When the Cincinnati Insurance Company sold a directors and officers liability policy to Citizens Bank covering breaches of fiduciary duty in the rural bank's tiny trust department, it probably seemed like a safe bet that the bank would never make a claim approaching the policy's $5 million limit. The odds changed dramatically, however, when Citizens' trust officer, without fully understanding the risks involved, began investing the vast majority of his customers' life savings in extremely volatile derivative instruments. Given this combination, before you could say "collateralized mortgage obligation" Cincinnati found itself facing a claim for $5 million. The insurer tried to fend it off, but after a 4-day bench trial in the district court, Judge Harry Leinenweber awarded the bank not only $4.9 million under the policy, but also attorneys fees as a sanction for Cincinnati's efforts to avoid payment.

Cincinnati now appeals the decision, arguing that the bank's conduct fell under exclusions covering actions taken in bad faith and activities related to the funding of trusts. It also claims that the judge abused his discretion in awarding attorneys fees because its conduct was neither "vexatious" nor "unreasonable." Citizens, in turn, cross-appeals, asking that we amend its award to include an extra $100,000 under the policy. We begin our review with the facts.

Citizens First National Bank is an agricultural credit bank located in the town of Princeton, Illinois (population 7,300). In addition to its credit operations, the bank runs a trust department for the farmers and other local residents who make up its customer base. Like any trustee, the bank owes the beneficiaries of the trusts it manages a fiduciary's strict duties of care and loyalty.

To protect itself against claims arising from a breach of these duties, in late 1993 Citizens purchased a year's worth of directors and officers liability insurance from the Cincinnati Insurance Company. Under the policy, Cincinnati agreed to cover losses arising from errors and omissions committed in the administration of the bank's trust accounts. More specifically, in the policy's Trust Department Errors and Omissions Endorsement, the insurer pledged to indemnify Citizens for any "Loss" that the bank incurred as a result of a "Wrongful Act" committed by Citizens as a fiduciary. "Wrongful Act" was defined to include, inter alia, breach of duty, neglect, and error. "Loss" described "any amount which the Insured is legally obligated to pay . . . for a claim or claims made against the Insured." Cincinnati agreed to cover 100 percent of such losses in excess of a $100,000 retention up to $5 million.

While the losses which eventually maxed out the policy took place during 1994, their roots stretch back to the bank's hiring of Randall Rimington nearly a decade before to head up its trust operations. At the time, Rimington seemed like a good choice--coming to Citizens from a top managerial position in a competitor's trust department, he had hoped that the new position would allow him to serve Princeton's small, close-knit community. In retrospect, the bank's decision to put Rimington in charge of its customers' trust accounts looks a lot like Mister Burns' determination that Homer Simpson would make a fine nuclear safety inspector.

Rimington's less-than-prudent nature emerged in the early 1990's when he began heavily investing in collateralized mortgage obligations (CMOs). A CMO is a security backed by a pool of real estate mortgages. Like a regular bond, it pays fixed interest on the underlying principal and its price fluctuates with interest rates--falling when rates go up, rising when rates decline. Unlike normal bonds, CMOs do not have a set life- span. Instead, investors recover their principal when the underlying mortgages get paid off. This means that the life of a CMO turns on interest rates: if rates rise, homeowners tend to hang onto their mortgages and the life of the CMO extends; if rates fall, people will likely prepay their existing obligations to take advantage of the lower rates and the life of the security will decrease. Thus, rising rates leave the owner of a CMO with an investment that is losing value and won't pay off anytime soon.

Mixed in with the regular CMOs, Rimington added a variety called "inverse floaters." These securities share all the characteristics of a basic CMO except that instead of paying a fixed return to holders for the life of the bond, the interest payments fluctuate inversely to changes in an index. In this case, the inverse floaters' yields were determined according to a formula keyed to the London Interbank Offering Rate (LIBOR)--a well-known index in the financial markets measuring interest rates. The precise formula governing the return of Citizens' inverse floaters lies outside the scope of this opinion, but it is worth noting its basic effect: even small interest rate changes would drastically affect the inverse floaters' returns; a sharp rise would drop the yield to zero.

Surveying the financial landscape, Rimington decided that CMOs, and particularly inverse floaters, offered a near-perfect (in his words, an "aggressively conservative") buying opportunity. He reasoned that since the securities provided high rates of return without risking the underlying capital--the holder would eventually recover the security's face value when the mortgages got paid off--he could meet his customers' income needs without taking any serious gambles. He also thought that since each variety of CMO was backed by a separate pool of mortgages, and these pools varied--each containing a group of mortgages of differing lengths from different regions of the country-- the economic and structural factors affecting each security would cancel out any risk inherent in creating a monolithic portfolio.

