of any kind to the public - had a tort duty to guard against negligently causing what we and others have called "purely economic loss[es]." ( Centinela , supra , 1 Cal.5th at p. 1013 , 209 Cal.Rptr.3d 280 , 382 P.3d 1116 ; see also 532 Madison Avenue Gourmet Foods, Inc. v. Finlandia Center, Inc. (N.Y. 2001) 96 N.Y.2d 280 , 727 N.Y.S.2d 49 , 750 N.E.2d 1097 , 1102 ( 532 Madison ).) We use that term as a shorthand for "pecuniary or commercial loss that does not arise from actionable physical, emotional or reputational injury to persons or physical injury to property." (Dobbs, An Introduction to Non-Statutory Economic Loss Claims (2006) 48 Ariz. L.Rev. 713 (Dobbs).) And although SoCalGas of course had a tort duty to guard against the latter kinds of injury, we conclude it had no tort duty to guard against purely economic losses.

In California, the "general rule" is that people owe a duty of care to avoid causing harm to others and that they are *638 thus usually liable for injuries their negligence inflicts. ( Cabral v. Ralphs Grocery Co. (2011) 51 Cal.4th 764 , 771, 122 Cal.Rptr.3d 313 , 248 P.3d 1170 ( Cabral ).) Under Civil Code section 1714, subdivision (a), "[e]veryone 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, except so far as the latter has, willfully or by want of ordinary care, brought the injury upon himself or herself." So at least in cases involving traditionally compensable forms of injury - like physical harm to person or property - we presume the defendant owed the plaintiff a duty of care and then ask whether the circumstances "justify a departure" from that usual presumption. ( Cabral , at p. 771, 122 Cal.Rptr.3d 313 , 248 P.3d 1170 .) In Rowland v. Christian (1968) 69 Cal.2d 108 , 70 Cal.Rptr. 97 , 443 P.2d 561 ( Rowland ), we identified several factors that, among others, may bear on that question: (1) "the foreseeability of harm to the plaintiff," (2) "the degree of certainty that the plaintiff suffered injury," (3) "the closeness of the connection between the defendant's conduct and the injury suffered," (4) "the moral blame attached to the defendant's conduct," (5) "the policy of preventing future harm," (6) "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 (7) "the availability, cost, and prevalence of insurance for the *399 risk involved." ( **886 Id. at p. 113, 70 Cal.Rptr. 97 , 443 P.2d 561 .) At core, though, the inquiry hinges not on mere rote application of these separate so-called Rowland factors, but instead on a comprehensive look at the " 'the sum total' " of the policy considerations at play in the context before us. ( Parsons v. Crown Disposal Co. (1997) 15 Cal.4th 456 , 472, 63 Cal.Rptr.2d 291 , 936 P.2d 70 ( Parsons ), quoting Ballard v. Uribe (1986) 41 Cal.3d 564 , 572, fn. 6, 224 Cal.Rptr. 664 , 715 P.2d 624 ; see also T.H. v. Novartis Pharmaceuticals Corp. (2017) 4 Cal.5th 145 , 164, 226 Cal.Rptr.3d 336 , 407 P.3d 18 .)

What Civil Code section 1714 does not do is impose a presumptive duty of care to guard against any conceivable harm that a negligent act might cause. No one doubts, for example, that a child suffers gravely when an accident permanently disables her parent. But in Borer v. American Airlines, Inc. (1977) 19 Cal.3d 441 , 138 Cal.Rptr. 302 , 563 P.2d 858 ( Borer ), we nevertheless treated the prospect of a child recovering for loss of consortium in precisely that circumstance as "a wholly new cause of action," rather than a presumptively viable one. ( Id. at p. 447, 138 Cal.Rptr. 302 , 563 P.2d 858 .) And we refused to recognize such a novel - though quite sympathetic - claim for emotional harm largely because that claim, unlike a loss of consortium claim brought by a spouse, threatened indeterminate and disproportionate liability. ( Id. at pp. 448-449, 453, 138 Cal.Rptr. 302 , 563 P.2d 858 ; see also Thing v. La Chusa (1989) 48 Cal.3d 644 , 666-668, 257 Cal.Rptr. 865 , 771 P.2d 814 ( Thing ) [strictly cabining recovery for negligent infliction of emotional distress to ensure meaningful limits on liability].) Disabled parents, after all, have parents of their own, along with "brothers, sisters, cousins, inlaws, friends, colleagues, and other acquaintances who," in some way, may "be deprived of [their] companionship." ( Borer , at p. 446, 138 Cal.Rptr. 302 , 563 P.2d 858 .) In declining to impose a tort duty to guard against such harms, we noted in Borer the "overwhelming approval" our conclusion enjoyed in other jurisdictions and rejected the argument that our analysis was somehow "inconsistent with the principles of tort law" established in our own. ( Id. at pp. 449-450, 138 Cal.Rptr. 302 , 563 P.2d 858 ; see also *639 Thing , at p. 668, fn. 11, 257 Cal.Rptr. 865 , 771 P.2d 814 .)

Plaintiffs do cite several cases where we presumed the defendant owed the plaintiff a duty of care and then asked whether the circumstances warranted a departure from that baseline presumption. But unlike Borer and Thing , every one of those cases involved a traditionally compensable form of harm: personal injury. (See Vasilenko v. Grace Family Church (2017) 3 Cal.5th 1077 , 1082, 224 Cal.Rptr.3d 846 , 404 P.3d 1196 [plaintiff was struck by a car when crossing a public street shortly after leaving defendant's parking lot]; Kesner v. Superior Court (2016) 1 Cal.5th 1132 , 1140-1141, 210 Cal.Rptr.3d 283 , 384 P.3d 283 ( Kesner ) [employee's household members were exposed to asbestos, causing personal injury and death]; Cabral , supra , 51 Cal.4th at p. 769 , 122 Cal.Rptr.3d 313 , 248 P.3d 1170 [plaintiff's husband died after colliding with a truck owned by defendant]; John B. v. Superior Court (2006) 38 Cal.4th 1177 , 1181-1183, 45 Cal.Rptr.3d 316 , 137 P.3d 153 [defendant infected plaintiff with HIV];

*400 Parsons , supra , 15 Cal.4th at p. 460 , 63 Cal.Rptr.2d 291 , 936 P.2d 70 [defendant's truck startled plaintiff's horse, causing the plaintiff to fall to the ground].) And in Rowland itself, a faulty faucet in the defendant's home mangled tendons and nerves in the plaintiff's hand. ( Rowland , supra , 69 Cal.2d at p. 110 , 70 Cal.Rptr. 97 , 443 P.2d 561 .) So yes, we have frequently begun our analysis by presuming a duty of care. But we have not universally done so.

