SYKES, Circuit Judge.

First National Bank of Manitowoc extended credit to a local used-car dealership based in part on the dealership's presentation of leases signed by its customers. Unbeknownst to the Bank, in many instances the dealership's president forged customers' signatures on leases that were fabricated or altered. The dealership eventually defaulted on the loans, and the Bank lost more than $1.7 million. The Bank filed a claim for the loss with its insurer, Cincinnati Insurance Company. The policy Cincinnati had issued to the Bank was similar but not identical to an outdated version of the standard Bankers Blanket Bond, now known as a Financial Institutions Bond. Generally speaking, these bonds provide coverage to financial institutions for losses caused by specified dishonest, fraudulent, or criminal acts.

Cincinnati denied the Bank's claim and this suit ensued. Both parties moved for summary judgment. The district court denied Cincinnati's motion and granted the Bank's in substantial part, rejecting only its claim for statutory interest. Both parties appealed. Because the Cincinnati policy covers the Bank's losses and statutory interest was properly denied, we affirm.

I. Background

First National Bank of Manitowoc is a national bank headquartered in Manitowoc, Wisconsin. In 1991 the Bank began doing business with West Town Auto, Inc., a used-car dealership also located in Manitowoc. The Bank had several lending relationships with West Town, including a line of credit through which West Town purchased vehicles to lease. West Town would enter into a preliminary lease agreement with a customer at the dealership, and Lee Kust, West Town's president, would procure a loan to purchase the vehicle.

Kust would call the Bank or fax it the lease terms and wait for the Bank's approval. Once the lease was approved (sometimes several days later), Kust would finalize the transaction with his customer and bring the signed lease agreement to the Bank. At that time Kust would execute several documents, including a business note, an assignment of lease payments, and a chattel security agreement granting the Bank a security interest in the vehicle. Under the terms of its line of credit with the Bank, West Town was responsible for making loan payments to the Bank; West Town's customers made their lease payments directly to West Town.

The facts surrounding Kust's fraud are undisputed. The scam worked in one of two ways: Kust either fabricated a lease agreement for a nonexistent vehicle and transaction or altered the terms of a valid lease agreement and submitted the altered version to the Bank. (As examples of the latter fraud, Kust would alter a vehicle's condition, make, or model, thereby enabling him to obtain a larger loan.) Under both scenarios, Kust forged customer signatures by tracing a valid signature onto a fabricated or altered lease form. In 2001 Kust suddenly disappeared and West Town defaulted on the loan. Until then, however, West Town had been making monthly payments as required, although the Bank had assessed late charges on several occasions.

Cincinnati does not suggest that any Bank employees were aware of Kust's fraudulent scheme, but the insurer does point to what it says are "red flags" during the course of the lending relationship that it believes ought to affect coverage. For example, the Bank did not have a copy of each vehicle's certificate of title and relied on Kust to record and perfect its security interest. Bank employees were aware that lien confirmations were not on file for many vehicles, and those that were on file contained discrepancies (the vehicle identification number ("VIN") on the confirmations did not always match the VIN number identified on the lease and loan documents). On several occasions one VIN number served as collateral for two separate loans. After Kust disappeared it only took a few phone calls to West Town lessees for the Bank to realize there was a problem.

After the Bank assessed its losses, it sought coverage under an insurance policy it had purchased from Cincinnati in 2001 called the Depository Institutions Blanket Bond, No. B80-534208. The Cincinnati Bond borrows from the Bankers Blanket Bond, Standard Form No. 24, an industry-standard insurance policy for commercial banks offered by several carriers. The standard Bankers Blanket Bond is "a two-party agreement between the underwriter and the insured financial institution, pursuant to which the underwriter agrees to indemnify the insured against loss sustained by reason of specific perils described under six `Insuring Agreements,' which are commonly referred to by the letter designating them in the bond." Peter I. Broeman, An Overview of the Financial Institution Bond, Standard Form No. 24, 110 BANKING L.J. 439, 439-40 (1993). The standard Bond also includes several exclusions which subtract from coverage provided by the insuring agreements. Cont'l Corp. v. Aetna Cos. Sur. Co., 892 F.2d 540, 546 (7th Cir.1989) ("[E]xclusions are expressly intended to modify coverage clauses and to limit their scope."); D Angela v. Cornell Paperboard Prods. Co., 59 Wis.2d 46, 207 N.W.2d 846, 849 (1973); Bulen v. West Bend Mut. Ins. Co., 125 Wis.2d 259, 371 N.W.2d 392, 394 (Ct.App. 1985).

