COFFEY, Circuit Judge.

In this diversity jurisdiction case, we are called upon to decide whether an insurance policy fails to reflect the true intent of the parties as to the limits of coverage, and so should be reformed, or, alternatively, whether the wording of the policy is so ambiguous that the court must divine its meaning. The district court found that the words in the policy were unambiguous, though they failed to reflect the true intent of the parties. Accordingly, the court granted summary judgment to the plaintiff and reformed the contract to reflect that intent. Board of Trustees of the University of Illinois v. Insurance Corp. of Ireland, 750 F. Supp. 1375 (N.D.Ill. 1990). We affirm.

BACKGROUND

In early 1984 the Board of Trustees of the University of Illinois (U of I) requested quotes for excess public, hospital, and medical professional liability insurance coverage for a one-year period beginning March 1, 1984. March McLennan, an insurance brokerage, issued a quote of the Insurance Corporation of Ireland (ICI). The U of I accepted ICI's quote for coverage, and the parties agreed to a policy term of three years, from March 1, 1984 to March 1, 1987. This three-year term is somewhat misleading, however, because the contract provided that the policy was cancelable at the end of any year, and thus did not provide a full three years of coverage. Essentially, the U of I was purchasing one year of insurance, and both it and ICI had the option to renew, cancel, or modify the policy and its terms before the renewal date each year.

Before the policy term began the parties agreed to change the coverage date a little so that the annual premium would be due (and the year of coverage would commence) on July 1, 1984, the start of the U of I's fiscal year. This accommodation, however, left a gap in coverage between March 1, 1984 and July 30, 1984. To fill this gap ICI amended the policy with Endorsement # 3, which provided that the policy would now run from March 1, 1984 to July 1, 1987 (assuming the policy was renewed at the end of each year). The U of I, in return, paid a $165,000 premium for the extra four months of coverage (from March 1, 1984 to June 30, 1984), exactly one-third of the yearly $495,000 premium.

As to the extent of the coverage provided by ICI, the policy stated:

Item 1: LIMIT OF LIABILITY

$5,000,000 Combined Single Limit each and every occurrence and in the aggregate.

Excess of Self Insured Retention of $100,000 each and every occurrence $1,000,000 overall annual aggregate.

The parties agree that the latter clause gave the U of I a $100,000 deductible, meaning it was obligated to pay the first $100,000 on any claim, up to a total of $1,000,000 in each year. They disagree, however, as to the time period covered by the $5,000,000 aggregate limit. The U of I argues that it actually has $10,000,000 in coverage; $5,000,000 for the four-month period from March 1, 1984 to June 30, 1984 (colorfully dubbed the "stub period" by the district court, a tag we adopt for the purposes of this opinion), and $5,000,000 for the one-year period from July 1, 1984 to June 30, 1985. ICI, in contrast, maintains that there is a single $5,000,000 limit covering the entire sixteen months the policy was in effect (March 1, 1984 to June 30, 1985). Resolution of this disagreement is important for two reasons. First, this is an occurrence policy, not a claims-made policy, so even though it has been canceled, there is still a chance that more claims could be brought against it. Second, ICI settled a $4,500,000 claim for an accident during the stub period. This is important because the $5,000,000 limit is an aggregate, meaning that ICI will pay up to, but no more than, $5,000,000 for claims from the relevant period, no matter how many claims are made. Thus, if ICI's view is correct and the $5,000,000 limit applies to the entire sixteen-month period, the U of I has only $500,000 left to cover claims arising from occurrences during that time. But if the U of I's interpretation is proper, it may be close to exhausting its coverage for the stub period, but it has a fresh $5,000,000 available to cover claims arising from incidents during the one-year period.