To take advantage of this "aggressively conservative" investment opportunity, Rimington began to buy. He purchased derivatives for his parents, other relatives and, most relevantly, his trust customers.1 One of his customers, an 80-year-old nursing home resident in poor health, had a portfolio made up of 83 percent CMOs; others had 100 percent exposure. Taken as a whole, by the end of Rimington's tenure (more to come), CMO's accounted for nearly 50 percent of Citizens' total trust assets, and almost one- third of these were inverse floaters.

For a time, the derivatives brought home nice returns, and neither Rimington nor anyone else at Citizens found fault with his strategy. In early 1994, however, interest rates started to rise and the trust accounts began to show huge paper losses. As customers watched their life savings vanish, several wrote letters to the bank stating that the investments were inappropriate and asking for their money back. These complaints, coupled with the mounting losses, caused Citizens' trust committee to finally focus in a little more closely on Rimington's activities. Horrified by what they found, in October of 1994 the bank alerted both the Office of the Controller of the Currency (OCC) and Cincinnati that the trust department was in deep trouble. The OCC promptly investigated, and in late February 1995 the agency issued a confidential report to Citizens stating that the bank had breached its fiduciary duty of care to its trust customers by purchasing such high-risk securities. The report concluded that Rimington's lack of investment expertise and training had caused him to become overly reliant on broker recommendations in making investment decisions and that he had purchased many of the securities "without full knowledge of the risks involved." The agency also decried the bank's oversight, stating that Citizens' trust committee had approved the purchase of the CMOs without considering their inherent risks. But despite its harsh criticism of the bank's multiple screw-ups, the OCC did not find that anyone at Citizens had engaged in any intentional wrongdoing or that the investments created any conflicts of interest between the bank or its personnel and its trust customers.

In light of its findings, the OCC told Citizens to liquidate the derivatives and reimburse its customers for any associated losses. Rimington, who still argued that the CMOs were "totally appropriate," strongly disagreed with both the report's conclusions and the OCC's directives. He argued that if the bank simply held onto the investments they would eventually pay off. In light of this stance, Citizens decided that Rimington's continued employment would only slow the trust department's recovery, so he was let go. The bank then followed the OCC's orders-- liquidating the inverse floaters for a loss of more than $4.8 million and unloading the remaining high-risk CMOs for additional losses of over $565,000.

With Rimington out of the picture, the derivatives out of the trust holders' portfolios, and the bank out more than five million bucks, the rest of our story concerns Citizens' attempt to recoup its losses from Cincinnati under the D&O policy. As we said, this effort began in October 1994 when the bank first alerted Cincinnati of the trust department's troubles. This notification was not a formal claim; Citizens simply forwarded to Cincinnati the customers' letters of complaint and mentioned that the problems which gave rise to the letters existed in a number of other accounts. In a prompt response, Cincinnati acknowledged Citizens' missive as a "notice of potential claims," stated that it would begin an investigation, and noted that "a determination of actual coverage applicable will have to await the actual filing of a lawsuit against the bank, directors and officers or plan administrators."

After meetings with Citizens in November and December (in both of which, Rimington assured Cincinnati that the investments were appropriate and that the claims might well go away), Cincinnati restated its determination that the situation presented a notice of potential claims and requested further information allowing it to evaluate its coverage obligations if and when the customers filed suit. However, the insurer also expressed an interest in resolving any coverage issues short of a customer suit if this would enable Citizens to reach a favorable settlement with its disgruntled customers. Finally, Cincinnati once again clearly set forth that it reserved all defenses under the policy if and when the claims were to arise.

When Citizens responded that the letters constituted claims under the policy, Cincinnati again disagreed. This time, however, the insurer decided that the parties should turn to case law to settle the dispute. Thus, in an April 1995 letter, Cincinnati explained that in Harbor Insurance Co. v. Continental Bank Corp., 922 F.2d 357 (7th Cir. 1991), "the Seventh Circuit intimated that a claim is not made until an insured is specifically named in a lawsuit." It then invited Citizens to reply in kind, stating that it would be happy to consider any cases upon which the bank relied to contest the point.

Rather than respond to this invitation, in August 1995 Citizens decided to skirt the problem by sending Cincinnati a "formal demand for indemnity" outlining its losses and the reasons it believed they fell under the policy. To this, Cincinnati replied that it did not possess enough information about the liquidation to determine whether it constituted the settlement of a claim. Instead, it stated that based on its current knowledge, the liquidation appeared to be a business decision designed to deflect the negative publicity that would accompany the release of the OCC's report and would thus not constitute a "Loss" derived from a claim. In addition, Cincinnati asserted that despite its limited understanding of the facts, it appeared that if the demand were to constitute a claim, it would fall under exclusion 4(c) which denied coverage for claims that arose from actions taken in bad faith. With that, the insurer once again requested the information it deemed necessary to process the claims.