Here, we are primarily concerned with Insuring Agreement E and Exclusion H. Insuring Agreement E covers loss resulting from a financial institution's good-faith reliance on forged or counterfeit documents. Cincinnati's version of Insuring Agreement E reads as follows:

E. ALL RISK FORGERY

Loss by reason of the Insured (a) having in good faith and in the usual course of business . . . extended any credit or assumed any liability or otherwise acted upon any security, document, or other written instrument which proves to have been a forgery or to have been altered or raised or counterfeited. . . .

. . . .

Actual physical possession of such security, document or other written instrument by the Insured . . . is a condition precedent to the Insured's having relied on the faith of, or otherwise acted upon, such security, document or, other written instrument.

Forgery is defined in the Cincinnati policy as "the signing of the name of another with intent to deceive." Exclusion H excludes coverage for "loss caused by an Employee, except when covered under Insuring Agreement A." (Insuring Agreement A covers losses "resulting directly from dishonest or fraudulent acts of an Employee.")

The Bank submitted a Proof of Claim to Cincinnati for coverage under Insuring Agreements D and E of the Policy. Cincinnati denied coverage and the Bank filed suit in state court. Cincinnati removed the case to federal court based on diversity jurisdiction, and both parties moved for summary judgment. The district court denied Cincinnati's motion and granted the Bank's motion in part, holding that Insuring Agreement E covered the Bank's loss but that questions of fact existed with respect to damages. To move the case along, the parties stipulated to damages (about $1.75 million), and the district court awarded the Bank common-law prejudgment interest at a rate of 5%. The court denied the Bank's request for statutory interest at the higher rate of 12%.

Both parties appealed. Cincinnati appealed from the district court's orders partially granting the Bank's motion for summary judgment on coverage, denying Cincinnati's motion for summary judgment, denying Cincinnati's motion to strike an affidavit, and awarding the Bank common-law prejudgment interest. The Bank appealed from the district court's order awarding interest, arguing that it is entitled to statutory prejudgment interest at a rate of 12% under section 628.46 of the Wisconsin Statutes.

II. Discussion

We review a district court's grant of summary judgment de novo. "With cross summary judgment motions, we construe all facts and inferences therefrom `in favor of the party against whom the motion under consideration is made.'" In re United Air Lines, Inc., 453 F.3d 463, 468 (7th Cir.2006) (quoting Kort v. Diversified Collection Servs., Inc., 394 F.3d 530, 536 (7th Cir.2005)). Summary judgment is appropriate if "there is no genuine issue as to any material fact and . . . the moving party is entitled to a judgment as a matter of law." FED.R.CIV.P. 56(c). The parties agree that Wisconsin law governs this diversity suit.

A. Coverage Under the Cincinnati Bond

The interpretation of an insurance policy is a question of law that is reviewed de novo. Cont'l Corp., 892 F.2d at 543 (citing Kraemer Bros. v. United States Fire Ins. Co., 89 Wis.2d 555, 278 N.W.2d 857, 860 (1979)); Folkman v. Quamme, 264 Wis.2d 617, 665 N.W.2d 857, 864 (2003). An insurance policy is construed to give effect to the intent of the parties as expressed in the language of the policy, which is interpreted as a reasonable person in the position of the insured would understand it. Folkman, 665 N.W.2d at 864; Danbeck v. Am. Family Mut. Ins. Co., 245 Wis.2d 186, 629 N.W.2d 150, 153 (2001). If the language of the policy is plain and unambiguous, it is enforced as written, without resort to rules of construction. Folkman, 665 N.W.2d at 864; Danbeck, 629 N.W.2d at 154. Policy language is interpreted not in isolation but in the context of the policy as a whole. Folkman, 665 N.W.2d at 866. If the policy language is ambiguous, it is construed against the insurer and in favor of coverage. Id. at 864; Frost ex rel. Anderson v. Whitbeck, 257 Wis.2d 80, 654 N.W.2d 225, 230 (2002) ("If terms in an insurance policy are ambiguous, they should be construed against the insurance company that drafted the policy."); Danbeck, 629 N.W.2d at 154.

This last principle, however, generally does not apply where the policy in question is a standard "fidelity" or Bankers Blanket Bond, drafted by representatives from both the banking and insurance industries. Tri City Nat'l Bank v. Fed. Ins. Co., 268 Wis.2d 785, 674 N.W.2d 617, 621-22 (Ct.App. 2003) (citing State Bank of Viroqua v. Capitol Indem. Corp., 61 Wis.2d 699, 214 N.W.2d 42, 43 n. 1 (Wis. 1974)) ("These bonds are not the usual contracts of adhesion and the familiar rule of interpreting a contract strictly against the insurer and liberally in favor of the insured should not apply."); Sharp v. Fed. Sav. Loan Ins. Corp., 858 F.2d 1042, 1046 (5th Cir.1988) (the principle that insurance contracts are to be construed against the underwriter does not apply when "the contract was in fact a joint effort of both insurers and the insureds"). As best we can tell, the language of the Cincinnati Bond appears to generally — but not uniformly — resemble an older version of the standard Bankers Blanket Bond. Accordingly, any ambiguity in language unique to the Cincinnati Bond may be resolved by reference to the general practice of construing ambiguities against the insurer/drafter and in favor of coverage.

