OPINION OF THE COURT

MAZZARELLI, J.P.

Plaintiff Jacobson Family Investments (JFI) is an investment management company that manages the assets and businesses of the 16 other plaintiffs, which are various limited liability companies, limited partnerships, foundations and trusts established by various members of the Jacobson family and another family. JFI manages the assets of these entities by selecting outside investment advisors. In 1998, JFI made the fateful decision to select Bernard L. Madoff and his firm, Bernard L. Madoff Investment Securities LLC (BLMIS), as one of those outside investment advisors.

JFI attempted to protect its investments by purchasing fidelity bonds insulating it from theft or other dishonest acts of the outside investment advisors. The primary bond at issue in this case was sold to JFI by defendant National Union and covered the policy period from October 19, 2007 to November 1, 2008, and was later extended to March 2009. JFI also purchased several layers of excess fidelity bonds (the excess bonds) from National Union and the remaining defendants.1 The bond had a $10 million single loss limit of liability, an aggregate limit of liability of $20 million, and a deductible of $3 million per single loss. The operative portion of the bond provided that coverage would be afforded for "[l]oss resulting directly from dishonest or fraudulent acts committed by an Employee2 acting alone or in collusion with others." The term "loss" is not defined anywhere in the bond.

As part of its application to National Union for the bond, JFI disclosed that the amount of assets being managed by BLMIS on behalf of the various entities was at the time $123,805,948. Unbeknownst to JFI or National Union, a substantial portion of that figure represented fictitious gains in JFI's initial investment with BLMIS. Rider 9 to the bond constituted JFI's representation that the information disclosed in the application was complete, true and correct and provided that the application "constitutes part of this policy."

After Madoff's fraud was exposed in December 2008, following his arrest, it was revealed that six of the plaintiffs had contributed more money to BLMIS than they had withdrawn from it, or, in the common parlance, were "net losers." The remaining plaintiffs were "net winners," because over time they had withdrawn more money than they had invested. When the "net wins" and "net losses" of the individual plaintiffs are aggregated, the entities are shown to have had a total net win of $3,142,677. Nevertheless, JFI submitted a single proof of loss to National Union in the amount of $107,619,369.33,3 which was based on the last account statement furnished by BLMIS prior to Madoff's arrest. That statement, of course, and the proof of loss which was based on it, included the millions of dollars in "gains" which Madoff infamously conjured out of thin air.

National Union denied coverage, asserting, inter alia, that JFI suffered no losses from Madoff's wrongdoing because "the non-existent profits that Mr. Madoff fraudulently attributed to his purported investments" did not constitute a loss under the bond. The excess insurers denied JFI's claim on the same grounds. JFI commenced this action, seeking a declaratory judgment that the entire claimed loss was covered by the bond, as well as damages for breach of contract and breach of the implied covenant of good faith and fair dealing.

Before discovery commenced, JFI moved for partial summary judgment declaring, inter alia, that the bond covered the full extent of the losses it claimed. Defendants jointly cross-moved for summary judgment declaring that coverage under the bond was limited to JFI's "actual losses" in its Madoff accounts, and also moved to dismiss JFI's claim for breach of the covenant of good faith and fair dealing. In an argument which they withdrew before the court decided the motions, defendants further asserted that, if their "actual losses" theory was adopted, the gains and losses of the various JFI entities should be aggregated. Because this would result in a total net win for JFI, defendants submitted that the entire complaint should be dismissed.

Supreme Court denied JFI's motion, and granted the cross motion, to the extent of finding that the bond limited coverage to JFI's "actual losses" (2011 NY Slip Op 33628[U], *8). The court concluded that because the fictitious gains recorded by BLMIS were "never owned" by JFI, they could not have been "lost" (id.). The court also dismissed JFI's claim for breach of the covenant of good faith and fair dealing, since defendants had an "arguable basis" for denying coverage (id. at *10).

