Lucker Manufacturing, a Unit of Amclyde Engineered Products, Inc. v. The Home Insurance Company

OPINION OF THE COURT BECKER, Circuit Judge. This is an appeal from summary judgment granted by the district court in favor of the defendant, The Home Insurance Company (“The Home”) and against the plaintiff, Lucker Manufacturing, a Unit of Amclyde Engineered Products, Inc. (“Lucker”), in an insurance coverage dispute arising under the diversity jurisdiction, 28 U.S.C. § 1332. The appeal raises a question of interpretation of language that appears in the industry-wide, standard-form liability insurance policy known as Comprehensive General Liability Insurance (“CGL”): does the clause “loss of use of tangible property that has not been physically injured” cover costs of preventing a defective component from becoming incorporated into a product that has been designed but has not yet been manufactured? The product at issue here is an anchoring system made by Lucker for the off-shore oil drilling industry and called a Lateral Mooring System (“LMS”). Because of a defect that Lueker discovered in a component of the LMS — castings manufactured by Milwaukee Steel Foundry, a division of Grede Foundries, Inc. (“Grede”) — Lueker was forced to increase the number of safety precautions in the manufacturing process for the LMS. The increased precautions ensured that the castings incorporated into the LMS would not be defective. When Lueker sued Grede for these costs, The Home, Grede’s insurer, asserted that these costs were not covered by the policy, and refused to defend or indemnify Grede. As part of a settlement agreement between Lueker and Grede, Grede assigned to Lueker any rights that it had against The Home for its failure to defend or indemnify. Lueker then sued The Home. The district court granted summary judgment for The Home because it believed that the additional safety precautions Lueker had to add to its manufacturing process did not represent a “loss of use” to Lueker of the LMS or LMS design, but rather represented a change in its customers’ acceptance of the original LMS and LMS design. Since this injury to Lueker did not constitute loss of use, the court held that The Home had not breached its duty to defend or indemnify Grede. In our view, however, loss of use can and should cover the added costs of preventing a defective component from being incorporated into a product, even if those added costs were incurred because of a change in customer preferences. As we discuss below, the distinction that the court drew between “loss of customer acceptance” and loss of use is arbitrary. Liability for costs incurred because of a change in demand for a product in the marketplace brought about by the insured’s wrongful act seems to be precisely the type of liability that loss of use coverage was designed to protect against. But the fact that the Lueker complaint adequately alleged a loss of use (and gets by summary judgment on that issue) does not end the inquiry because, under the CGL policy, the loss of use must have been to “tangible property.” When The Home withdrew coverage, it knew that the LMS itself had not physically existed at the time Gre-de’s casting failed but was only in the design stage. Lueker contends, however, that its LMS design, which did exist, was tangible property within the meaning of the policy. We disagree because under current Wisconsin and Pennsylvania law, a system design like that of the LMS is not tangible property as that term is used in the standard-form CGL policy. Since an insurer has no duty to defend or indemnify claims that fall outside the coverage of the policy, The Home had no duty to defend or indemnify Grede. Consequently, we will affirm the judgment of the district court. I. FACTS AND PROCEDURAL HISTORY In 1989, Lueker contracted with Shell Oil Company to design and manufacture the *811 LMS. An LMS is, in essence, a huge permanent anchor. It is fixed to the ocean floor and holds in place ships, oil platforms, and other large structures floating on the surface. Among its components are “castings,” large metal objects that attach to the ocean floor and hold the cables connected to the ship or platform. Lucker purchased a number of these castings from a foundry in Milwaukee owned by Grede. Before putting these east-ings into the LMS, Lucker decided to test their strength, and it arranged an “equipment load test” at Lehigh University. Confident that the test would impress its customer, Lucker invited a Shell representative to watch. The test, however, was a disaster. To everyone’s horror, a casting involved in the test suffered a catastrophic failure. Had it been incorporated into the LMS and put into operation, Shell’s ships and platforms would have floated off to sea. So Shell told Lucker that, although it still wanted the LMS, Lucker had to maintain tighter control over the production and testing of the steel for the eastings. Lucker complied at a cost of $600,000. At the time of the failure of the castings, the LMS was only in the design phase and had not yet been built. After it completed the LMS, Lucker sued Grede on both tort and contract theories for the cost of compliance with Shell’s instructions. It is undisputed that the castings were defective and that the defect was Grede’s fault. Grede had a CGL policy with The Home which insured it against any “property damage” Grede would be legally obligated to pay Lucker. Although Lucker never claimed that the LMS was physically injured, physical injury was not a prerequisite of coverage. According to the terms of the policy, “property damage” included “[l]oss of use of tangible property that is not physically injured.” Such a provision typically covers an interruption of income that is caused by a wrongful act not accompanied by physical injury. The Home conditionally defended Grede during the early part of the lawsuit under a reservation of rights. But when the district court held that the tort claims sought purely economic losses which are not recoverable under Pennsylvania law and that Lucker could pursue its contract but not its tort claims, the Home withdrew its defense of Grede and disclaimed all liability under the policy. In a letter to Grede, The Home claimed that contract-based economic damages were not loss of use of tangible property and that there was no basis for coverage. In addition, The Home took the position that the losses were excluded by the “business risk,” “failure to perform,” and “sistership” exclusions in the policy. 