A brief history of the standard Bankers Blanket Bond is in order. The Surety Association of America drafted the first American Bankers Blanket Bond in 1916. Edward G. Gallagher, A Concise History of Standard Form No. 24, 1986 Edition, in ANNOTATED FINANCIAL INSTITUTION BOND 5, 5 (Michael Keeley ed., 2d ed.2004). By 1941 the Bond had undergone several revisions — with input from the American Bankers Association and other trade groups — and was termed the "Bankers Blanket Bond, Standard Form No. 24." Broeman, An Overview of the Financial Institution Bond supra, at 443; Gallagher, A Concise History, supra, at 6. Additional revisions were made over the years, and beginning with the 1986 revision, the Bond was renamed the "Financial Institution Bond, Standard Form No. 24." Gallagher, A Concise History, supra, at 5. The standard Bond contains six Insuring Agreements (Insuring Agreements A-F) which cover the insured financial institution against loss arising from specified dishonest, fraudulent, or criminal acts. Broeman, An Overview of the Financial Institution Bond supra, at 440.

Because there are several revisions of the Bond in circulation, courts initially ought to determine which version, if any, the policy in question adopts; case law interpreting one revision may be unhelpful or even irrelevant to the task of interpreting another. The language of the Cincinnati Bond is not entirely consistent with the standard Bond version, the 1986 revision, in use at the time it was issued. For example, the Cincinnati Bond adds eight Insuring Agreements to the standard Bond's six, covering such additional risks as "Directors' and Officers' Expenses in Defending Suits," "All Risk Safe Deposit Box," and "Audit Expense." Insuring Agreement E, at issue here, is called "All Risk Forgery" and most closely resembles Insuring Agreement E as it appeared in the 1951 and 1969 revisions of the standard Bond, though it does not precisely track the language of either of these versions.

As noted above, Insuring Agreement E in the Cincinnati Bond covers "loss by reason of the Insured . . . having in good faith and in the usual course of business . . . extended any credit or assumed any liability or otherwise acted upon any security, document, or other written instrument which proves to have been a forgery or to have been altered or raised or counterfeit." Cincinnati argues that the "in good faith and in the usual course of business" language imposes a duty on the Bank to follow "sound banking practices" in connection with the events underlying the claim, and that this is a condition for coverage. We disagree. Although we interpret the policy as a whole, to interpret "in good faith" and "in the usual course of business" as together imposing a prerequisite normative standard of banking conduct ignores the independent meaning of each phrase. Rabinovitz v. Travelers Ins. Co., 11 Wis.2d 545, 105 N.W.2d 807, 811 (1960) ("Some meaning must be given to each sentence, phrase, and word used, and when this may fairly and properly be done, no part of the language used can be rejected as superfluous or unmeaning.").

Neither phrase is defined by the policy, but the Bank points to section 401.201(19) of the Wisconsin Statutes, adopting the UCC definition of "good faith" as "honesty in fact in the conduct or transaction concerned." We have stated that "`good faith' usually establishes a subjective standard," and pointed out that "[m]any negligent acts are committed with pure hearts and empty heads." State Bank of the Lakes v. Kan. Bankers Sur. Co., 328 F.3d 906, 909 (7th Cir.2003). Cincinnati asserts there are material issues of fact regarding whether the Bank was "selectively ignorant" in extending credit to Kust; however, its corporate designee conceded that Bank employees acted honestly and in good faith, with no knowledge of Kust's fraudulent scheme. We hold the good-faith requirement does not impose a "sound business practices" prerequisite to coverage.