A case in point is liability in negligence for purely economic losses, which is "the exception, not the rule" under our precedents. ( Quelimane Co. v. Stewart Title Guaranty Co. (1998) 19 Cal.4th 26 , 58, 77 Cal.Rptr.2d 709 , 960 P.2d 513 ( Quelimane ).) And that holds true even though Civil Code section 1714 does not, by its terms, "distinguish among injuries to one's person, one's property or one's financial interests." ( J'Aire , supra , 24 Cal.3d at p. 806 , 157 Cal.Rptr. 407 , 598 P.2d 60 ; see Centinela , supra , 1 Cal.5th at p. 1013, 209 Cal.Rptr.3d 280 , 382 P.3d 1116 ; Dobbs, supra , 48 Ariz. L.Rev. at p. 713 [explaining that "[n]egligently inflicted economic loss that results from some other kind of injury may be recoverable, but recovery for stand-alone economic loss is frequently rejected"].)

**887 The primary exception to the general rule of no-recovery for negligently inflicted purely economic losses is where the plaintiff and the defendant have a "special relationship." ( J'Aire , supra , 24 Cal.3d at p. 804 , 157 Cal.Rptr. 407 , 598 P.2d 60 .) What we mean by special relationship is that the plaintiff was an intended beneficiary of a particular transaction but was harmed by the defendant's negligence in carrying it out. Take, for example, Biakanja v. Irving (1958) 49 Cal.2d 647 , 320 P.2d 16 ( Biakanja ). There, we held that the intended beneficiary of a will could recover for assets she would have received if the notary had not been negligent in preparing the document. ( Id. at pp. 650-651, 320 P.2d 16 .) A special relationship existed between the intended beneficiary and the notary in Biakanja , we emphasized, because "the 'end and aim' of the transaction" between the nonparty decedent and the notary was to ensure that the decedent's estate passed to the intended beneficiary. ( Id. at p. 650, 320 P.2d 16 .)

For similar reasons, in J'Aire we held that a special relationship existed between a restaurant operator and a contractor hired by a third-party property owner to renovate the space rented by the restaurant operator. ( J'Aire , supra , 24 Cal.3d at pp. 804-805, 157 Cal.Rptr. 407 , 598 P.2d 60 .) So when the contractor negligently failed to complete the construction work on time, the restaurant operator could recover *640 purely economic losses it suffered as a result. ( J'Aire , at pp. 804-805, 157 Cal.Rptr. 407 , 598 P.2d 60 .) *401 Discerning whether there is a special relationship justifying liability of this sort can nonetheless be a subtle enterprise. In both Biakanja and J'Aire we emphasized that our duty determination rested not just on (i) "the extent to which the transaction was intended to affect the plaintiff," but also on a subset of the Rowland factors relevant to the circumstances before us in those cases: (ii) "the foreseeability of harm to the plaintiff," (iii) "the degree of certainty that the plaintiff suffered injury," (iv) "the closeness of the connection between the defendant's conduct and the injury suffered," (v) "the moral blame attached to the defendant's conduct," and (vi) "the policy of preventing future harm." ( J'Aire , supra , 24 Cal.3d at p. 804 , 157 Cal.Rptr. 407 , 598 P.2d 60 , citing Biakanja , supra , 49 Cal.2d at p. 650 , 320 P.2d 16 .)

Our subsequent decision in Bily , however, underscored for negligence cases involving purely economic losses what is true of all negligence cases. Deciding whether to impose a duty of care turns on a careful consideration of the " 'the sum total' " of the policy considerations at play, not a mere tallying of some finite, one-size-fits-all set of factors. ( Bily , supra , 3 Cal.4th at p. 397 , 11 Cal.Rptr.2d 51 , 834 P.2d 745 , quoting Dillon v. Legg (1968) 68 Cal.2d 728 , 734, 69 Cal.Rptr. 72 , 441 P.2d 912 ( Dillon ).) In Bily , investors in a failed company sued the company's auditor for financial losses they allegedly suffered due to the auditor's negligent preparation of a public report on the company's financial well-being. (See Bily at pp. 376-379, 11 Cal.Rptr.2d 51 , 834 P.2d 745 .) We rejected those claims. (See id. at p. 376, 11 Cal.Rptr.2d 51 , 834 P.2d 745 .) Despite acknowledging that financial losses to investors from negligently prepared audit reports are "certainly" foreseeable, we held that an auditor "owes no general duty of care regarding the conduct of an audit to persons other than the client." ( Bily , at pp. 376, 398, 11 Cal.Rptr.2d 51 , 834 P.2d 745 .)

