[*969] Opinion

Plaintiff Martin J. Morris appeals from a summary judgment entered in favor of The Paul Revere Life Insurance Company. Morris alleged that Revere acted in bad faith in denying his claim for benefits under his disability policy. The court granted summary judgment in favor of Revere, concluding that its interpretation of the policy language and the California statute governing incontestability clauses in disability insurance policies was reasonable in light of the evolving state of the law. We agree and affirm the judgment.

Revere issued disability policies to Morris in 1989 and 1990. Among the policies’ coverage limitations was one for disabilities caused by a “sickness or disease which first manifests itself after the Date of Issue and while Your Policy is in force.”

The policies also included an incontestability clause, as required by California Insurance Code section 10350.2, which provided: “a. After Your Policy has been in force for 2 years, excluding any time You are Disabled, We cannot contest the statements in the application, [f] b. No claim for loss incurred or Disability that starts after 2 years from the Date of Issue will be reduced or denied because a sickness or physical condition not excluded by name or specific description before the date of loss had existed before the Date of Issue.”

The policies also included a provision ensuring conformity with state laws: “Any provision of this Policy which, on its Date of Issue, conflicts with the laws of the state in which You reside on that date is amended to meet the minimum requirements of such laws.”

After the policies had been in effect for over two years, Morris filed a claim for benefits, because of a disability resulting from multiple sclerosis (MS). Although Revere initially paid benefits to Morris for about six months under the policies, it subsequently notified him that it was denying further benefits on the ground that his MS was not a covered disease because it had manifested itself before the effective dates of the policies.

Morris sued to obtain the benefits he claimed were owed to him. He also alleged that in light of the incontestability clause, Revere had acted in bad faith in denying the claim. The trial court granted summary judgment in favor of Revere on the coverage issue, concluding that the undisputed evidence demonstrated Morris’s MS had manifested itself prior to the date the policies had issued and was thus not a covered sickness. Having agreed[*970] with Revere on the coverage issue, the trial court had no occasion to address the allegation of bad faith.

Morris appealed (Aprahamian v. Paul Revere Life Ins. Co., D036827), and while that appeal was pending, our Supreme Court considered and decided the identical coverage issue in Galanty v. Paul Revere Life Ins. Co. (2000) 23 Cal.4th 368 [97 Cal.Rptr.2d 67, 1 P.3d 658]. In Galanty, the court noted courts around the country had split on the issue. Decisions interpreting substantially identical policy and statutory provisions—as to whether a coverage provision excluding sicknesses which had manifested prior to the date of policy issuance was inconsistent with an incontestability clause that precluded an insurer from denying coverage on the ground the disease had preexisted—had gone both ways.

According to the Supreme Court, seven jurisdictions, led by the Fifth Circuit Court of Appeals in Massachusetts Casualty Insurance Co. v. Forman (5th Cir. 1975) 516 F.2d. 425, had found in favor of the insurer and concluded there was no conflict. (Galanty v. Paul Revere Life Ins. Co., supra, 23 Cal.4th at pp. 378, 384.) Those courts reasoned that the “manifestation” of a disease referred to the point when the insured has reason to know of or discover the condition, whereas the “existence” of the disease referred to the point when it first took hold, without regard to whether the insured had any reason to suspect, or even could have discovered, its existence. As exemplified in Forman, those courts held that distinguishing between those points gave the parties the benefit of their bargain and was consistent with the purposes of an incontestability clause: “The statutory [incontestability] clause protects both parties from uncertainty where coverage turns on when a disease first existed, occurred or arose. Under such polices coverage may be uncertain even long after the policy was procured because a disease may exist but not be discoverable by reasonable medical investigation. This clause resolves the uncertainty in favor of coverage beginning two years after the policy date. Where coverage turns upon manifestation rather than existence, the insurance policy itself eliminates this uncertainty because manifestation is normally discoverable by reasonable medical investigation.” (Massachusetts Casualty Insurance Co. v. Forman, supra, 516 F.2d at pp. 429-430.)