In hopes of resolving this dispute, the U of I sued ICI, asking the court to either; (1) enter a declaratory judgment holding that the policy's description of coverage is ambiguous, and construe the ambiguity in its favor by finding that each of the two periods has a separate $5,000,000 limit, or (2) if the policy is not ambiguous, find that it does not reflect the true intent of the parties and reform it to provide separate $5,000,000 limits for the stub and one-year periods. The U of I then moved for summary judgment, but the district court found that the policy was not ambiguous, reasoning that as written the policy clearly provided for a single $5,000,000 limit. Nevertheless, the court held that the policy failed to reflect the parties' true intent as to coverage. Thus, the court reformed the policy to provide separate $5,000,000 limits for both the stub period and the one-year period.

ICI appeals, asserting that a genuine issue of material fact remains as to the intent of the parties regarding coverage under the policy. The U of I cross-appeals (though really it is just proposing an alternative ground for affirmance), arguing that even if reformation was improper, the policy is ambiguous and should be interpreted in its favor.

DISCUSSION

We review summary judgments de novo, viewing the record and all reasonable inferences that may be drawn from it in the light most favorable to the nonmovant. Karazanos v. Navistar Int'l Transp. Corp., 948 F.2d 332, 335 (7th Cir. 1991). Summary judgment is proper when "the pleadings, depositions, answers to interrogatories, and admissions on file, together with the affidavits, if any, show that there is no genuine issue as to any material fact and that the moving party is entitled to judgment a matter of law." Fed.R.Civ.P. 56(c). The movant must point to the portions of these papers demonstrating the lack of any genuine issue of material fact, and if the movant succeeds, the non-movant must go beyond the pleadings and find parts of affidavits, depositions, answers to interrogatories, or admissions on file demonstrating that a genuine issue does in fact remain for trial. Celotex Corp. v. Catrett, 477 U.S. 317, 323-24, 106 S.Ct. 2548, 2552-53, 91 L.Ed.2d 265 (1986). Further, in deciding whether the movant has met its burden, the court "must be guided by the substantive evidentiary standards that apply to the case." Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 255, 106 S.Ct. 2505, 2513-14, 91 L.Ed.2d 202 (1986). The relevant standard in contract reformation cases is clear and convincing evidence. Magnus v. Barrett, 197 Ill.App.3d 931, 145 Ill.Dec. 482, 485, 557 N.E.2d 252, 255 (1990).

We begin by examining the substantive law regarding the reformation of contracts, noting that insurance policies may be reformed for the same reasons as any other written contract. National Ben Franklin Insurance Co. v. Davidovitch, 123 Ill. App.3d 88, 78 Ill.Dec. 577, 580, 462 N.E.2d 696, 699 (1984). Normally contracts are to be taken at face value, and the written agreement is presumed to express the parties' intent. Beynon Building Corp. v. National Guardian Life Ins. Co., 118 Ill.App.3d 754, 74 Ill.Dec. 216, 223, 455 N.E.2d 246, 253 (1983). In certain cases, however, this presumption may be rebutted, and the court may at times step in and reform the contract. An Illinois appellate court recently described when reformation is proper.

[R]eformation of a contract should only be allowed when clear and convincing evidence compels the conclusion that the instrument as it stands does not properly reflect the true intention of the parties, and that there has been either a mutual mistake or a mistake by one party and fraud by the other.

Magnus, 145 Ill.Dec. at 485, 557 N.E.2d at 255. Put another way, "reformation is available when the parties, having reached an agreement and having then attempted to reduce it to writing, fail to express it correctly in the writing." Restatement (Second) of Contracts, § 155, comment a. The purpose of reformation is "to make a writing express the agreement that the parties intended it should." Id. It is an equitable remedy, performed by the court, provided that the intended agreement is certain enough to accurately rewrite the contract. Id. Where applied too liberally, reformation can do more harm than good, and therefore the plaintiff is required to prove the need for the remedy by clear and convincing evidence. See National Ben Franklin Ins. Co., 78 Ill.Dec. at 580, 462 N.E.2d at 699; see also Mutual of Omaha Ins. Co. v. Russell, 402 F.2d 339, 344 (10th Cir. 1968) ("reformation is an extraordinary remedy, and courts exercise it with great caution.") In the insurance context, reformation often occurs where the insurer and insured reach an agreement as to the terms of the policy, but for some reason these terms are either omitted or inaccurately expressed in the final written product. See Magnus, 145 Ill.Dec. at 485, 557 N.E.2d at 255.