1 The dispute between Lucker and Grede eventually went to trial, and Lucker won a jury award of approximately $500,000. 2 A few months later, Lucker and Grede settled the lawsuit: Grede paid Lucker $600,000 and assigned to Lucker its rights against The Home to recover for defense costs and indemnification under the policy. Standing in Grede’s shoes, Lucker sued The Home claiming that The Home was in breach of both its duty to defend and to indemnify Grede, and that its breach was in bad faith. Both The Home and Lucker moved for summary judgment, agreeing that there were no factual disputes. The district court granted summary judgment for The Home. Lucker Mfg., Unit of Amclyde Engineered Prods., Inc. v. Home Ins. Co., 818 F.Supp. 821 (ED.Pa.1993). The court held that The Home did not breach its duty to defend Grede because Lucker’s complaint against Grede did not allege damages for “loss of use” of the LMS or LMS design, or for any other form of property damage covered by the policy. Id. at 828. It also held that The Home had no duty to indemnify Lucker because none of Lucker’s damages fell within the policy’s cov *812 erage. Id. Additionally, it held that The Home did not act in bad faith. Id. at 830. In deciding as it did, the court looked both to the language of the CGL policy and the language of the complaint Lucker had filed in its lawsuit against Grede. The CGL policy in this case provided that The Home: will pay those sums that the insured [Gre-de] becomes legally obligated to pay because of “bodily injury” or “property damage” to which the insurance applies. The policy defined “property damage” as a. Physical injury to tangible property including all resulting loss of use of that property; or b. Loss of use of tangible property that is not physically injured. Lucker conceded that there had been no claim in the underlying action for actual physical injury to property other than the eastings, which (it also conceded) were not covered by the policy. Instead, it characterized its loss as one for damages resulting from the loss of use of its own tangible property — the LMS design. 3 In fending off the Home’s claim that its pleadings were inadequate, Lucker contended that paragraphs 14-17 of its complaint potentially stated a claim for loss of use of the design of the LMS. Paragraph 14 alleged that Shell had imposed stricter standards for the production and testing of the steel for the LMS; paragraph 15 alleged that Lucker had lost a competitive advantage in bidding on Shell projects and had sustained damage to its reputation; paragraph 16 alleged that had the castings not failed, the additional tests would not have been necessary; and paragraph 17 alleged that Lucker had suffered increased testing costs, increased costs in performing the Shell contract, lost profits, and lost future profits. According to the district court, however, paragraphs 14-17 merely averred that “due to the failure of the castings, [Lucker] could no longer sell the original LMS or LMS design to Shell because Shell imposed additional requirements; and it was the cost of complying with these additional requirements that Lucker sought to recover.” 818 F.Supp. at 825 (footnote omitted). Nowhere did Lucker claim that the LMS as designed could no longer perform its intended function because of the defect discovered in the castings. Rather, Lucker acknowledged that the LMS as originally conceived could still hold floating objects in place, but argued that its “use” of the LMS design and the LMS was to sell the LMS to Shell and other customers, and that when it was forced to change the manufacturing process and the design, it lost the use of the LMS design. Id. at 827. The court rejected this argument: There is no indication in the underlying complaint, either by inference or otherwise, that Lucker was ever unable, due to the failure of the castings or otherwise, to offer the original LMS for sale to Shell or to offer use of the original LMS design to other customers. Lucker simply complained that when offered, Shell and perhaps other customers wanted something different because of the failure of the castings. Thus Lucker did not allege a loss of an intended use of the original LMS as merchandise for sale or the original LMS design as a means of producing such merchandise; Lucker alleged a loss of expected customer acceptance of its product and design. And loss of customer acceptance of a product or design is in no way the equivalent of “loss of use” of a product or design under the policy. To equate the two would be to link CGL coverage to the vagaries of customer desire and make insurers of liabilities into guarantors of markets for goods and services. The CGL policy at issue here committed The Home to no such *813 undertaking. The Home contracted to defend and indemnify Grede for damages resulting from the loss of use of tangible property only. Id. at 828 (footnote omitted). On this basis, the district court granted summary judgment for The Home on the duty to defend issue. Once it determined that The Home was not in breach of its duty to defend, its conclusions on the duty to indemnify and bad faith claims followed almost as a matter of course, and it granted The Home summary judgment on both of those issues. Id. at 830. This appeal followed. Because the material facts are not disputed, we have plenary review of the district court’s decision. Pacific Indemnity Co. v. Linn, 766 F.2d 764, 760 (3d Cir.1985). II. CHOICE OF LAW Faced with the possibility that either Pennsylvania or Wisconsin law applies to this diversity ease, a choice of law question looms on the horizon. Before a choice of law question arises, however, there must actually be a conflict between the potentially applicable bodies of law. See Oil Shipping B.V. v. Sonmez Denizcilik Ve Ticaret A.S., 10 F.3d 1015, 1018 (3d Cir.1993). Where there is no difference between the laws of the forum state and those of the foreign jurisdiction, there is a “false conflict” and the court need not decide the choice of law issue. In re Complaint of Bankers Trust Co., 752 F.2d 874, 882 (3d Cir.1984) (“If the foreign law to which the forum’s choice-of-law rule refers does not differ from that of the forum on the issue, the issue presents a ‘false conflict.’ ”); Lambert v. Kysar, 983 F.2d 1110, 1114-15 (1st Cir.1993) (“We need not resolve the [conflict of law] issue ... as the outcome is the same under the substantive law of either jurisdiction.”). Neither party has pressed a choice of law question on this appeal because neither has been able to identify any differences between Wisconsin and Pennsylvania law on the questions of an insurer’s duty to defend and indemnify. Our own research has not identified any relevant differences either. As far as we can tell, the outcome of this lawsuit should be the same under either Wisconsin or Pennsylvania law. Since there is no conflict of law under such circumstances, we will avoid the choice of law question. Cf. Melville v. American Home Assur. Co., 584 F.2d 1306, 1311 (3d Cir.1978) (warning courts to avoid dicta on conflicts questions when not put in issue). We therefore will interchangeably refer to the laws of Wisconsin and Pennsylvania in discussing the law governing The Home’s duty to defend and indemnify. 4 III. THE DUTY TO DEFEND A. General Principles and The Pleading Question Under the governing law, an insurance company is obligated to defend an insured whenever the allegations in a complaint filed against the insured potentially fall within the policy’s coverage. This duty to defend remains with the insurer until facts sufficient to confine the claims to liability not within the scope of the policy become known to the insurer. See Imperial Casualty & Indem. Co. v. High Concrete Structures, Inc., 858 F.2d 128, 131-32 (3d Cir.1988); Sola Basic Indus., Inc. v. United States Fidelity & Guar. Co., 90 Wis.2d 641, 280 N.W.2d 211, 213 (1979) (“[I]t is necessary to determine whether the complaint alleges facts which, if proven, would give rise to liability covered under the terms and conditions of the policy.”); Stidham v. Millvale Sportsmen’s Club, 421 Pa.Super. 548, 618 A.2d 945, 953-54 (1992) (if indemnification depends upon the existence or nonexistence of disputed facts, the insurer has a duty to defend until the claim is narrowed to one patently outside the policy coverage). Before considering whether a complaint is potentially covered by a policy, it is necessary to determine the coverage of the policy in the first instance. In both Pennsylvania and Wisconsin - (as in the majority of jurisdictions), the inquiry into coverage is *814 independent of, and antecedent to, the question of duty to defend. 5 See Erie Ins. Exch. v. Transamerica Ins. Co., 516 Pa. 574, 533 A.2d 1363, 1368 (1987) (first construing the terms of the policy, and then determining whether the complaint alleged facts which, if proven, would come within the scope of the policy as construed); U.S. Fire Ins. Co. v. Good Humor Corp., 173 Wis.2d 804, 496 N.W.2d 730 (1993) (same). Traditional principles of insurance policy interpretation control the inquiry into coverage. The policy language must be tested by what a reasonable person in the position of the insured would have understood the words to mean. See Imperial, 858 F.2d at 131. We must construe ambiguous language to provide coverage. Id. A provision is ambiguous if reasonable persons considering the relevant language in the context of the entire policy could honestly differ as to its meaning. . Id.; see also Harford Mutual Ins. Co. v. Moorhead, 396 Pa.Super. 234, 578 A.2d 492, 503 (1990) (the policy must unequivocally indicate coverage or non-coverage), appeal denied, 527 Pa. 617, 590 A.2d 757 (1991). Nevertheless, a court should be careful not to create an ambiguity and, likewise, it should avoid rewriting the policy language in such a way that it conflicts with the plain meaning of the language. Imperial, 858 F.2d at 131. Once coverage of the policy is determined, the court then looks to the underlying complaint to see if it triggers coverage. The underlying complaint need not track the policy language for there to be coverage: under the liberal rules of notice pleading, Lucker’s complaint needed only to indicate the type of litigation involved so that the defendant would have fair notice of the claim and its defenses. See First State Underwriters Agency of New England Reinsurance Corp. v. Travelers Ins. Co., 803 F.2d 1308, 1315 (3d Cir.1986); Ollerman v. O’Rourke Co., 94 Wis.2d 17, 288 N.W.2d 95, 98 (1980); see also Western Casualty & Sur. Co. v. Budrus, 112 Wis.2d 348, 332 N.W.2d 837, 839-40 (1983) (liberally construing a complaint to include a claim for loss of . use of tangible property). Thus, the fact that Lucker did not include the magic words “loss of use” or “tangible property” in its complaint does not relieve The Home of its duty to defend. As the district court recognized, the complaint, reduced to its relevant essentials, averred that the failure of the Grede castings prevented Lucker from being able to sell the LMS design to Shell. Lucker, 818 F.Supp. at 825. The main question in this appeal, then, is not whether there was adequate notice of such a claim, but rather whether, given traditional principles of insurance policy interpretation, such a claim is properly considered a claim for loss of use of tangible property. With these general principles in mind, we discuss first the question whether Lucker’s damages represented a loss of use of the LMS design. We then turn to the question of whether the LMS design was tangible property. B. Loss of Use One of the two principal issues in this appeal — whether the damages Lucker suffered because Shell wanted Lucker to beef up its quality control of the steel in the LMS are potentially “loss of use” damages as that term is used in the CGL — turns on a conflict between the connotations of the term “use” on the one hand and the objectives of insurance on the other. As has been mentioned, the district court agreed with The Home that a “loss of use” within the meaning of the CGL form contract does not occur when a customer demands from the injured party added safeguards in the manufacturing process after a defect in the insured’s product is discovered. *815 Critical to the district court’s decision that Lucker’s damages were not from a “loss of use” was the fact that Lucker never claimed that the LMS design could not work after the castings had failed at the Lehigh test. In the court’s view, the failure of the castings merely had the undesirable consequence of “reducing the acceptability of the LMS and LMS design to Shell.” Lucker, 818 F.Supp. at 826 n. 12. Apparently, the district court saw a distinction between the loss of the ability to physically use the LMS design and