This leaves the question of whether the phrase "in the usual course of business" means "consistent with sound business practices" or otherwise imposes a particular standard of conduct on the Bank for its loss to be covered. On its face, the phrase does not suggest a duty of care but, rather, a certain category of acts — i.e., those usually conducted in the banking business. Because the language of the Cincinnati Bond is not standard in this respect, bond-specific case law provides little guidance. However, Wisconsin courts that have addressed this phrase in other contexts have understood it to mean actions normally taken by a bank. Fid. Deposit Co. of Md. v. Peoples Exch. Bank of Thorp, 270 Wis. 415, 71 N.W.2d 290, 292 (1955) ("The check here was complete and regular on its face and, so far as the . . . bank was concerned, it had no notice of any infirmity in it or any defect in the title of the person cashing it, and took it in the usual course of business."); Banking Comm'n v. First Wis. Nat'l Bank of Milwaukee, 234 Wis. 60, 290 N.W. 735, 749 (1940) ("The usual course of business upon a bank loan is to credit the account of the borrowing customer and respond to checks.").

This is the interpretation the district court adopted in its well-reasoned opinion, and we agree. Because the Bank acted "upon the kinds of documents that it would normally act upon in its business, such as leases, checks, securities, etc., rather than documents outside that usual course," the Bank acted in the usual course of business. Fid. Deposit Co. of Md., 71 N.W.2d at 292.

Cincinnati also argues that Insuring Agreement E covers only losses directly caused by forgery and not losses arising from loans made on forged documents describing nonexistent assets or transactions. Here, Cincinnati argues, the forged signatures on the leases did not directly cause the Bank's loss, the absence of collateral did. This argument ignores the plain language of Insuring Agreement E, which covers "loss by reason of the Insured having in good faith and in the usual course of business . . . extended any credit or . . . otherwise acted upon any . . . document . . . which proves to have been a forgery." The Bank's loss easily comes within this language; the Bank sustained a loss because it extended credit to West Town based on vehicle leases which proved to be forgeries.

Insuring Agreement D in the Cincinnati Bond, entitled "Forgery, Alteration and Unauthorized Signatures," covers (among other things) "loss resulting directly from . . . Forgery or alteration of, on or in any Negotiable instrument . . ., Acceptance, withdrawal order, receipt for the withdrawal of Property, Certificate of Deposit or Letter of Credit." Insuring Agreements D and E thus cover similar and potentially overlapping categories of loss, but the language of each is distinct: the former covers loss resulting directly from the forgery or alteration of certain documents, the latter covers loss "by reason of" the Bank having "extended any credit" or otherwise "acted upon" a document "which proves to have been a forgery." The coverage granted in Insuring Agreement E does not simply duplicate the coverage granted in Insuring Agreement D; Cincinnati's interpretation essentially conflates the two.

Cincinnati asserts that "courts have overwhelmingly held" that Insuring Agreement E does not cover losses from loans based on forged documents describing fictitious transactions or assets. This is not true. Of the cases Cincinnati cites, only four are appellate decisions. Two of the four concerned not loss causation but the so-called "actual physical possession" pre-requisite to coverage under Insuring Agreement E. See Republic Nat'l Bank of Miami v. Fid. Deposit Co. of Md., 894 F.2d 1255, 1262-63 (11th Cir.1990) (holding that the condition precedent of actual physical possession was unmet and suggesting in dicta that the standard Bankers Blanket Bond imposes a requirement of "commercially reasonable" reliance before Insuring Agreement E will cover a loss); Nat'l City Bank of Minneapolis v. St. Paul Fire Marine Ins. Co., 447 N.W.2d 171, 177 (Minn. 1989) (holding that the actual physical possession condition was not met and likewise suggesting in dicta a "sound business practices" reliance requirement). Another of Cincinnati's cited cases held that the record evidence established a fraud but not a forgery because the documents in question were not signed in the name of another. Charter Bank Nw. v. Evanston Ins. Co., 791 F.2d 379, 382 (5th Cir.1986). Here, it is undisputed that the leases were forgeries.

The remaining appellate decision Cincinnati cites on this point was scantily reasoned. Georgia Bank Trust v. Cincinnati Insurance Co., 245 Ga.App. 687, 538 S.E.2d 764, 766 (2000), involved a Cincinnati Bond similar to the one at issue here. Georgia Bank Trust extended credit based on forged documents confirming the existence of certain accounts that served as collateral for the loan. When the debtor defaulted, Georgia Bank filed a claim for its loss with Cincinnati. The Georgia court of appeals cited both Insuring Agreements D and E in its very brief opinion; without specifically addressing the language of either, the court accepted Cincinnati's argument that its Bond does not cover losses resulting from the nonexistence of assets assigned by a forged instrument. The court concluded that "the blanket bond did not protect the bank from its bad business deal. Even if the signature on the confirmation was authentic, the bank would have suffered the loss, because the assets did not exist." Id. This conclusion ignores the practical reality of the situation; but for the forged documents purporting to verify the existence of the collateral, credit would not have been extended in the first place, and there would have been no loss. It also ignores the plain language of Insuring Agreement E, which covers loss "by reason of" the Bank "having . . . extended any credit . . . or otherwise acted upon any . . . document" that "proves to have been a forgery." As here, the loss at issue in Georgia Bank easily fit within this coverage language. The case is unpersuasive and we decline to follow it.