The U of I attempts to meet its summary judgment burden with affidavits from James Gallivan, the university's Risk Manager; Paul Burston, the Marsh McLenan employee who acted as the middleman between the U of I and ICI in procuring the insurance policy; William Fleming, ICI's United States Manager who negotiated the policy with the U of I; and Frederick Radichel, ICI's Assistant United States Manager who performed the mechanical task of preparing the policy for typing by cutting and pasting clauses from old policies. Gallivan, Burston, and Fleming all stated under oath that the policy was intended to provide $5,000,000 in coverage for the stub period and another $5,000,000 for the one-year period. Radichel swore that he understood the policy to provide separate coverage limits for each billing period in Endorsement # 3, though his affidavit failed to state what he thought those limits were.

Attempting to show that the affidavits do not tell the whole story, ICI looks to the deposition responses of these four men. It notes: (1) that Gallivan admitted there was no separate bidding procedure to establish coverage during the stub period, and did not recall any specific instances when he had discussed the policy limits, (2) that Burston's affidavit may have been based on an interpretation of the policy, not his memory of actual negotiations, and (3) that Fleming remembered nothing of the negotiation of the policy and based his affidavit responses on a review of the policy, March McLennan's bid, and correspondence between the U of I, Marsh McLennan, and ICI. ICI argues that these facts undercut the strength of the affidavits, and that therefore a factual issue as to the parties' intent remains.

While the U of I's case would obviously be more convincing if these men were able to recollect some specific discussions concerning the coverage question. ICI has not shaken their credibility enough to create a triable issue of fact. "Where, as here, the movant's witnesses have been examined by the nonmovant in depositions, the nonmovant ordinarily must identify specific factual inconsistencies in the witness' testimony in order to withstand a motion for summary judgment." Corrugated Paper Products, Inc. v. Longview Fibre Co., 868 F.2d 908, 914 (7th Cir. 1989). Gallivan and Fleming's deposition testimony was consistent with the statements in their affidavits. Both affidavits stated that the affiant had reviewed certain documents and believed that the policy was to provide separate $5,000,000 limits for the stub and one-year periods. The depositions simply revealed that the coverage statements were based on the document review, rather than on a refreshed recollection of specific discussions during negotiation of the policy. This is not a factual inconsistency, nor does it indicate that the witnesses were being less than truthful or that it is possible that the parties had a different understanding as to coverage. Compare DF Activities Corp. v. Brown, 851 F.2d 920, 923 (7th Cir. 1988) (cautioning that where an affidavit is clear, ambiguous deposition responses should not be used to prevent summary judgment). Gallivan, Fleming, and Burston were the only three people involved in negotiating and implementing the policy, and they all swore that the policy was to provide separate $5,000,000 limits. Absent recantation of the affidavits in total or in part, the deposition testimony presented here is insufficient to create an issue of credibility, and thus is insufficient to bar the entry of summary judgment. Cf. Boston Five Cents Savings Bank v. Department of Housing, 768 F.2d 5, 8 (1st Cir. 1985) (where evidence as to contracting parties' intent is entirely one-sided, judge may take the matter from the jury).

The Fourth Circuit has decided a similar case, interpreting Illinois law. In American Employers Ins. Co. v. St. Paul Fire Marine Ins. Co., 594 F.2d 973 (4th Cir. 1979), two marine insurers argued over how much St. Paul was obligated to pay toward damages caused by exploding barges. The question was whether St. Paul had agreed to provide coverage on a "per occurrence" or "per vessel" basis, the latter exposing it to much greater liability. As written, the insurance policy seemed to provide "per vessel" coverage, but St. Paul argued that this was the result of a mistake by the trainee who drafted the policy, and that the contract should be reformed. Representatives of both St. Paul and the insured barge owner stated that the coverage was meant to be "per occurrence," although they had never discussed the issue with each other during negotiations and such coverage was unusual in the industry. The defendant argued that reformation was improper because the insured had never communicated its intent to the insurer. The court, however, found that evidence of past conversations was unnecessary where both negotiating parties were in agreement that their intent was other than that reflected in the writing.