the loss of the ability to sell the LMS design. One was use and the other was non-use. Cf. Eljer Mfg. v. Liberty Mut. Ins. Co., 972 F.2d 805, 810 (7th Cir.1992) (discussing physical injury in the context of CGL standard policies), ce rt. denied, — U.S. -, 113 S.Ct. 1646, 123 L.Ed.2d 267 (1993). In everyday English, the district court’s distinction makes sense. The term “use” conjures the idea of some kind of physical application of property, as when a carpenter uses a hammer. If someone does not want to buy the hammer but it can still pound nails into wood, the hammer can still be “used.” That a customer does not want to buy it has no effect on its usefulness. But such a distinction has little to do with the objectives of parties to insurance contracts. See Eljer Mfg., 972 F.2d at 809. As we see it, the focus of the insurance coverage determination should be on whether a customer’s unwillingness to use a hammer, regardless of whether it is physically possible to use the hammer, is a type of lost use for which a risk-averse business pays its premium. “Ordinary language” interpretations of phrases are not the only plausible interpretations of insurance contracts, especially when the contract is between sophisticated business entities. It is important to ask what function “loss of use” was intended to perform in a CGL policy before relying on a common sense or lay distinction between physical use and other uses. See Erie Ins. Exch., 533 A.2d at 1367 (stating that the term “use” must be considered with regard to the setting in which it is employed). The 1966 version of the CGL defined “property damage” as “injury to or destruction of tangible property.” Yet cases interpreting that policy often afforded coverage for non-physical injuries, like diminution in value. On the other hand, a number of cases excluded injuries where there was no physical contact between the injurer and the property injured. See Sola Basic, 280 N.W.2d at 214-15 (citing cases). In 1973, the CGL policy was revised to clear up these internal tensions. The new CGL policy replaced the old definition of property damage with the two-part definition contained in the policy involved in this case. The first part repeats the old definition, except that the word “physical” is put before “injury.” The second part is the “loss of use” definition, which covers injury which is not physical because the insured’s property does not physically contact that of the injured party. The revised language, then, allows coverage for both physical and non-physical injuries. 6 Of course, Lucker did not suffer physical injury to the LMS (since it did not exist) or the LMS design. Nor did it lose the physical use of the LMS or LMS design. Indeed, Lucker could still have manufactured it and offered it for sale, and, even according to Lucker, the original LMS design would still have worked properly. Nevertheless, Luck-er did lose the economic use of the original LMS design: because of the defective castings, Shell was no longer willing to buy the product, and Lucker could no longer use the LMS design as a source of income. The question in this case, therefore, reduces to whether the lost “use” has to have been a lost physical use of the property, or whether it can also include a lost non-physical or economic use of the property. The district court thought that loss of use should cover only lost physical use, and that customer acceptance simply was not a “use.” Luck- *816 er, 818 F.Supp. at 828. We believe, however, that both the purposes behind liability insurance and the case law interpreting liability insurance suggest that the loss of a nonphysical use of a product, such as offering it for sale, should be considered a “loss of use”; and that the decreased value of a product because of loss of customer acceptance of the product is a “loss of use” within the meaning of the standard CGL policy. We test our understanding against the baseline case for both parties, Sola Basic. In Sola Basic, 280 N.W.2d at 211, the insured was sued when a transformer it had manufactured for use in an electrical steel furnace owned by Thunder Bay Manufacturing was damaged by the insured’s employee who had been sent to repair the transformer. Id. at 212-13. Because of the damage to the transformer, Thunder Bay could not operate the furnace. The Supreme Court of Wisconsin held that Thunder Bay’s loss of use of the furnace was a covered loss. Id. at 217. Sola Basic suggests that the use of a product as merchandise for sale or as a means of production is a “use” within the meaning of a CGL form policy. In real terms, such a loss of use affects the injured party’s ability to supply a product. In Sola Basic, demand for Thunder Bay’s product remained the same, but its ability to supply its product — that is, to supply the product at a competitive price — was impaired because the furnace could not work. It is as if the supply curve was suddenly shifted to the left due to the negligence of the insured, and the liability insurance loss of use concept made up the difference. 7 The district court accepted the Sola Basic decision as authority but distinguished it from this case because the change in customer acceptance did not affect the ability of Lucker to supply the product. A change in customer acceptance is a change in the demand curve, not the supply curve. Apparently, the court believed that the change in demand should be analyzed differently from a change in supply. But it is not clear to us why this should be the case. When a manufacturer supplies a defective product that injures a third party, the injury from that particular defect may, depending on the reaction of the injured third party, affect either the demand for the third party’s goods or the supply of those goods. Differing factors may influence whether the effect is on supply or demand. For example the time of discovery of the defect is important. If a manufacturer of enamel finishes produces a defective batch and sells it to a manufacturer of widgets, the discovery of the defect before the enamel is applied to the widgets will affect the supply of widgets the manufacturer has available to sell because production must be halted while new enamel finish is obtained. If the discovery is made after the enamel is applied, the demand for widgets will be affected because customers won’t want to buy widgets with blotchy finishes. Another factor that may influence whether supply or