In requiring more than mere foreseeability for imposing a duty of care in Bily , we appreciated the need to safeguard the efficacy of tort law by setting meaningful limits on liability. ( Bily , supra , 3 Cal.4th at pp. 398-399, 11 Cal.Rptr.2d 51 , 834 P.2d 745 .) Citing decisions from our court limiting recovery for emotional harms based on similar concerns, **888 we explained that although foreseeability " 'may set tolerable limits for most types of physical harm, it provides virtually no limit on liability for nonphysical harm.' " ( Id. at p. 398, 11 Cal.Rptr.2d 51 , 834 P.2d 745 , quoting Thing , supra , 48 Cal.3d at p. 663 , 257 Cal.Rptr. 865 , 771 P.2d 814 .) After all, on " 'clear judicial days' " courts " 'can foresee forever.' " ( Bily , at p. 399, 11 Cal.Rptr.2d 51 , 834 P.2d 745 , quoting Thing , at p. 668, 257 Cal.Rptr. 865 , 771 P.2d 814 .) So although exposure to liability often provides an important incentive for parties to internalize the social costs of their actions, we were *641 concerned that allowing the countless people who rely on public audit reports to recover "pure economic loss suffered" due to a shoddy audit would "raise[ ] the spectre of vast numbers of *402 suits and limitless financial exposure." ( Bily , at p. 400, 11 Cal.Rptr.2d 51 , 834 P.2d 745 .) The resulting universe of potential claims would not only raise difficult line-drawing questions for courts, it might deter socially beneficial behavior. ( Id. at pp. 400, 404, 11 Cal.Rptr.2d 51 , 834 P.2d 745 .) That result was unacceptable. ( Id. at p. 406, 11 Cal.Rptr.2d 51 , 834 P.2d 745 .) We therefore limited auditor liability to claims for negligent misrepresentation brought by plaintiffs who - like the plaintiffs in Biakanja and J'Aire -were "specifically intended beneficiaries" of the defendant's conduct. ( Bily , at pp. 406-407, 11 Cal.Rptr.2d 51 , 834 P.2d 745 .)

To be sure, several additional considerations cut further in favor of strictly circumscribing recovery in Bily . In the audit context, "[t]he client typically prepares [the] financial statements" on which the auditor relies in preparing a report - and that report "is not a simple statement of verifiable fact" but instead "a professional opinion based on numerous and complex factors." ( Bily , supra , 3 Cal.4th at pp. 399-400, 11 Cal.Rptr.2d 51 , 834 P.2d 745 .) The plaintiffs in Bily were also particularly "sophisticated" and had "efficient means of self-protection," such as diversifying their investment portfolios or conducting their own due diligence. ( Id. at p. 406, 11 Cal.Rptr.2d 51 , 834 P.2d 745 .) More fundamentally, purely economic losses flowing from a financial transaction gone awry - which were at issue in Biakanja , J'Aire , Bily , and our other negligence cases to date about purely economic losses - "are primarily the domain of contract and warranty law or the law of fraud, rather than of negligence." ( Aas v. Superior Court (2000) 24 Cal.4th 627 , 636, 101 Cal.Rptr.2d 718 , 12 P.3d 1125 ( Aas ), superseded by statute on other grounds as stated in Rosen v. State Farm General Ins. Co. (2003) 30 Cal.4th 1070 , 1079-1080, 135 Cal.Rptr.2d 361 , 70 P.3d 351 ( Rosen ).)

We nonetheless acknowledged in Bily the "need to limit liability for [purely] economic loss[es]" even in the absence of those additional considerations. ( Bily , supra , 3 Cal.4th at p. 400, fn. 11 , 11 Cal.Rptr.2d 51 , 834 P.2d 745 .) In doing so, we pointed to a hypothetical scenario similar in many ways to the case now before us. We considered a situation where "a defendant negligently causes an automobile accident that blocks a major traffic artery such as a bridge or tunnel." ( Ibid. ; see also Kinsman Transit Co. (2d Cir. 1968) 388 F.2d 821 , 825, fn. 8 [using a similar illustration]; Rabin, Tort Recovery for Economic Loss: A Reassessment (1985) 37 Stan. L.Rev. 1513 , 1536-38 [same].) That defendant would of course be liable for "personal injuries and property damage suffered in such an accident." ( Bily , at pp. 400-401, fn. 11, 11 Cal.Rptr.2d 51 , 834 P.2d 745 .) But would "any court," we continued, "allow recovery by the myriad [other] third parties who might claim [purely] economic losses because the bridge or *403 tunnel" was blocked? ( Id. at p. 401, fn. 11, 11 Cal.Rptr.2d 51 , 834 P.2d 745 .) Based on concerns about limitless liability and unending *642 litigation, as well as on long-standing legal consensus, we considered that prospect "doubtful." ( Ibid .) **889 B.

What we recognized in Bily fits with numerous decisions from other jurisdictions - as well as the Restatement of Torts. That consensus cuts sharply against imposing a duty of care to avoid causing purely economic losses in negligence cases like this one: where purely economic losses flow not from a financial transaction meant to benefit the plaintiff (and which is later botched by the defendant), but instead from an industrial accident caused by the defendant (and which happens to occur near the plaintiff).

Concerned about line-drawing problems and potentially overwhelming liability, courts across the country have rejected recovery for purely economic losses stemming from man-made calamity. Take the New York Court of Appeals' decision in 532 Madison . There, part of a 39-story office tower collapsed, shutting down more than a dozen bustling blocks of midtown Manhattan for several weeks. (See 532 Madison , supra , 727 N.Y.S.2d 49 , 750 N.E.2d at p. 1099 .) Among the plaintiffs in 532 Madison were local businesses who alleged that would-be customers "were unable to gain access to their stores" due to the disaster, forcing the plaintiffs to shut down for an extended period of time. ( Id. , 727 N.Y.S.2d 49 , 750 N.E.2d at pp. 1099-1100.) They sued on behalf of themselves and "all other business entities" operating within the affected city blocks. ( Ibid. )

The plaintiffs in 532 Madison sought compensation for the income they lost from the tower collapse. The New York Court of Appeals responded by declining to hold "that a landowner owes a duty to protect an entire urban neighborhood against purely economic losses." ( 532 Madison , supra , 727 N.Y.S.2d 49 , 750 N.E.2d. at pp. 1102.) It instead "limit[ed] the scope of defendants' duty to those who ha[d], as a result of th[e] [collapse], suffered personal injury or property damage." ( Id. 727 N.Y.S.2d 49 , 750 N.E.2d at p. 1103 .) The court explained that this limitation provided "a principled basis for reasonably apportioning liability" that was necessary to prevent potentially crushing liability to "an indeterminate group in the affected areas" who could prove "financial losses directly traceable to the" collapse. ( Ibid. ) Adopting that rule, the court added, was what "historically courts ha[d] done" with similar negligence claims. ( Ibid. )