On the other hand, the Supreme Court noted nine other jurisdictions had ruled in favor of the insured, concluding the incontestability clause precluded the insurer from arguing a disease had premanifested once the incontestability period had commenced. (Galanty v. Paul Revere Life Ins. Co., supra, 23 Cal.4th at p. 378.)

Finally, the Supreme Court stated there was only one published opinion on point in this state, McMackin v. Great American Reserve Ins. Co. (1971) 22[*971] Cal.App.3d 428, 439-440 [99 Cal.Rptr. 227]. The McMackin court had decided the issue in the insured’s favor, albeit “without useful discussion.” On the other hand, two other California appellate courts had decided the issue in the insurer’s favor, including the lower court in Galanty, and the court in Callahan v. Mutual Life Ins. Co * (Cal.App.), in which review was granted and the case held pending the Galanty opinion. (Galanty v. Paul Revere Life Ins. Co., supra, 23 Cal.4th at p. 377.)

As part of a comprehensive review of the history and development of incontestability laws, the Supreme Court explained that the concept was first developed in the context of life insurance. As the incontestability law developed in that area, it was interpreted by some courts as wiping out the ability of insurers to place limits on the risks they undertook. In at least one case, the insurer had excluded death caused by air travel from coverage under the policy. When the insured later died in a plane crash, the court held that the exemption had been nullified when the policy became incontestable, and ordered the insurer to pay the benefits. (Jordon v. Western States Life Ins. Co. (1952) 78 N.D. 902 [53 N.W.2d 860].)

Other courts, recognizing the incongruity of such results, were careful to differentiate the concepts of coverage and incontestability. As explained by Chief Judge Cardozo in Metropolitan Life Ins. Co. v. Conway (1930) 252 N.Y. 449 [169 N.E. 642], “The provision that a policy shall be incontestable after it has been in force during the lifetime of the insured for a period of two years is not a mandate as to coverage, a definition of the hazards to be borne by the insurer. It means only this, that within the limits of the coverage, the policy shall stand, unaffected by any defense that it was invalid in its inception, or thereafter became invalid by reason of a condition broken.” (Id. at p. 452.) That opinion, described as “very influential” by our Supreme Court (Galanty v. Paul Revere Life Ins. Co., supra, 23 Cal.4th at p. 380), inspired insurance law reforms designed to ensure that concepts of coverage and incontestability remained separate in the interpretation of life insurance policies.

It was in that context that insurers began to offer—and state laws began to mandate—incontestability clauses in disability policies as well. Although the language of such clauses was substantially different from that found in life insurance policies (Galanty v. Paul Revere Life Ins. Co., supra, 23 Cal.4th at pp. 381, 383), some courts nonetheless interpreted them in accordance with the principles developed in the life insurance cases. Our Supreme Court unanimously concluded such an analysis was incorrect.

[*972] The court reasoned, “[The] argument has serious flaws, beginning with the assumption on which it depends. In the present context, the assumption that an incontestability clause cannot affect coverage is erroneous. The assumption, as noted, finds its origin in Chief Judge Cardozo’s explanation that an incontestability clause ‘is not a mandate as to coverage, a definition of the hazards to be borne by the insurer. . . .’ (Metropolitan Life Ins. Co. v. Conway, supra, 169 N.E. at p. 642.) This explanation makes perfect sense in its original context: a conflict between an incontestability clause in a life insurance policy, providing simply that the policy ‘shall be incontestable’ after two years, and a provision excluding coverage for a specific hazard, namely, travel by aircraft. (Ibid.) But the proposition that no incontestability clause, regardless of wording, can ever affect coverage has no logical force in the present context of a statutorily required (§ 10350.2) clause that expressly bars the insurer, after two years, from denying coverage ‘because a sickness or physical condition not excluded by name or specific description before the date of loss had existed before the Date of Issue.’ ” (Galanty v. Paul Revere Life Ins. Co., supra, 23 Cal.4th at pp. 384-385.)