demand is affected is the need of the widget producer to move merchandise. He may decide he would rather sell blotchy widgets at a lower price in order to keep up cash flow. The type of injury, whether caused by the widget producer’s loss through reduction in supply or in demand, may thus be beyond the control of the enamel finish manufacturer. However, in purchasing CGL coverage, the manufacturer will want protection from liability for the widget producer’s loss in either event. For the same reason, if Grede were supplying steel to the widget producer, Grede would want CGL coverage whether the defective steel caused a reduction in supply because the production of widgets was halted or a reduction in demand because customers doubted the durability of widgets. Consequently, a manufacturer worried about liability has no reason to see a difference in the type of coverage he has bought unless the insurance contract says otherwise, and the standard CGL insurance contract does not say otherwise. CGL insurance appears to protect the insured from making up the difference in the injured party’s revenue, regardless of whether the liability is for shift *817 ing supply or demand. Therefore, coverage for both sorts of injuries appears to be within the reasonable expectations of the insured. The relevant ease law supports our view that no difference exists between coverage for wrongful acts that affect supply, and those that affect demand. One case from this Court interpreting Pennsylvania law apparently held that a change in demand for a product may be considered a “loss of use.” See Imperial Casualty, 858 F.2d at 128. In Imperial Casualty, Keystone, a manufacturer of washers, had a contract with Nice Bearing Company to supply it with a special kind of washer. Id. at 130-31. Keystone bought special steel from High Concrete to make the washers. Id. Nice discovered that Keystone’s washers were defective and rejected them. Id. Apparently the steel High Concrete had supplied was defective. Id. Keystone then sued High Concrete for breach of warranty, seeking “loss of use” damages in the form of lost profits and other incidental charges, including added freight charges. Id. at 135. This court held that these damages were potentially recoverable under High Concrete’s CGL policy as “loss of use.” Id. at 135-36. In Imperial, we allowed recovery for loss of use despite the fact that Keystone could have produced all of the washers made from the defective High Concrete steel that it wanted. Nothing stopped Keystone from supplying washers made from the High Concrete steel except for the fact that its customer refused to accept them. As in this case, a change in customer acceptance of the product caused the loss of use to Keystone, which had no use for the product other than selling it. To the extent there was a loss of use, it had to have been that Keystone’s lost ability to sell the product was due to decreased demand. The Home’s attempt to distinguish Imperial on the ground that in Imperial there was physical injury to the washers does not change the fact that this Court found a loss of use that was attributable solely to a change in customer acceptance of the product. The cases upon which The Home relies do not specifically address the contours of the loss of use provision. The Home relies on McDowell-Wellman Eng’g Co. v. Hartford Acci. & Indem. Co., 711 F.2d 521 (3d Cir.1983), and Trio’s, Inc. v. Jones Sign Co., 151 Wis.2d 380, 444 N.W.2d 443 (1989), for the proposition that loss of use does not include the type of loss Lucker suffered. In each case, the insured’s product failed (McDowell-Wellman involved an ore bridge 8 and Trio’s involved a restaurant sign) and the insured incurred liability for a number of costs associated with the product’s failure. In McDowell-Wellman the costs included the expenses incurred to keep the steel plant operational without a functional ore bridge, and in Trio’s the costs included lost revenues for the time the restaurant went without a sign. Although the courts denied recovery for those costs in both cases, neither court did so because there was no loss of use. Rather, in both cases the courts thought that the costs recovered represented loss of use of the insured’s product, costs which the policies, specifically excluded from coverage. McDowell-Wellman, 711 F.2d at 526-27; Trio’s, 444 N.W.2d at 444-45. 9 Because both cases were decided on the basis of policy exclusions, and not on the basis of loss of use, they are inapposite. Apparently at the heart of the district court’s opinion in this case was a fear that allowing coverage for Lucker would link insurance coverage to the “vagaries of customer desire” and turn liability insurers into *818 “guarantors of markets for goods and services.” Lucker, 818 F.Supp. at 828. That fear in the context of this case is exaggerated: the loss of use provision is triggered only when the change in preference is due to the insured’s wrongful act, and obviously not all changes in customer preferences are due to an insured’s wrongful act. To be sure, identifying the cause or magnitude of damages due to changes in customer preference might prove difficult in some cases, but substantive principles of tort and contract law account for such difficult inquiries with doctrines that shield defendants (the insureds) from liability where appropriate. For example, the economic loss doctrine, by denying recovery under the rubrics of duty and proximate cause, protects defendants from unanticipated plaintiffs and disproportionate liability. 10 Even though reasonable minds may disagree with our construction of the language, the language is, at the very least, reasonably susceptible to more than one construction from the viewpoint of the insured and is therefore ambiguous. Since under Pennsylvania and Wisconsin law we must resolve any ambiguity in favor of coverage, we hold that loss of use includes the loss of customer acceptance that Lucker suffered in this case. C. Tangible Property We need not reverse, however, if the LMS was not tangible property within the meaning of the insurance contract. Although the underlying complaint may have potentially alleged that the LMS existed as a tangible entity, at the time The Home disclaimed coverage for Lucker’s complaint against Gre-de it was apparent that the LMS itself was not tangible at the time the castings failed. The LMS design did exist, however, and the complaint potentially pled that Lucker had lost the use of its design after the castings had failed. 