Indeed: the Illinois Supreme Court, for example, reached the same result for similar reasons in litigation flowing from a flood caused by human error that inundated downtown Chicago in 1992. (See *404 In re Chicago Flood Litigation (1997) 176 Ill.2d 179 , 223 Ill.Dec. 532 , 680 N.E.2d 265 , 268, 276.) Consider also the West Virginia Supreme Court of Appeals' decision in Aikens v. Debow (2000) 208 W.Va. 486 , 541 S.E.2d 576 , the Iowa Supreme Court's decision in Nebraska Innkeepers, Inc. v. Pittsburgh-Des Moines Corp. (Iowa 1984) 345 N.W.2d 124 , the Massachusetts Supreme Judicial Court's decision in Stop & Shop Companies, Inc. v. Fisher (1983) 387 Mass. 889 , 444 N.E.2d 368 , and the Seventh Circuit's decision applying Wisconsin law in Leadfree Enterprises, Inc. v. U.S. Steel Corp. (7th Cir. 1983) 711 F.2d 805 . Those cases all concerned bridge accidents similar to the hypothetical we discussed in Bily - and those cases all rejected attempts by affected businesses to recover in negligence for purely economic losses resulting from those accidents. (See Aikens , at pp. 579, 589 , ; Nebraska Innkeepers , at pp. 125, 128 ; *643 Stop & Shop , at pp. 371-373 ; Leadfree Enterprises , at pp. 806, 809 ; see also American Petroleum and Transport, Inc. v. City of New York (2d Cir. 2013) 737 F.3d 185 , 187, 196-197 [rejecting under federal maritime law recovery for purely economic losses stemming from a drawbridge malfunction].) Among their concerns were the endless "ripple effects of a negligence claim based upon pure economic loss." ( Aikens , at p. 591 , ; see also Dundee Cement Co. v. Chemical Laboratories, Inc. (7th Cir. 1983) 712 F.2d 1166 , 1172 [opining that allowing recovery in negligence for purely economic losses may unleash "multiversant possibilities" for litigation that "are staggering to the imagination"].)

Similar rationales buttressed the Court of Appeals for the District of Columbia's decision in Aguilar v. RP MRP Washington Harbour, LLC (D.C. 2014) 98 A.3d 979 ( Aguilar ) and the Connecticut Supreme Court's decision in **890 Lawrence v. O & G Industries, Inc. (2015) 319 Conn. 641 , 126 A.3d 569 . Like the California Court of Appeal's decision in Adams , those cases rejected claims for lost wages brought by employees whose workplaces were forced to close by a man-made disaster - a flood in Aguilar and an explosion in Lawrence . (See Aguilar , at pp. 981 , 983 ; Lawrence , at pp. 571, 585 ,.) In fact, more than a half century ago we ourselves approved a decision from an intermediate appellate court in Ohio that arrived at the same conclusion on very similar facts - another explosion causing the closure of a nearby workplace. (See Fifield Manor v. Finston (1960) 54 Cal.2d 632 , 636, 7 Cal.Rptr. 377 , 354 P.2d 1073 , citing Stevenson v. East Ohio Gas Co. (Ohio Ct.App. 1946) 73 N.E.2d 200 , 201-204.)

Federal courts sitting in admiralty have dealt with industrial accidents perhaps most like the one before us: maritime spills of oil and other pollutants. Leaving aside one narrow exception not applicable here, they too have refused to impose a duty of care to guard against purely economic losses. To wit: in State of Louisiana ex rel. Guste v. M/V TESTBANK (5th Cir. 1985) 752 F.2d 1019 ( Testbank ), two ships collided in the Mississippi River *405 Gulf. ( Id. at p. 1020.) Some 12 tons of a toxic chemical called pentachlorophenol rushed into the water and caused the suspension of fishing, shrimping, and other maritime activities across four hundred square miles of marsh and waterways. ( Ibid. ) Among the plaintiffs were businesses like boat rental operators, seafood restaurants, and tackle and bait shops. ( Id. at pp. 1020-1021.) They sued to recover "for economic loss unaccompanied by physical damage" that the spill had inflicted. ( Id. at p. 1021.)

The Fifth Circuit rejected those claims. (See Testbank , supra , 752 F.2d at pp. 1028-1029.) The court echoed concerns about "wave upon wave of successive economic consequences" and stressed that "[t]hose who would delete the requirement of physical damage have no rule or principle to substitute," save perhaps letting the trier of fact determine case-by-case, whim-by-whim which claims for purely economic losses warrant recovery. ( Id. at p. 1028.) The Fifth Circuit further explained that "to the extent that economic analysis" mattered, it favored rejecting recovery for purely economic losses. ( Id. at p. 1029.) That was because defendants in industrial accident cases - despite their frequently deep pockets - will have more difficulty obtaining third-party insurance coverage against purely economic losses than will individual plaintiffs seeking comparable first-party insurance. (See ibid. ) Defendants' potential liability for purely economic losses in such cases is massive and indeterminate. ( Ibid. ) So insurance companies cannot feasibly offer them comprehensive coverage - or even fix a sensible premium based on actuarial measurement.

*644 ( Ibid. ) Plaintiffs' "own potential losses," by contrast, "are finite and readily discernible." ( Ibid. ) They can therefore obtain insurance to cover them - perhaps relatively cheaply. ( Ibid. ; see also Posner, Common-Law Economic Torts: An Economic and Legal Analysis (2006) 48 Ariz. L.Rev. 735 , 737-738.)