The court then rejected the suggestion that a preexisting condition was different from a premanifested condition. “Paul Revere’s argument . . . depends upon an unexpected and inobvious, if not unnatural, definition of the term ‘existed.’. . . [f]. . .In saying that something exists, one does not normally entertain unarticulated mental reservations about manifestation.” (Galanty v. Paul Revere Life Ins. Co., supra, 23 Cal.4th at p. 386.) The Supreme Court thus concluded that a policy provision excluding from coverage any disability caused by a disease that had manifested prior to the date of policy issuance was inconsistent with the statutorily mandated incontestability clause, and must be disregarded. It consequently held that the insured was entitled to coverage as a matter of law, notwithstanding any evidence of premanifestation, and remanded the case to the Court of Appeal for determination of whether the insurer had acted in bad faith. The Supreme Court itself expressed no opinion on the bad faith issue. (Galanty v. Paul Revere Life Ins. Co., supra, 23 Cal.4th at p. 389.)

In light of the opinion in Galanty, the parties stipulated to a reversal of the original judgment and to a remand of this case to the superior court. And, as in Galanty, the sole remaining claim to be adjudicated was whether Revere had acted in bad faith by even arguing that the defense of prior manifestation could be asserted in light of the language of the incontestability clause.

Revere moved for summary judgment on the issue, asserting that in light of the substantial split of authority nationwide, and the relative dearth of case law in California prior to Galanty, the issue was not sufficiently settled[*973] to render its argument unreasonable. The trial court agreed and entered summary judgment in favor of Revere on the bad faith claim.

I

In an appeal from a summary judgment, our standard of review is de novo. (Certain Underwriters at Lloyd’s of London v. Superior Court (2001) 24 Cal.4th 945, 972 [103 Cal.Rptr.2d 672, 16 P.3d 94].) Moreover, even if this appeal had not arisen from a summary judgment, the issue of Revere’s liability depends solely on whether it acted in bad faith when it interpreted Morris’s policy to preclude coverage for disability caused by MS, despite the provisions of California’s incontestability law. (Ins. Code, § 10350.2.) That issue, which turns on the reasonableness of the legal argument advanced by Revere, presents a pure question of law that we would, in any case, review de novo. ^[1]

“[T]he ultimate test of [bad faith] liability in the first party cases is whether the refusal to. pay policy benefits was unreasonable.” (Austero v. National Cas. Co. (1978) 84 Cal.App.3d 1, 32 [148 Cal.Rptr. 653], italics omitted, overruled on other grounds in Egan v. Mutual of Omaha Ins. Co. (1979) 24 Cal.3d 809, 824, fn. 7 [169 Cal.Rptr. 691, 620 P.2d 141].) “[B]efore an [insurer] can be found to have acted tortiously, i.e., in bad faith, in refusing to bestow policy benefits, it must have done so ‘without proper cause.’” (California Shoppers, Inc. v. Royal Globe Ins. Co. (1985) 175 Cal.App.3d 1, 54-55 [221 Cal.Rptr. 171].)

Morris suggests there are unresolved factual issues pertaining to Revere’s subjective understanding of the law and its intent in shaping the law to suit its own ends, which preclude summary judgment in this case. However, if the conduct of Revere in defending this case was objectively reasonable, its subjective intent is irrelevant. An insurer is not a fiduciary, and owes no obligation to consider the interests of its insured above its own. (Love v. Fire Ins. Exchange (1990) 221 Cal.App.3d 1136, 1148 [271 Cal.Rptr. 246].) “An insurer . . . may give its own interests consideration equal to that it gives the interests of its insured (Egan v. Mutual of Omaha Ins. Co., supra, 24 Cal.3d at pp. 818-819); it is not required to disregard the interests of its shareholders and other policyholders when evaluating claims (Austero v. National Cas. Co., supra, 84 Cal.App.3d at p. 30); and it is not required to pay noncovered claims, even though payment would be in the best interests[*974] of its insured (Native Sun Investment Group v. Ticor Title Ins. Co. (1987) 189 Cal.App.3d 1265, 1278 [235 Cal.Rptr. 34].)” (Id. at pp. 1148-1149.)

In short, Revere was entitled to argue for whatever interpretation of the law and policy language most benefited its own interests. The sole issue then, is whether the position it took was objectively reasonable in light of the law that existed prior to Galanty. With that in mind, we turn to the merits.