11 Thus, we must determine whether a design is tangible property as that term is used in a CGL. 12 We conclude that the language of the policy, analogous case law from Pennsylvania and Wisconsin, and case law from other jurisdictions compel the conclusion that the term tangible property does not include non-tangible property like system designs. Under Pennsylvania and Wisconsin law, tangible property is property that can be felt or touched, or property capable of being possessed or realized. In re Estate of MacFarlane, 313 Pa.Super. 397, 459 A2d 1289, 1291-92 (1983); see also United States Fidelity & Guar. Co. v. Barron Indus., Inc., 809 F.Supp. 355, 360 (M.D.Pa.1992) (the term tangible in a CGL covers things which are physical — capable of being touched and objectively perceivable); Holsum Foods, Div. *819 of Harvest States Coop. v. Home Ins. Co., 162 Wis.2d 563, 469 N.W.2d 918, 921 (1991) (similar approach). In contrast, intangible property is defined as property that does not have intrinsic value but which is merely representative or evidence of value, like stock certificates. MacFarlane, 459 A.2d at 1292; Barron Indus., 809 F.Supp. at 360 (under Pennsylvania law CGL policy does not cover intangible property, such as property that represents value but has no intrinsic marketable value of its own (e.g., stock, investments, copyrights, promissory notes); property regarded as intangible rights (e.g., goodwill and reputation); or economic interests (e.g., overhead, profits, investment value, and productivity)); Columbia Gas Transmission Carp. v. Commonwealth, 19 Pa.Cmwlth. 523, 339 A.2d 912, 918 (1975) (natural gas is tangible property for tax purposes even though it cannot be felt because it is capable of being perceived as materially existent; “[ijntangible properties in the law are such incorporeal rights as shares of capital stock, choses in action, copyrights and the like”), error dismissed, 22 Pa.Cmwlth. 467, 350 A.2d 193 (1975), rev’d on other grounds, 468 Pa. 145, 360 A.2d 592 (1976); Palmolive Co. v. Conway, 43 F.2d 226, 227 (D.Wis.1930) (trademarks, trade secrets, and goodwill not tangible property), aff'd, 56 F.2d 83 (7th Cir.1932), cert. denied, 287 U.S. 601, 53 S.Ct. 8, 77 L.Ed. 524 (1932); American Tel. & Telegraph Co. v. Department of Revenue, 143 Wis.2d 533, 422 N.W.2d 629, 631 (1988) (cash, shares of stock, notes, bonds, etc., not tangible as defined in tax statute). But see Man, Levy & Nogi, Inc. v. School Dist. of Scranton, 31 Pa. Cmwlth. 75, 375 A.2d 832, 834 (1977) (insurance premiums are tangible property for tax purposes). 13 Because these principles are so well settled, Lucker does not argue that the concept of the LMS is tangible property, for such an idea cannot be touched and is not materially existent. Rather, Lucker contends that a design which is reduced to a tangible medium, like a blueprint or a computer disk, should be considered tangible property. The Home, on the other hand, argues that where the real value of a design is in the idea, not in the physical plans that memorialize it, any loss in value of the design represents a loss in the value of the idea, which is not a loss of use of tangible property. We believe that The Home’s position is the one best supported by the relevant case law. The Home principally relies on a taxation case from Wisconsin which held that the sale *820 of computer keypunch cards was not a sale of tangible property for purposes of the Wisconsin sales tax. See Janesville Data Center, Inc. v. Wisconsin Dep’t of Revenue, 84 Wis.2d 341, 267 N.W.2d 656, 658-59 (1978). In Janesville, the Wisconsin Supreme Court reasoned that the information on the card, rather than the card itself, was the object of the transaction, and that the tangible medium keeping the information was merely incidental to the transaction. Id. Therefore, the court held, the sale of the keypunch cards was not a sale of tangible property. Id. 14 Lucker has cited no authority from the relevant jurisdictions. Instead it has countered with a Minnesota case, Retail Sys., Inc. v. CNA Ins. Cos., 469 N.W.2d 735 (Minn.Ct. App.1991), which held that a computer tape that stored information was tangible property covered under a liability policy, and a case from the United States District Court for the Northern District of Georgia, State Farm Fire & Casualty Ins. Co. v. White, 777 F.Supp. 952 (N.D.Ga.1991), which held that architectural plans in blueprint form were tangible property covered under a CGL policy. Neither Retail Systems nor White extended the concept of tangible property as far as Lucker would have us do here, however. In Retail Systems the court limited the coverage to the considerable value of the computer tape as a storage medium, disallowing recovery for the value of the data it stored. Similarly, in White, a case in which developers sought coverage for costs they incurred in converting architectural drawings, the district court recognized that the only recovery due the developers under the policy was for the value of the paper and ink, and not the value of the ideas the paper and ink embodied. 777 F.Supp. at 954-55. Both cases drew a sharp distinction between recovery for the value of a tangible medium storing ideas, and recovery for the ideas themselves. 15 To the extent that the damage had been merely to the value of the idea, it was not damage to “tangible” property. In this ease, none of the losses Lucker sought from Grede represented a loss in value of the storage medium in which the design for the LMS was embodied or in the costs in reducing the design to blueprints or computer tape (e.g. the costs of having engineers draw up the plans for the system). The recovery Lucker sought was for the loss of use of the design itself — for the loss in usefulness of the original concept of the LMS. The loss of use of this concept, however, was not loss of use of something which could be touched or felt. For this reason, we hold that Lucker’s loss of use of the LMS design was not loss of use of tangible property. We note, however, that the “tangibility” limitation in the standard form CGL seems to be in tension with what we believe is its underlying rationale. As far as we can tell, the CGL limits coverage to “tangible property” to avoid indemnifying the insured for any liability the insured faces for damage caused to stocks, bonds, copyrights and the like, items for which either the insurer is arguably in no better position to spread risk than the insured, or which would dramatically increase the premiums. 