Faced with an oil spill diverting a container ship at substantial cost, the First Circuit in Barber Lines A/S v. M/V Donau Maru (1st Cir. 1985) 764 F.2d 50 denied recovery in negligence for those purely economic losses. The First Circuit's analysis in many ways mirrored the Fifth Circuit's reasoning in Testbank. (See Barber Lines , at pp. 50-52.) Through the pen of then-Judge Breyer, the First Circuit explained that the "number of persons suffering foreseeable financial harm in a typical accident" - like a car crash - "is likely to be far greater than those who suffer traditional (recoverable) physical harm." ( Id. at p. 54.) And when it comes to industrial accidents, that proliferation of potential liability for purely economic losses is even more dramatic. (See ibid. ) An oil spill, for instance, "foreseeably harms" not just those whose property is "covered with oil," but also "blockaded ships, marina merchants, suppliers of those firms, the employees of marina businesses and suppliers, the suppliers' suppliers, and so forth." ( Ibid. ) That indeterminate liability, the First Circuit continued, made third-party insurance coverage against purely economic losses less feasible than first-party insurance. ( Ibid. )

*406 It also risked over-deterring socially productive activities. ( Id. at p. 55.) And unable to "distinguish between, say, oil spill accidents and tunnel accidents," the First Circuit rejected the idea of adopting different duty rules depending **891 on the particular "industrial context" at issue. ( Id. at p. 57.)

The narrow exception mentioned earlier, to which we now turn, does not help Plaintiffs. Applying maritime law and California law alike in Union Oil Co. v. Oppen (9th Cir. 1974) 501 F.2d 558 , the Ninth Circuit held that commercial fishermen could recover in negligence for the "diminution of aquatic life" caused by an oil spill. ( Id. at pp. 563, 570.) But that was because theirs was "a pecuniary loss of a particular and special nature" grounded in the time-worn principle that "seamen are the favorites of admiralty." ( Id. at pp. 567, 570 ; see also Curd v. Mosaic Fertilizer, LLC (Fla. 2010) 39 So.3d 1216 , 1228.) Recovery in Union Oil was therefore tightly circumscribed: it was "limited to the class of commercial fishermen" whose livelihoods depend on the flourishing of aquatic life in the commons of the sea and thus did not include, for example, recreational fisherman whose " 'Sunday piscatorial pleasure' " depended on angling in the same waters. ( Union Oil , at p. 570, quoting Oppen v. Aetna Ins. Co. (9th Cir. 1973) 485 F.2d 252 , 260.) Indeed, the Ninth Circuit further cautioned that its narrow holding based on unique features of the maritime context did "not open the door to claims" from others "whose economic or personal affairs were discommoded by the oil spill." ( Union Oil , at p. 570.) Not "every decline in the general commercial activity of every business" nearby, the court reasoned, was "a legally cognizable injury for which the defendants may be responsible." ( Ibid. ) So in Union Oil - as in every case discussed so far - recovery in negligence for purely economic losses was the exception, not the rule.

Against all these decisions, only the New Jersey Supreme Court's opinion in People Express Airlines, Inc. v. Consolidated Rail Corp. (1985) 100 N.J. 246 , 495 A.2d 107 ( People Express ) cuts definitively the other way. In People Express , a railroad fire forced a nearby terminal at Newark International Airport to shut down for *645 twelve hours - a terminal housing the plaintiff's business. (See id. at p. 108.) The plaintiff brought a negligence claim for income lost as a result - a claim the New Jersey Supreme Court permitted to proceed. ( Id. at pp. 108, 116.) The court imposed a tort duty to guard against purely economic losses where there is "an identifiable class with respect to whom [the] defendant knows or has reason to know are likely to suffer such damages from its conduct." ( Id. at p. 116.) The court stressed "that an identifiable class of plaintiffs is not simply a foreseeable class of plaintiffs" - such as happenstance bystanders - but instead a class that is "particularly foreseeable in terms of the type of persons or entities comprising the class, the certainty or predictability of their presence, the approximate numbers of those in the class, as well as the type of economic expectations disrupted." ( Ibid. ) *407 Yet decades after the demise of the airline that gave the case its name, People Express remains "a lonely outpost." (Rabin, Respecting Boundaries and the Economic Loss Rule in Tort (2006) 48 Ariz. L.Rev. 857 , 858.) Its relatively ad hoc standard, embodied in a fact-intensive " 'particular foreseeability' " test, has been avoided by other courts with - as one scholar put it - "a striking degree of unanimity." ( Ibid. ; see also 532 Madison , supra , 727 N.Y.S.2d 49 , 750 N.E.2d at p. 1103 [declining to follow People Express ]; Aguilar , supra , 98 A.3d at p. 984 [same].)

Little wonder the Restatement of Torts takes the dominant view. Although acknowledging that "[d]uties to avoid the unintentional infliction of economic loss" exist in certain recognized circumstances, the latest Restatement provides that there is "no general duty to avoid the unintentional infliction of economic loss on another." (Rest.3d, Torts, Liability for Economic Harm (Tent. Draft. No. 1, Apr. 4, 2012) § 1 (Restatement T.D. 1).)

**892 In justifying that position, the Restatement echoes widespread judicial concern that purely economic losses "proliferate more easily than losses of other kinds" and "are not self-limiting" in the same way. (Restatement T.D. 1, § 1, com. c.) Those characteristics, the Restatement explains, threaten "liabilities that are indeterminate and out of proportion to [a defendant's] culpability," and with them "exaggerated pressure to avoid an activity altogether." (Restatement T.D. 1, § 1, com. c.) For centuries, in fact, similar concerns have justified strict limits on private recovery for a public nuisance. (See 4 Blackstone, Commentaries 167 [noting that a public nuisance is usually not privately actionable because "it would be unreasonable to multiply suits by giving every man a separate right of action"]; accord Rest.3d Torts, Liability for Economic Harm (Tent. Draft. No. 2, Apr. 7, 2014) § 8, com. c. (Restatement T.D. 2); Civ. Code, § 3493 [originally enacted in 1872].)

Only when the foregoing considerations are "weak or absent" - such as in Biakanja and J'Aire , but not in Bily - does a duty to guard against purely economic losses exist under the Restatement approach to negligence claims. (See Restatement T.D. 1, supra , § 1, com. d; see also *646 Restatement T.D. 2, supra , § 7, com. a [using 532 Madison 's facts and the *408 court's holding as an illustration of the Restatement view].) But in this case, as in the mine run of man-made disaster cases, those rationales apply with full force.