II

Morris makes several arguments in support of his contention that Revere’s legal argument was unreasonable even before the Supreme Court ruled on the issue in Galanty. First, he argues that the incontestability law had long been settled in California. As Morris points out, Insurance Code section 10350.2 was enacted in 1951, and has never been significantly changed. In his view, then, the law has been settled since that time. We cannot agree. The enactment of a statute establishes its language, but not always its meaning. If there is any uncertainty, the meaning of a statute is ultimately determined by the courts. (Bialo v. Western Mutual Ins. Co. (2002) 95 Cal.App.4th 68, 76-77 [115 Cal.Rptr.2d 3].) In such a case, it cannot be said that the law is settled until the courts have ruled.

Thus, Morris alternatively suggests the law became settled in California in 1971, when McMackin v. Great American Reserve Ins. Co., supra, 22 Cal.App.3d 428, was published. However, as noted by the Supreme Court in Galanty, the McMackin court offered its conclusion “without useful discussion.” Indeed, someone reading McMackin. might well have considered it dicta.

In McMackin, the issue to be decided was whether substantial evidence supported the trial court’s determination that plaintiffs disability stemmed from a particular work-related injury. The court’s primary holding was that there was sufficient evidence to support that conclusion: “The testimony of plaintiffs physician alone would support the trial court’s finding that the total disability was due to the January 4 incident.” (McMackin v. Great American Reserve Ins. Co., supra, 22 Cal.App.3d at p. 439.) Only after upholding the factual determination that plaintiffs entire disability had been caused by the covered accident, did the court turn to the issue of incontestability. It noted that any prior injuries suffered by the plaintiff were irrelevant in any case, because the insurer failed to assert them within two years of issuing the policy. “Not having challenged plaintiffs application within the two-year period, defendant may be said to have taken the plaintiff as it found him and cannot now urge that plaintiffs disability resulted from a[*975] pre-existing disease, illness or injury not covered by the policy.” (McMackin v. Great American Reserve Ins. Co., supra, 22 Cal.App.3d, at p. 440.) That latter holding, being unnecessary to the decision, could well be regarded as dictum.

Moreover, according to the opinion in Galanty, the McMackin dictum had been disagreed with by at least two other Courts of Appeal, including the lower court in Galanty itself. The split of opinion among the Courts of Appeal, even if all were not published, suggests the issue was far from settled in our courts.

Finally, although perhaps not determinative, the very fact that our Supreme Court accepted Galanty for review suggests it did not consider the issue argued by Revere to have been settled. At the time Galanty was decided, California Rules of Court, rule 29(a) provided: “Review by the Supreme Court of a decision of a Court of Appeal will be ordered (1) where it appears necessary to secure uniformity of decision or the settlement of important questions of law; (2) where the Court of Appeal was without jurisdiction of the cause; or (3) where, because of disqualification or other reason, the decision of the Court of Appeal lacks the concurrence of the required majority of qualified judges.” Because there is no assertion that the Court of Appeal in Galanty was either without jurisdiction or improperly constituted, it is clear that Supreme Court review in that case was based upon the apparent necessity “to secure uniformity of decisions or the settlement of important questions of law.”

In any event, if we accepted Morris’s argument that a sole published Court of Appeal decision which addresses an insurance issue, in no matter how cursory a fashion, constitutes unassailable “controlling law” on that point, we would be setting a very troubling precedent. We cannot ignore the fact that intermediate Courts of Appeal sometimes disagree with each other on points of law, including the interpretation of statutes. Such disagreement is a specific basis for Supreme Court review. And sometimes the first court to tackle an issue does not devise the optimal analysis. In some cases, a later court, having had the benefit of time to study the earlier case, and perhaps the reactions to it in the legal community, offers a better-reasoned result.

Under Morris’s scenario, however, it is unlikely that any court would even be given the opportunity to weigh in after one court decides a legal issue in favor of an insured. In that circumstance, no insurer could ever offer a rationale supporting a different result to another court without exposing itself to bad faith damages in the event it does not prevail. There would be, in essence, a strict liability penalty for incorrect argument—a penalty to[*976] which few insurers could be expected to subject themselves. And if insurers were so thoroughly discouraged from ever challenging an appellate court decision—even a bad one—there would be little opportunity for the courts to refine the law. We cannot support such a result.