[C]ourts routinely determine the mutual intention of contracting parties without any evidence that their intent was mutually expressed or that there was a relevant custom or usage in the trade. Consequently, where, as here, the insured and the insurer both credibly testify that they intended to have insurance providing coverage different from that in the written policy, the court must reform the policy to conform to their mutual intention[.]

Id. at 978 (citation omitted). Because the agents of the parties agreed as to their intent, the court granted summary judgment and reformed the contract.

Likewise, it is our belief that the parties to this contract have clearly expressed their intent that separate $5,000,000 limits apply to both the stub and one-year periods, and that the facts demonstrate that the policy as written does not reflect that intent. That they are unable to recall a specific discussion of this issue is irrelevant, for the negotiating individuals now agree that their intent at the time of the negotiations was to have separate $5,000,000 limits for each period, and ICI has failed to undermine the witnesses' statements as to their respective recollections.

Alternatively, ICI asserts that the deposition testimony of Radichel contradicts the others' affidavits, and thus creates a genuine issue of material fact. Radichel admitted in his deposition that he was not involved in the negotiation of the policy and no idea what the coverage for the stub period was supposed to be. Nevertheless, ICI contends that his opinion is entitled to special weight because he was responsible for drafting the policy, and thus must have intimate knowledge of its provisions. His opinion, based on his experience as an insurance professional, was that the parties could never have intended to provide a $5,000,000 limit for a four-month policy because: (1) ICI did not normally issue four-month policies, and (2) the premium should have been higher to justify that much exposure for ICI, as it would be irrational to give that much coverage for so little. (The premium, recall, was one-third of the yearly premium charged for $5,000,000 in coverage.) Radichel first adopted his position when ICI paid the $4,500,000 claim, writing letters to Burston indicating that the $5,000,000 limit was for the entire life of the policy, and that separate limits did not apply to any shorter period.

While Radichel's opinion as an ICI employee may or may not coincide with regular industry practice (apart from Radichel's statements — which did not include any specific examples — there was no other evidence as to what the regular practice was, and the trial court disagreed with his opinion), the plain fact is that his testimony is irrelevant. We have noted previously that he was not involved in negotiating the terms of the policy, and, not having been a party to the negotiating discussions, he had no knowledge of what Fleming's intentions were. So far as he knew, the issue of coverage during the stub period was never considered until after the $4,500,000 claim had been settled. Further, Radichel admitted that what we have called his "drafting" of the policy was purely mechanical in nature; that is, it consisted primarily of cutting and pasting provisions from an old policy and preparing and proofing it for typing. This is hardly the type of analytical exercise that would entitle his interpretation to extra weight. More, the mere fact that Radichel did not believe that ICI would normally issue a four-month policy or provide a separate aggregate limit for such a policy does not create a genuine issue of material fact where the mutual intent of the parties is clear and unequivocal. See American Employers, 594 F.2d at 978. The relevant inquiry here is not what ought to have been agreed to, but rather what was agreed to and whether, because of a mutual mistake, that agreement failed to be adequately expressed in the policy Radichel helped piece together. The negotiators of the policy stated that there was a mistake and that the policy as drafted did not accurately reflect their intent; Radichel failed to specifically contradict this testimony, and therefore ICI has failed to establish a genuine issue of material fact. Finally, Radichel's position in his letters to Burston was, as he stated, part of a "hardball" stance taken on behalf of his employer. There is nothing wrong with this, of course, but it also has little bearing on the intent of the parties when entering the contract and deciding what coverage the policy was to provide. Because Radichel has no firsthand knowledge of the parties' intent as to coverage, his opinion falls short of creating a triable issue of fact.

Accordingly, because ICI has not contradicted the U of I's affidavits with sufficient evidence to demonstrate that a genuine issue of material fact remains, the district court's decision to grant summary judgment and reform the contract is

AFFIRMED.