The allegations before us underscore the ineluctable difficulty associated with imposing a duty to guard against purely economic losses in negligence cases like this one. It may be possible to quantify the profits any one business lost because of an industrial accident, but imposing such a duty would nevertheless create line-drawing problems across - quite literally - space and time. So although our duty determination must ultimately "occur[ ] at a higher level of generality" than would a jury's analysis of fact-intensive issues like breach and causation ( Kesner , supra , 1 Cal.5th at p. 1144, 210 Cal.Rptr.3d 283 , 384 P.3d 283 ), we examine some particulars of Plaintiffs' claims to illustrate those two sets of persistent line-drawing problems.

We lack clear spatial bounds within which to cabin claims like those asserted here.

This case does not involve a so-called special relationship under our precedents. Plaintiffs concede - as they must - that their only relevant ties to SoCalGas are having the misfortune of operating near the Aliso Facility. Accordingly, they propose to limit the class they seek to represent based on geographic proximity alone. Putative class members here are businesses operating "in the area within five miles" of the leak, a space which Plaintiffs characterize as "the precise area from which residents were evacuated."

What is far from clear is why the five-mile line means anything. Others beyond that boundary were also affected. We discern no compelling basis for us to let a business operating 4.9 miles away recover its lost profits but deny such recovery to another business operating 5.1 miles away. Nor is it clear what we should do about a third business operating 6 miles away whose balance sheet was hit just as hard by the leak and ensuing evacuation - or perhaps a fourth business operating 10 miles away, whose income depends on supplying Porter Ranch businesses or offering services to its residents. Similar questions arise regarding employees of businesses operating within *409 the five-mile mark but who live outside it - or even well outside it. (This is Los Angeles we're talking about.) They might have lost wages during a temporary business slowdown - or even lost their jobs if their employers were forced to cut back permanently. Those employees might not be included in Plaintiffs' proposed class, but their losses are foreseeable, too. They could come to court **893 next in lawsuits of their own. And if we were to permit recovery for purely economic losses in this case, we don't see how we could justify denying it in that one.

Most of the foregoing difficulties emerge even when an evacuation zone is set in stone. But here the lines drawn were traced in sand. Plaintiffs' own complaint acknowledges that, a few weeks after the leak was detected, the evacuation zone was extended beyond the initial five-mile mark. Why businesses operating outside the original *647 boundary but inside the new one should not get to recover their equally real and foreseeable financial losses we do not know.

Using the boundary of an evacuation zone as a liability line might not just lack predictability. In certain circumstances, it could also inject a dangerous incentive into disaster response efforts. Consider how a company taking after Justice Oliver Wendell Holmes's infamous "bad man" - that is, a company that "cares nothing for an ethical rule" and thus cares "only for the material consequences" of its actions - might respond to an evacuation zone rule. (Holmes, The Path of the Law (1897) 10 Harv. L.Rev. 457 , 459, 461 ; see also Exxon Shipping Co. v. Baker (2008) 554 U.S. 471 , 501-502, 128 S.Ct. 2605 , 171 L.Ed.2d 570 [looking to "Justice Holmes's 'bad man' " in a tort case brought under federal maritime law].) If companies face liability in negligence only for traditionally compensable harms, their financial incentive with respect to evacuations points in one direction: caution. To minimize the risk of, and their liability for, harm to people and property, companies under that legal regime may indeed seek (or at least not try to avoid) large evacuation zones. But imposing liability for purely economic losses - bounded only by the size of the evacuation zone - would blunt that otherwise sharp financial incentive for caution. The larger the evacuation zone, the larger a company's potential liability for purely economic losses. So under that rule, a company taking after Justice Holmes's bad man would face a newly vexing cost-benefit analysis: will an evacuation prevent enough physical damage to offset the purely economic losses it is sure to cause? The calculus of a ruthlessly self-interested company would tend to prioritize maximizing its own economic return, not minimizing the risk of harm to people and property. And where it expects an evacuation to harm its bottom line, our proverbial "bad company" might take steps to confine or prevent one.

Such steps might include, most obviously, overt pressure on public officials to roll back or eliminate a proposed evacuation. But that's not the only possibility.

*410 Public officials must often rely on company information to know what scale of risk the community faces. Case in point: during the very disaster at issue here, authorities allegedly demanded from SoCalGas real-time data about the leak - and a timeline for ending it. So public officials might simply be kept in the dark. That's bad enough when, as here, the public health concerns are things like nausea and nosebleeds. But it would be much worse when, on different facts, the stakes are life and death.

Nor is it always simple to decide what counts as an evacuation, or to resolve claims for purely economic losses where the disaster in question never triggered an evacuation. Some evacuations are mandatory, others are voluntary. And sometimes public officials issue public safety warnings without telling people to leave the area. An evacuation zone rule would require a coherent way to decide which sorts of government action count and which ones don't. We do not see one. What is more, the utility of an evacuation zone rule depends on there being at least some sort of evacuation. So adopting an evacuation zone rule would be of no help in cases where nothing remotely approaching an evacuation happens, but the economic effects are nevertheless severe. (Consider, for instance, an oil spill at sea that leaves dry land mostly untouched.) Faced with all this potential for negative consequences and doctrinal confusion, "we would be acting rashly to adopt a rule treating" evacuation zones as talismanic. ( Intel Corp. v. Hamidi (2003) 30 Cal.4th 1342 , 1363, 1 Cal.Rptr.3d 32 , 71 P.3d 296 [declining to extend trespass *648 liability into cyberspace based on similar doctrinal and practical concerns].)

Without adopting a (not so) bright-line evacuation zone rule, the alternative is applying a fact-intensive, case-by-case standard à la People Express . But we have already experimented **894 with an analogous approach regarding recovery for negligent infliction of emotional distress. It did not go well. In Thing , we lamented the "arbitrary results" and the "inconsistent and often conflicting" body of law that approach produced. ( Thing , supra , 48 Cal.3d at p. 662 , 257 Cal.Rptr. 865 , 771 P.2d 814 .) Which is why we retreated from an ad hoc standard and imposed instead a hard-and-fast rule. ( Id. at pp. 667-668, 257 Cal.Rptr. 865 , 771 P.2d 814 .)