Convinced that no amount of information could get Cincinnati to pay the claim voluntarily, Citizens failed to provide its insurer with any new facts. This began a standoff in which Citizens continually repeated its demands for repayment while Cincinnati responded that it could not assess the claims without a better understanding of the context in which the losses arose.

In June 1996 the logjam broke when a frustrated Citizens filed suit. At trial, Cincinnati acknowledged that the bank had breached its fiduciary obligation to its trust customers. It also abandoned its position that Citizens had failed to make a claim. Instead, the insurer continued to assert that the claims were barred by exclusion 4(c) and added two new defenses: (1) that exclusion 4(a) operated to deny the claims because they arose out of "dishonest, fraudulent, criminal, or malicious acts"; and (2) that exclusion 4(j) precluded Citizens' recovery because the claims arose from the "method of funding" the trust accounts.

As we said, Judge Leinenweber rejected these defenses. He decided that exclusions 4(a) and 4(c) did not apply because Rimington acted with "a pure heart, but an empty head" and that neither he nor the bank had engaged in any sort of intentional wrongdoing. He then found that since the meaning of "funding" was ambiguous, Cincinnati could not rely on exclusion 4(j) to deny coverage. He thus awarded Citizens $4.9 million under the policy. Then, stating that he believed Cincinnati had "unreasonably and vexatiously" denied the bank's claim, the judge hit the insurer with Citizens' attorneys fees and costs.

The first issue for review is whether Judge Leinenweber's finding that Rimington possessed a "pure heart, but empty head" constituted clear error. Cincinnati claims that the record offered no support for the judge's finding and ample support for its opposite. In its most basic form, its argument boils down to an assertion that, given the level of understanding Rimington displayed about the structure and behavior of the derivatives during trial, he must have known he was playing Russian roulette with his customers' life savings. As this would show a deliberate disregard for his customers, his fiduciary responsibilities, and Citizens' internal rules, Cincinnati believes that "Rimington could not possibly have acted with a pure heart."

Judge Leinenweber's determination that Rimington's heart was pure is a finding of fact that we review only for clear error. We will not reverse his determination unless it strikes us as wrong with the force of a 5-week-old, unrefrigerated, dead fish. See Parts and Elec. Motors, Inc. v. Sterling Elec., Inc., 866 F.2d 228, 233 (7th Cir. 1988) (Bauer, J.). Cincinnati cannot clear this titanic olfactory hurdle. While it strains credulity to believe that Rimington could have understood how the derivatives worked (he did) and still think them an appropriate, "conservatively aggressive" investment, it appears that this was just the case. Not only were his pure intentions underlined by his willingness to pile his family's and colleagues' fortunes into the securities, but it seems he might have held onto his job had he admitted his mistakes. Instead, he stuck by his risk analysis and went down with the ship. In fact, Rimington still hasn't lost his affinity for the CMOs. At trial, after a passionate defense of the derivatives, he stated, "I believed in those securities and I still do." Judge Leinenweber accepted this explanation, and as he had ample evidence to do so, his finding was not clearly erroneous.

Cincinnati next appeals the court's interpretation of the term "bad faith" in exclusion 4(c). In his opinion, Judge Leinenweber looked to Black's Law Dictionary to find that bad faith "implies some level of intentional malfeasance." He then explained that if Cincinnati had intended the term to cover innocent breaches of fiduciary duty, it could have drafted the policy to expressly address such situations. Without such a clause, he found that the insurer's efforts to rely on the exclusion were simply an attempt to equate bad faith with breach of fiduciary duty--an untenable position for a policy expressly covering fiduciary breaches.

Cincinnati argues that Judge Leinenweber's analysis not only robs bad faith of its natural meaning within the context of the contract, but that it contradicts a number of trust cases holding that a trustee who has recklessly disregarded the limits on her powers cannot defend a suit for breach of duty on good faith grounds (i.e., she's liable because she acted in bad faith). See, e.g., United States Nat'l Bank & Trust Co. v. Sullivan, 69 F.2d 412, 415 (7th Cir 1934). Based on this line of authority, the insurer asserts that in the context of trusts, bad faith should cover a reckless breach of duty. And since it believes both Rimington and the trust committee acted recklessly, Cincinnati concludes that the exclusion must preclude the bank's recovery.