So the Bank's loss is covered by Insuring Agreement E, and we are left with Cincinnati's argument that Exclusion H applies because Bank employees caused the loss. This argument is a nonstarter. Exclusion H states: "The Bond does not cover loss caused by an Employee. . . ." Cincinnati insists the Bank's employees caused the loss by failing to properly investigate the collateral presented by Kust. Had they done so, Cincinnati argues, they would have discovered the forgeries. As the district court noted, however, this interpretation of Exclusion H would eliminate coverage under Insuring Agreements D and E in all cases, as bank employees are intermediaries in every forgery-related bank loss. Exclusions are intended to subtract from or limit coverage in specified circumstances. Cont'l Corp., 892 F.2d at 546; Bulen, 371 N.W.2d at 394. They do not operate as complete cancellations of coverage granted in the insuring agreements. To the contrary, under Wisconsin law exclusions are narrowly construed, especially if their effect is uncertain. Am. Family Mut. Ins. Co. v. Am. Girl, Inc., 268 Wis.2d 16, 673 N.W.2d 65, 73 (2004); Cardinal v. Leader Nat'l Ins. Co., 166 Wis.2d 375, 480 N.W.2d 1, 3 (1992). We reject Cincinnati's expansive interpretation of Exclusion H; the exclusion does not apply here.

B. Prejudgment Interest

The district court's award of common-law prejudgment interest is reviewed for abuse of discretion. The district court awarded common-law prejudgment interest at a rate of 5%. See McRoberts Software, Inc. v. Media 100, Inc., 329 F.3d 557, 572 (7th Cir.2003). A plaintiff is entitled to prejudgment interest when the amount owed to it is "readily determinable" or when there is a "reasonably certain standard of measurement by . . . which one can ascertain the amount he owes." Olguin v. Allstate Ins. Co., 71 Wis.2d 160, 237 N.W.2d 694, 698 (1976). Cincinnati argues that because there was no reasonably certain standard of measurement — as evidenced by the Bank's changing its calculation midway through the litigation — the Bank is not entitled to prejudgment interest. This argument misses the mark. Amending a damages amount is not tantamount to conceding there is no reasonably certain standard of measurement. The Bank's losses can be calculated by analyzing the forged leases, the vehicles' Black Book and Kelley Blue Book values, and the amount the Bank was able to recoup by selling the remaining existing vehicles. Admittedly, it may be a cumbersome process, but this does not render the Bank's losses indeterminate.

Tellingly, any disputes the parties had with respect to damages dealt with the substantive issue of the Bank's entitlement to them, not their calculation. More specifically, Cincinnati argued that in certain instances the Bank acquired physical possession of the forged leases only after it issued funds on the line of credit — not before, as required by Insuring Agreement E. These losses are not covered. The parties were able to resolve these factual disputes by stipulating to a damages amount after the district court entered summary judgment. The district court did not abuse its discretion in awarding common-law prejudgment interest.

For its part, the Bank is not satisfied with 5% interest and argues it is entitled to statutory prejudgment interest at a rate of 12% pursuant to section 628.46(1) of the Wisconsin Statutes. Because the district court correctly interpreted section 628.46(1), we review the court's application of the statute to the facts for clear error. Thomas v. Gen. Motors Acceptance Corp., 288 F.3d 305, 307 (7th Cir.2002). Under section 628.46(1), an insurer must pay on a proof of claim within thirty days of receipt, with all overdue payments subject to interest at the rate of 12%, unless the insurer has "reasonable proof to establish that [it is] not responsible for the payment." Reasonable proof of nonresponsibility exists when coverage is "fairly debatable." Kontowicz v. Am. Standard Ins. Co. of Wis., 290 Wis.2d 302, 714 N.W.2d 105, 117 (2006). The district court held that coverage was "fairly debatable" here and we agree. The claim raised disputed questions about the scope of coverage under the forgery insuring agreements of an older and somewhat nonstandard Bankers Blanket Bond. There are no Wisconsin cases directly on point and little persuasive extrajurisdictional case law exists. This makes the claim fairly debatable. The Bank was not entitled to statutory prejudgment interest.

For the foregoing reasons, the Bank's losses are covered under Cincinnati's version of Insuring Agreement E, Exclusion H does not apply, and the Bank is entitled to common-law prejudgment interest at a rate of 5%. The orders of the district court are AFFIRMED.