16 But by making “tangibility” the touchstone of coverage, the CGL excludes a significant class of property *821 for which liability insurance reasonably could be provided — property like system designs or computer software. The “tangibility” limitation was probably a reasonable way to separate insurable from non-insurable property interests in 1973 when the CGL standard policy was drafted. But the tremendous increase in automation, and the concomitant increase in demand for intangible products like computer software and system designs during the past twenty years, has made such a limitation of questionable value. As a matter of risk spreading, we see no qualitative difference between the need for insurance to protect a manufacturer from liability incurred because its product shuts down a furnace, damages a computerized billing system, or, as in this case, devalues a system design. 17 Nevertheless, we are bound by the language of the policy, and we cannot stretch it to include non-tangible property like the LMS design. Unlike “loss of use,” which can plausibly be construed to include loss of customer acceptance, it would require too great a departure from the meaning of “tangible” to hold that a system design is tangible property covered under the policy. Therefore, because the LMS design was not tangible property, there was no “property damage” and thus no coverage under the policy for Lueker’s loss. As a result, we agree with the district court that The Home did not breach its duty to defend Grede when it disclaimed coverage. 18 IV. THE DUTY TO INDEMNIFY In light of our holding on the duty to defend, we may dispose of the duty to indemnify summarily. An insurer has a duty to indemnify its insured only if it is established that the insured’s damages are actually within the policy coverage. Safeguard Scientics v. Liberty Mut. Ins. Co., 766 F.Supp. 324, 334 (E.D.Pa.1991), aff'd in part, rev’d in part without opinion, 961 F.2d 209 (3d Cir.1992). Lucker recovered from Grede as a result of the jury verdict: 1) $32,934 for the actual cost of the castings; 2) $200,007 in “Test Project Costs”; and 3) $251,337 for “Costs Absorbed to reproduce Shell’s casting to a higher specification.” Lucker cannot recover any of these costs because neither the LMS nor the LMS design was tangible property, and hence there was no property damage covered by the policy. 19 The judgment of the district court will be affirmed. Footnotes [1] The business risk exclusion excludes from coverage damage to the castings; the failure to perform exclusion excludes from coverage nonphysical injury to other property arising from a breach of contract unless the damage comes from a sudden and accidental injury; the sister-ship exclusion excludes from coverage the costs of inspection, repair, or replacement of the castings. [2] Whether the jury correctly found Grede liable for the costs Lucker incurred due to Shell’s demands is not before us. [3] Lucker actually argued to the district court that it had lost the use of two things: 1) the LMS and 2) the design of the LMS. Because the district court decided there was no loss of use, it assumed that the LMS and LMS design both existed. It appears from the record before us, however, that the LMS itself never existed, but was still in the design stage at the time the castings failed. While the complaint potentially pled that the LMS itself existed, the record indicates that The Home knew these facts when it disclaimed coverage. We believe that since the LMS was no more than a design at the time, Lucker’s claims that it lost the use of the LMS and the LMS design actually claim the same thing — loss of use of the design. [4] The only real choice of law issue in this case involves the applicability of Pennsylvania's insurer bad faith statute, 42 Pa.Cons.Stat. § 8371. We need not reach this issue given our resolution *814 of the duty to defend and duty to indemnify questions. [5] A minority of courts have held that where the question of coverage is an open question the insurer has a duty to defend. See St. Paul Fire & Marine Ins. Co. v. National Computer Sys., Inc., 490 N.W.2d 626, 632 (Minn.Ct.App.1992) (where the question whether binders with confidential information in them were covered was an open one, the insurer had a duty to defend); see also Centennial Ins. Co. v. Applied Health Care Sys., Inc., 710 F.2d 1288, 1291 (7th Cir.1983) (applying California law, the court determined that the question whether information stored in a data processing system could be tangible property was irrelevant for purposes of determining the duty to defend because it was an unresolved question). [6] The classic example of a loss of use injury is a case in which a manufacturer of construction cranes sells a defective crane which collapses in front of a restaurant, thereby impairing the restaurant's income. If the restaurant sues the manufacturer and recovers the lost income, the manufacturer would be covered by the "loss of use” component of the CGL policy. See Eljer, 972 F.2d at 810; George H. Tinker, "Comprehensive General Liability Insurance — Perspective and Overview," 25 Federation of Ins. Counsel Q., 217, 232 (1975). [7] Another way to think of Sola Basic is by viewing the furnace as an income stream. When the furnace was shut down, Thunder Bay was deprived of the income stream from the furnace for the period it was inoperative. That loss was lost use and was within the scope of the coverage. [8] An ore bridge is part of a system in a steel plant that carries raw materials to a blast furnace for processing. McDowell-Wellman, 711 F.2d at 523. [9] Both McDowell-Wellman and Trio’s are essentially cases about allocating damages between the loss of use of the malfunctioning component and loss of use of the other property. Both employ a sort of "but for" causation approach to deny coverage: "but for” the malfunctioning ore bridge, the steel company would not have incurred the additional costs; and "but for” the malfunctioning sign, the restaurant would have lost no revenues. Such an approach essentially reads the term "loss of use” right out of the policy. Whenever there is a malfunctioning component, all loss of use of other property is caused at some level by the malfunctioning product