We have not forgotten that experience. Today, we are confronted with hundreds of claims brought by hundreds of businesses stemming from one industrial accident - and that's just the artificially limited class Plaintiffs seek to represent, not the full universe of potential claimants whose pocketbooks were adversely (and foreseeably) affected by the leak. We see no workable way to limit geographically who may recover purely economic losses. Without one, the dangers of indeterminate liability, over-deterrence, and endless litigation are at their apex.

*411 2.

Nor do we see a viable way to limit temporally what purely economic losses could be recovered here.

Plaintiffs allege that they "have been and continue to be heavily impacted by the gas leak." (Italics added.) That is possible because Plaintiffs complain not of being forced to shut down during the disaster - no named plaintiff squarely alleges that - but of losing customers due to the exodus of neighborhood residents. And even though the leak is over, they allege that, for as long as the Aliso Facility remains in use, "business will never return to Porter Ranch as usual." (Italics added.) So Plaintiffs are, in effect, seeking pro rata recovery for the past, present, and future economic toll the leak allegedly had, has, and will have on Porter Ranch. These are claims without end.

True: we could conceivably cabin recovery for purely economic losses to those suffered during the disaster alone. Or we could allow recovery only for such losses suffered during a business closure, not merely for systemic hits to economic demand. Yet upon closer inspection, the alluring simplicity of both approaches quickly proves to be a mirage.

The "during the disaster" option would require a way of determining precisely what the words "during" and "disaster" mean in a given case. That will not always be easy. Even assuming the beginning and end of most disasters can be easily fixed by the closing of a wayward valve or its equivalent, distinctions between one disaster (say, a leak of flammable fluid) and another (a fire) can be unstable. Moreover, disasters like the gas leak at issue here happen over an extended period of time, but other industrial accidents (like tower collapses or railroad explosions) happen in an instant. So for the latter sort of disaster, we might have to use the duration of any subsequent evacuation (if there is one) to time-bound the ensuing claims for purely economic losses. But doing that would inject into disaster response efforts the very same dangerous incentives and other problems discussed above.

The "business closure" option, for its part, would likely prove self-defeating. Requiring affected businesses to close as a prerequisite for recovery in negligence would lock them into a dilemma: shut down and lose any income you might have *649 earned - or stay open and lose any tort claim you might have brought. Difficult though the choice could be for some, many businesses might rationally decide they are better off shutting down. Plaintiff Mediterranean Bistro, for example, would presumably be reluctant to keep its 80-seat restaurant open to serve a handful of customers if doing so meant forfeiting a potentially valuable tort claim. Encouraging businesses to close *412 could thus catalyze the very economic stagnation we want to minimize. Better instead to encourage businesses to continue their economic activity where they can.

None of this is to say that denying recovery for those who did not suffer injury to person or property is a perfect solution in negligence cases like this one. Far from it. It is only the least-worst rule out there.

Like other courts, we acknowledge that denying recovery for purely economic losses under circumstances like these has "the vice of creating results in cases at its edge that **895 are ... 'unjust' or 'unfair' " - or even "seemingly perverse." ( Testbank , supra , 752 F.2d at p. 1029 ; see also 532 Madison , supra , 727 N.Y.S.2d 49 , 750 N.E.2d at p. 1103 [acknowledging that this rule is "to an extent arbitrary because ... invariably it cuts off liability to persons who foreseeably might be plaintiffs"].) The courthouse doors are open for people who experience slight physical injury - yet closed to others who suffer devastating purely economic losses. That line may appear arbitrary in some sense. Yet so are the alternatives we have considered and rejected - and those alternatives, as we've explained, have further flaws of their own.

At any rate, "drawing arbitrary lines is unavoidable if we are to limit liability and establish meaningful rules for application by litigants and lower courts." ( Thing , supra , 48 Cal.3d at p. 666 , 257 Cal.Rptr. 865 , 771 P.2d 814 .) And as we have explained, the ripple effects of industrial catastrophe on this scale in an interconnected economy defy judicial creation of more finely tuned rules. Hence the admittedly imperfect legal regime that governs in most jurisdictions - and that we now confirm governs in ours.

The Legislature, however, may be able to improve that regime in ways that would be exceptionally difficult, if not impossible, for us. To name one example: after we rebuffed homeowners' efforts to recover for purely economic losses stemming from construction defects in Aas , the Legislature responded to popular calls for a more forgiving rule in that context. (See Rosen , supra , 30 Cal.4th at p. 1079, 135 Cal.Rptr.2d 361 , 70 P.3d 351 , citing Aas , supra , 24 Cal.4th at p. 646 , 101 Cal.Rptr.2d 718 , 12 P.3d 1125 .) It enacted a detailed statutory mechanism specifically designed for homeowners seeking redress against negligent builders. ( Rosen , at p. 1079, 135 Cal.Rptr.2d 361 , 70 P.3d 351 , citing Civ. Code, § 895 et seq. ) To name another: in view of "the economic and social disruptions arising out of the Lake Davis Northern Pike Eradication Project," the Legislature set up a special process for people to recover for, among other things, purely economic losses suffered due to that environmental protection effort. ( Gov. Code, §§ 998, 998.2.)

*413 With the economic consequences in this case allegedly so severe, and the number of people affected allegedly so large, the Legislature could be spurred yet again to act. To be sure, purely economic losses caused by a natural gas leak may present their own set of challenges. But so too, we can only presume, of those caused by an oil spill. And in that context the Legislature has already interceded. It enacted legislation permitting those "who derive[ ]

*650 at least 25 percent" of their income from activities that utilize "natural resources" to recover - without regard to fault - for "[l]oss of profits or impairment of earning capacity due to the injury, destruction, or loss of ... natural resources" from a spill. ( Gov. Code, § 8670.56.5, subd. (h)(6).) Perhaps there's a basis for further industry-specific legislative or regulatory action. And through the democratic process, the Legislature can bring to bear a mix of expertise while considering competing concerns to craft a solution in tune with public demands.