Morris next asserts that Revere’s reliance upon the precedents favoring its position is somehow disingenuous, because it was a party to many of the cases, and thus had a direct hand in creating the authority it now relies upon to establish the reasonableness of its position. But of course, litigants do not create precedent, courts do. We cannot adopt Morris’s implied premise that Revere somehow hoodwinked the courts and is now attempting to hide behind the barricade of bad law it created. As explained above, Revere is entitled to be an advocate for its own interests. The fact that it had such success doing so in prior cases cannot be twisted into a justification for requiring it to stop.

And the fact that so many courts had concluded that a premanifestation defense was not precluded by substantially identical incontestability clauses, strongly suggests the contention was not an unreasonable one to make prior to Galanty. This is not a situation where an insurer is attempting to spin its one-shot success in convincing a single trial court of its position into proof that the position was reasonable. (See Filippo Industries, Inc. v. Sun Ins. Co. (1999) 74 Cal.App.4th 1429, 1441 [88 Cal.Rptr.2d 881] [rejecting such an attempt, because it would have the “practical effect of denying the insured its right to appeal the trial court ruling because, even if the trial court were reversed, the initial finding would preclude bad faith as a matter of law”].) In this case, Revere had the highest courts of several jurisdictions, as well as two California Courts of Appeal, in its comer. We cannot just dismiss such a congerie of cerebrates as “unreasonable.”

Moreover, we must also reject the suggestion in Filippo Industries that any court opinion issued after the insurance company made its initial decision to deny coverage could not be considered in determining whether the decision was reasonable. If, as in this case, the coverage issue turns upon analysis of a legal point—and assuming the governing law has not changed in the interim—the fact that a court had interpreted that law in the same manner as did the insurer, whether before or after, is certainly probative of the reasonableness, if not necessarily the ultimate correctness, of its position. ^[2]

To be sure, the logic of the Galanty opinion is compelling, and may cause some other jurisdictions to reconsider their positions, but neither those courts[*977] nor Revere had the benefit of the opinion’s clarity prior to its issuance. It would be unfair to conclude, implicitly, that Revere should have seen it coming.

Finally, Morris argues that Revere’s bad faith is demonstrated by the fact that it discontinued his disability payments prior to obtaining any ruling on its prior manifestation defense, rather than continuing payments under a reservation of rights while pursuing a declaratory relief claim. He contends it was the preferred practice in the industry for insurers to simply reserve their rights while litigating coverage issues, and that Revere’s own personnel had acknowledged that approach should be taken if the prior manifestation defense was to be pursued in California. However, Morris offers no authority for his assertion that the failure to continue payments under a reservation of rights can constitute bad faith even if the coverage denial was reasonable, and we find it unpersuasive.

In our view, an insurer’s decision to continue payments under a reservation of rights, rather than cut off benefits entirely where it concludes there is no coverage, goes to the issue of damages, not liability. As long as the insurer’s coverage decision was reasonable, it will have no liability for breach of the covenant of good faith and fair dealing. An insurer which denies benefits reasonably, but incorrectly, will be liable only for damages flowing from the breach of contract, i.e., the policy benefits. (State of California v. Pacific Indemnity Co. (1998) 63 Cal.App.4th 1535, 1551 [75 Cal.Rptr.2d 69].) On the other hand, if the insurer is found to have acted unreasonably, its liability for breach of the covenant of good faith and fair dealing will sound in tort, exposing it to a much broader array of damages. (PPG Industries, Inc. v. Transamerica Ins. Co. (1999) 20 Cal.4th 310, 314-315 [84 Cal.Rptr.2d 455, 975 P.2d 652].) Because of that risk, many insurers reasonably choose to pay the disputed benefits pending a court decision, as a strategy to mitigate whatever damages might later be claimed by the insured, including possible punitive damages, should the insurer ultimately lose the coverage dispute. Such a strategy, while perhaps beneficial to an insured in the short run, is primarily a self-protective measure, not an obligation.

The judgment is affirmed. Respondent is to recover its costs on appeal.

Aronson, J., and Fybel, J., concurred.