or else there would simply be no injury due to the insured's product to which liability may attach. [10] Concerns about vastly increased liability are not at all implicated in this case. Quite to the contrary, when Lucker replaced the castings and added safeguards so that faulty castings would not be installed in the LMS, it was mitigating the consequences of Grede having supplied it with a defective product. It was thus curtailing rather than expanding potential liability. [11] The LMS designs and plans existed prior to the catastrophic failure of Grede’s product. They were used previously in AmClyde's Placid Oil Project, the Conoco Project, and the Akashi Bridge Project in Japan. The existence of the design was also apparent from the complaint. [12] The Home has argued vigorously that since the underlying complaint sought only loss of profits and other economic losses, Lucker was not seeking damages for loss of use of tangible property. Economic harms, The Home argues, are simply not tangible. While such an observation may be correct, it misunderstands the nature of the losses Lucker was seeking from Grede. Lucker was not seeking compensation for economic losses qua economic losses; rather, Luck-er was pointing to its economic losses as a proxy for the value of the lost use of its LMS design. Even the cases on which The Home relies recognize that an intangible economic loss, such as the diminution of value of a fixed asset, is recoverable if it provides a measure of damage to the tangible property. See Liberty Bank of Montana v. Travelers Indem. Co., 870 F.2d 1504, 1509 (9th Cir.1989); Giddings v. Industrial Indem. Co., 112 Cal.App.3d 213, 219, 169 Cal.Rptr. 278, 281 (1980) (a complaint seeking to recover for economic losses "falls within the scope of the insurance coverage only where these intangible economic losses provide a measure of damages to physical property which is within the policy's coverage") (internal quotation omitted). In this case, Lucker lost the use of its LMS design. The absorbed costs oí beefing up its specifications for the steel in the LMS and the other costs it recovered from Grede are a proxy for the value of the lost use of the original LMS design. As such, they are recoverable under the policy if the LMS design was tangible property. [13] The distinction between tangible and intangible property made by these cases tracks the definitions found in Black’s Law Dictionary. Black’s defines tangible property as "property that has physical form and substance and is not intangible" and intangible property as "such property as has no intrinsic and marketable value, but is merely the representative or evidence of value, such as certificates of stock, bonds, promissory notes, copyrights, and franchises.” Black’s Law Dictionary (6th ed. 1990). Insurance companies reasonably might want to exclude coverage for damage to such intangible interests because estimating the potential liability for purposes of setting the premium might be very difficult, or even if the premium could be calculated, insuring against such liability might expose the company to such increased costs because of a great variance in liability that a CGL policy might become prohibitively expensive. It may also be that insurance companies are in no better position to insure against such losses than the insured. For example, assuming that the stock is publicly traded, one can insure against changes in market price by purchasing options. We note, however, that it is difficult to explain why liability for copyright or patent infringement would be included among the interests not covered by a CGL policy. There is no obviously increased moral hazard problem (an insufficient incentive to be careful) with respect to copyright or patent infringement as compared to other types of injuries. Nor does it appear to raise the possibility of huge liability, or liability that is difficult to calculate. And it does not appear that the marketplace provides an efficient alternative to an insurance policy as it does with things like stocks. Perhaps exclusion of coverage for copyright or patent violations can be explained by the fact that the CGL policy is a standard form and most customers of such policies are not as risk averse with respect to copyright and patent violations as they are with other types of tort damages and so they do not demand coverage for such injuries. At all events, it appears sensible to presume that purchasers of liability insurance, who are principally concerned with more conventional forms of tort damage that their product may cause a third party, reasonably would be willing to bear the risk of loss to traditionally intangible interests in exchange for lower premiums. [14] The other case on which The Home places principle reliance, Gulf Insurance Co. v. L.A. Effects Group, Inc., 827 F.2d 574 (9th Cir.1987), does not really advance The Home's position. In that case, L.A. Effects was sued by Twentieth Century Fox for failing to perform adequately in designing the special effects for the film Aliens, Although L.A. Effects' argument that Fox’s damages amounted to a loss of use of Aliens was rejected, Fox did not allege as damage any diminution in value to the film. Id. at 577-578. Thus the issue presented here was not before that court and consequently that court did not hold that the loss of value of the film could not be loss of use of tangible property. [15] Other courts have also seen such a distinction. See, e.g., Commonwealth of Massachusetts v. Rizzuto, 1980 WL 4637 (Mass.Super.1980) (Commonwealth could not prosecute for theft a defendant that copied someone else's idea for a film because the idea, although reproduced in tangible form and capable of being reproduced into tangible form, was not itself tangible; distinction must be drawn between cause of value and thing of value). [16] See note 13 above. [17] A preferable way to approach the problem might be for the insurer to eliminate the over-broad "tangibiliiy” requirement from the definition of property damage and instead specifically exclude traditional intangible property interests, like stocks, copyrights and goodwill. [18] Because we hold that there was no “property damage,” we need not construe the policy's exclusions. [19] Because we find that The Home was in breach of neither its duty to defend nor its duty to indemnify, it did not act in bad faith and did not violate 42 Pa.Cons.Stat. § 8371.