Since interpreting an insurance contract is a question of law, we review Judge Leinenweber's decision de novo. See River v. Commercial Life Ins. Co., 160 F.3d 1164, 1169 (7th Cir. 1998). To begin this inquiry, we note that both sides agree that bad faith has no set meaning within the context of D&O insurance policies covering fiduciary breaches in Illinois. However, Illinois law is settled on one relevant point: exclusionary clauses may only deny coverage where their application is clear and free from doubt. See, e.g., Continental Cas. Co. v. McDowell & Colantoni, Ltd., 282 Ill. App. 3d 236, 241, 668 N.E.2d 59, 62 (1996). If two reasonable interpretations of a policy term exist, the ambiguity must be settled in favor of coverage. See e.g., River, 160 F.3d at 1169. Thus, we examine Cincinnati's arguments to determine if its interpretation is so obvious that Citizens would have been unreasonable to believe that pure-hearted, empty-headed conduct would fall under the policy.

Cincinnati asserts that since exclusion 4(a) covers "dishonest, fraudulent, criminal, or malicious acts"--all of which require some degree of intentional wrongdoing-- reading bad faith to impose the same requirement would render 4(a) unnecessary. As Illinois law requires that each clause of an insurance policy must be presumed to have been inserted for a purpose, see Home and Auto Ins. Co. v. Scharli, 10 Ill. App. 3d 133, 136, 293 N.E.2d 914, 916 (1973), Cincinnati concludes that we must define bad faith so as to give both exclusions 4(a) and 4(c) distinct meanings.

This argument would be persuasive if the term had two equally reasonable (and readily apparent) definitions in the context of the parties' contract. But Cincinnati's assertion that bad faith covers reckless breaches of fiduciary duty does not offer such an alternative. The trust cases it cites in support of this definition simply hold that a fiduciary who has acted recklessly cannot avoid liability because she has performed inappropriate acts in good faith. In these cases, the good faith defense protects trustees who have carried out their responsibilities in accordance with their fiduciary duty but have made mistakes. In other words, if they acted in good faith they did not breach a fiduciary duty. Thus, even if we accept Cincinnati's assertion that bad faith is simply the mirror image of good faith, the cases it cites would define bad faith to describe a simple breach of fiduciary duty. Because the contract primarily covers such breaches, Cincinnati does not push this position. Instead, it argues that bad faith refers only to reckless breaches. But since the insurer points to nothing that supports a definition that distinguishes between ordinary fiduciary breaches and reckless ones, its argument fails.

In so holding, we do not attempt to define bad faith for Illinois insurers--that is a job best left to the state's courts. Rather, we conclude only that Citizens' (and Judge Leinenweber's) proposed definition of the term to include some sort of intentional wrongdoing was reasonable under the circumstances. Exclusion 4(c) was insufficiently clear to preclude recovery under Illinois law.

Cincinnati's next claim--that exclusion 4(j) should have operated to deny coverage--falls victim to the same analysis. Cincinnati argues that "method of funding" includes anything related to investing a trust account's money. Citizens counters that the term typically refers only to the initial method of placing money into the trust. Both sides agree that no Illinois court has interpreted the exclusion.

Once again, we need not define the term to decide the case. We must only determine if Citizens attached a bizarre meaning to the phrase or if Cincinnati's proposed alternative was pellucid. Because we agree with Judge Leinenweber--both parties offered reasonable interpretations--Cincinnati cannot look to the exclusion to deny coverage under Illinois law. And the law makes sense: If Cincinnati wants to rely on exclusions to deny coverage--especially where it seeks to reject a claim arising from an inherent risk in the insured's core activity--it should either write its exclusions with terms whose meanings are not subject to debate, or define the terms within the policy.2

Next, we take up Judge Leinenweber's award of attorneys fees and costs under sec. 155 of the Illinois Insurance Code, a decision we review for abuse of discretion. Smith v. Equitable Life Assurance Soc'y, 67 F.3d 611, 618 (7th Cir. 1995). The judge found that Cincinnati "unaccountably mishandled the bank's claim." He explained that while exclusion 4(j) provided a legitimate defense at trial, Cincinnati could not explain why it waited until suit was filed to press it. Instead, he noted that Cincinnati only turned to exclusion 4(j) because it's primary defense--that Citizens could not recover because no claim had been made against the bank under the policy--was "found to be of no value whatsoever." Also, he noted that Cincinnati never tried to cast doubt on the amount of Citizens' losses.