A partial solution leveraging the insurance market may also prove feasible, at least for some businesses. Although many business interruption insurance policies presently available might not cover the purely economic losses alleged here (see Buxbaum v. Aetna Life & Casualty Co. (2002) 103 Cal.App.4th 434 , 448-449, 126 Cal.Rptr.2d 682 ), private insurance companies could conceivably see a profit-making opportunity in today's decision. Now certain that a lawsuit seeking purely economic losses of this sort will not succeed, businesses operating near a natural gas storage facility - or a dam, shipping lane, oil well, and so forth - may be more inclined to buy insurance covering profits they stand to lose if disaster strikes. (See, e.g., Testbank , supra , 752 F.2d at p. 1029 [observing that a local business's "own potential losses" in the event of an industrial accident "are finite and readily discernible," which may enable them to obtain insurance "at a relatively low cost"].) If so, private insurance companies might expand their policy offerings accordingly.

Finally, we recognize Plaintiffs' concern that SoCalGas's alleged negligent behavior will go insufficiently deterred if we deny **896 recovery here. But SoCalGas is not getting off scot-free. At oral argument, the company represented that some 50,000 claimants have alleged in other litigation that they suffered property damage caused by the leak - several hundred of whom are local businesses. It further informed us, and we have no reason to doubt, that the company has spent some $450 million on remedial measures and agreed to pay another $120 million as part of a settlement with local authorities. SoCalGas, operating in a heavily regulated domain, also remains under investigation - and may face further consequences in the future. *414 III.

Risks from industrial accidents raise grave concerns for society, and we have no doubt the accident precipitating this case caused significant hardships. To compensate those harmed and to deter those who do the harming, our society assigns tort law a pivotal role. But that does not mean society's interests are best served by extending its scope indefinitely. Meaningful limits on tort liability, along with the incentives they set, are crucial to the functioning of our economy and of our courts. Where such limits leave gaps in our social fabric, tort does not stand alone: insurance also compensates, regulation also deters. And where gaps persist, the Legislature can act.

The better part of a century has passed since then-Judge Cardozo warned that permitting recovery in negligence for purely economic losses can threaten indeterminacy-cubed: "liability in an indeterminate amount for an indeterminate time to an indeterminate class." ( *651 Ultramares Corp. v. Touche (1931) 255 N.Y. 170 , 174 N.E. 441 , 444.) Courts across the country have since heeded that warning, by and large denying recovery in negligence cases like this one even though purely economic losses inflict real pain. That prevailing rule of no recovery is, like society itself, imperfect. Yet nearly everyone follows a rule that few (if any) entirely like. California does, too. So we affirm the Court of Appeal's judgment.

We Concur:

CANTIL-SAKAUYE, C. J.

CHIN, J.

CORRIGAN, J.

LIU, J.

KRUGER, J.

GROBAN, J.

We refer to the named plaintiffs in this action collectively as "Plaintiffs." The Court of Appeal held in Adams that employees could not sue a railroad for lost wages even though, allegedly, the railroad's negligence caused an explosion that destroyed the employees' workplace, a nearby factory. (See Adams , supra , 50 Cal.App.3d at pp. 39-41, 123 Cal.Rptr. 216 .) The "ordinary standards of demurrer review still apply" even though this case "arrived at the Court of Appeal by the unusual path of a writ petition challenging an order overruling a demurrer." ( City of Stockton v. Superior Court (2007) 42 Cal.4th 730 , 746-747, 68 Cal.Rptr.3d 295 , 171 P.3d 20 .) Having concluded in J'Aire that recovery for foreseeable purely economic losses "should not be foreclosed simply because it is the only injury that occurs," we disapproved the Court of Appeal's decision in Adams "[t]o the extent that [it] h[eld] that there can be no recovery for negligent interference with prospective economic advantage." ( J'Aire , supra , 24 Cal.3d at pp. 806-807 & fn. 3, 157 Cal.Rptr. 407 , 598 P.2d 60 .) So as the Court of Appeal recognized, J'Aire disapproved Adams only to the extent it purported to impose an absolute rule that a plaintiff can never recover for negligently inflicted purely economic losses. We of course determine foreseeability not by reference to specific parties but instead based on the general sort of conduct at issue. (See Kesner , supra , 1 Cal.5th at p. 1145.) (See, e.g., Goonewardene v. ADP, LLC (2019) 6 Cal.5th 817 , 820, 243 Cal.Rptr.3d 299 , 434 P.3d 124 [purely economic losses stemming from payroll company's alleged miscalculation of wages]; Centinela , supra , 1 Cal.5th at p. 1013, 209 Cal.Rptr.3d 280 , 382 P.3d 1116 [purely economic losses stemming from health care plan's delegation of financial responsibility to pay emergency service claims to third parties]; Quelimane , supra , 19 Cal.4th at pp. 57-60, 77 Cal.Rptr.2d 709 , 960 P.2d 513 [purely economic losses stemming from defendant's refusal to issue title insurance policies on real property].) Although the Alaska Supreme Court discussed People Express in a positive light in Mattingly v. Sheldon Jackson College (Alaska 1987) 743 P.2d 356 , it did so while allowing a contractor to recover purely economic losses suffered from the collapse of a trench that was dug specifically "so that his employees could work in it." ( Id. at pp. 359-361.) Mattingly thus resembles our special relationship precedents far more so than People Express - or this case. We express no view on whether, or to what extent, these line-drawing problems persist (or dissipate) in cyberspace. (See, e.g., Dittman v. UPMC (Pa. 2018) 196 A.3d 1036 , 1038 [considering whether to impose a tort duty to guard against "purely pecuniary" losses stemming from a data breach].) The United States Congress has passed similar legislation. (See 33 U.S.C. § 2702 (b)(2)(E).)