Section 155 provides that an award of attorneys fees and costs is appropriate if insurers' actions are "vexatious and unreasonable." 215 Ill. Comp. Stat. 5/155 (West 1999). Because this statute is "penal in nature" its provisions must be strictly construed. Morris v. Auto-Owners Ins. Co., 239 Ill. App. 3d 500, 509, 606 N.E.2d 1299, 1305 (1993). Attorneys fees may not be awarded simply because an insurer takes an unsuccessful position in litigation, but only where the evidence shows that the insurer's behavior was willful and without reasonable cause. Id. This means that an insurer's conduct is not vexatious and unreasonable if: (1) there is a bona fide dispute concerning the scope and application of insurance coverage, Green v. International Ins. Co., 238 Ill. App. 3d 929, 935, 605 N.E.2d 1125, 1129 (1992); (2) the insurer asserts a legitimate policy defense, Cummings Foods, Inc. v. Great Central Ins. Co., 108 Ill. App. 3d 250, 259, 439 N.E.2d 37, 44 (1982); (3) the claim presents a genuine legal or factual issue regarding coverage, Lazzara v. Esser, 622 F. Supp. 382, 386 (N.D. Ill. 1985); or (4) the insurer takes a reasonable legal position on an unsettled issue of law. Martz v. Union Labor Life Ins. Co., 573 F. Supp. 580, 586 (N.D. Ill. 1983), rev'd on other grounds, 757 F.2d 135 (1985).

Under these standards, the sec. 155 ruling constituted an abuse of discretion. The judge did not find that Cincinnati failed to present reasonable defenses at trial, but rather that the insurer could not wait until trial to raise a valid defense. But this was not the case. While the judge correctly concluded that Cincinnati could likely have asserted its position with respect to exclusion 4(j) earlier, the point is only relevant if the insurer's reasons for denying coverage up to that point were baseless. They were not.

From August 1995 until trial, Cincinnati asserted that without a better understanding of the facts it could not determine whether the bank's claims fell under the policy. Citizens asserts that it cannot be blamed for Cincinnati's lack of knowledge because it could not disclose the contents of the OCC report. But while the contents of the report were restricted, this does not excuse the bank's less than stellar efforts to provide Cincinnati with the information it requested. Cincinnati had a right to try to figure out what happened in the trust department. Waste Management Inc. v. International Surplus Lines Ins. Co., 144 Ill. 2d 178, 204, 579 N.E.2d 322, 333 (1991). And it was right to inquire-- Judge Leinenweber stated that Rimington's behavior "seemed suspicious" and that the losses could be traced to the trust department's "ineffective . . . oversight, lax supervision, weak risk management practices, inadequate internal audit function, lack of expertise, and poor portfolio practices." While the judge correctly determined that the losses stemming from these activities fell under the policy, this required him to decide a number of hotly contested legal and factual issues. To expect Cincinnati to make its coverage decision without access to the facts which gave rise to such close questions would be, we think, a bit too much to ask.

Prior to August 1995 Cincinnati argued that the customer's letters did not constitute claims. Its argument in support of this position would not have set the legal community on fire, but it was presented with reasoned support. Furthermore, the insurer did not vexatiously assert the defense to deny Citizens' claims. On the contrary, Cincinnati more than once statedthat it would be interested in exploring any settlement opportunities in spite of its position, and it extended an invitation to the bank to provide it with support for Citizens' alternate interpretation of the term. Cincinnati cannot be faulted because Citizens declined to follow up on these offers.

It seems to us that the judge's assessment of Cincinnati's assertion that the letters did not present claims was driven largely by Cincinnati's failure to continue to press the defense at trial. But by that point Citizens had responded to Cincinnati's argument from Harbor, and the insurer realized that it had superior defenses in its arsenal. Cincinnati was entitled to change strategy.

Cincinnati's failure to contest the amount of Citizens' losses also cannot support the sec. 155 award. As stated, the insurer asserted meritorious defenses at trial once it possessed a full understanding of the facts. In light of these defenses, its decision not to contest Citizens' losses is irrelevant.

This brings us to Citizens' cross-appeal. The bank claims that Judge Leinenweber erred in subtracting the $100,000 retention from its award since it presented over $5.4 million in losses. Cincinnati concedes that this was error but, without citation, argues that since Citizens failed to file a corrective motion to the trial court, it should lose its ability to contest the award. We know of no such rule.

Judge Leinenweber correctly found Cincinnati liable under the policy, but the award should have been for $5 million, not $4.9 million. However, Citizens' award of attorneys fees and costs under sec. 155 of the Illinois Insurance Code is vacated. So, more than 90 percent of this judgment is affirmed, and the case is remanded for the entry of an amended judgment consistent with this opinion. Each side shall bear its own costs.

SO ORDERED.