After the parties had begun discovery, JFI moved for leave to renew its motion on the basis that it had gained access to information in the possession of defendants and third parties which warranted a different conclusion than the court had reached. The first new development presented by JFI had to do with the fact that a fidelity bond that National Union had issued to JFI in 2003 was expressly limited to losses of JFI's "investment interest" because of the dishonest acts of outside investment advisors, that is, the actual amount of cash which JFI entrusted to them. JFI claimed that it had discovered that National Union representatives had specifically insisted on the inclusion of that limitation in the 2003 bond, and argued that the omission of this limitation in the bond at issue was thus a purposeful act which established that National Union expected to cover more than simply lost "investment interest." The second new development which formed the basis of the renewal motion was JFI's having learned that, in calculating the bond's premium, National Union multiplied the premium by 375% to cover the total "assets at risk." According to JFI, this established that National Union understood that it was insuring the entire value of assets being managed by BLMIS at the time the bond was purchased, which, albeit unbeknownst to the parties, included tens of millions of dollars of fictitious profits.

Defendants simultaneously moved for summary judgment dismissing the complaint, reviving the claim made in the original motion that, on a net, aggregated basis, plaintiffs suffered no losses, as they collectively withdrew more money than they invested with Madoff. In making the argument that plaintiffs had to be viewed as having submitted a single claim, defendants noted that JFI submitted just one proof of claim. They also cited three provisions in the bond. The first was rider 8, which listed all of the individual entities covered by the bond and stated that they constituted the "Complete Named Insured." The second provision on which defendants relied was the definition of the term "Single Loss" in the section of the bond related to limits of liability. Because "Single Loss" was defined by the policy as "all covered loss ... resulting from [various acts of malfeasance]," defendants argued that all of the various losses suffered by the individual plaintiff entities constituted one aggregate loss. Finally, defendants relied on the Bond's "Joint Insured" provision, which stated:

As further evidence that plaintiffs collectively did not suffer a loss, defendants submitted JFI's settlement agreement with the trustee for the Securities Investor Protection Corporation (SIPC), which was charged with marshalling the assets of the bankrupt BLMIS and compensating Madoff's victims. While the net loser plaintiffs were entitled to have their claims paid by SIPC, to the extent funds were available, the net winner plaintiffs were subject to "clawback" claims from SIPC. The net losers agreed with SIPC to forego approximately $25 million worth of potential recovery in return for SIPC's release of its right to bring clawback claims against the net winners. Defendants argued that this constituted a "recovery" to JFI, and triggered the bond provision that

Defendants also argued that any payment to the net losers should be offset by a $2.5 million dollar cash payment to them from SIPC, as well as a $2.2 million payment to them from two of the net winners. Finally, defendants asserted that, if the court were to reject their argument and find that the net losers' claims were separate from each other, it should declare that the $3 million policy deductible applied to each of those claims.

The court granted JFI's motion for leave to renew based on the newly discovered evidence, and adhered to its original determination that the bond covered only the investment interest of each plaintiff in its individual BLMIS account (2012 NY Slip Op 33013[U]). With respect to defendants' motion, the court found that the plain language of the cited bond provisions did not compel aggregation of plaintiffs' net wins and losses. The court specifically found that a "single loss" was defined as "all covered losses," not all "net losses" (id. at *13). It found that the Joint Insured provision merely created an organized procedure for the 160 separate insureds to make claims under the bond. The court rejected defendants' argument that filing a single proof of loss suggested that JFI intended for the individual investing entities to be treated as a single insured, finding that by filing a single proof of claim JFI behaved consistently with the Joint Insured provision.

The court also rejected defendants' argument that the settlement with SIPC supported their claim that plaintiffs collectively suffered no loss, on the basis that the plain language of the bond precluded the consideration of extrinsic evidence. However, even reviewing the settlement agreement as potential evidence of a recovery by plaintiffs, the court held that it was unreasonable to consider the intangible net value JFI received as a result of the settlement agreement as a "recovery." The court also held that the $2.2 million payment to two of the net losers from two of the net winners did not constitute a "recovery" within the meaning of the bond. Because the issue of which plaintiffs suffered losses was no longer in dispute, the court dismissed the complaint as to the net winners, as they did not suffer actual losses. However, the court rejected defendants' argument that the $3 million deductible applied to eight of the net losers, finding that Madoff's fraud was one single act of malfeasance, and all claims of the net losers would therefore be subject to one deductible. Thus, the court determined that the net losers were entitled to actual losses, minus any cash payments by the SIPC Trustee and the $3 million deductible, the exact amount to be determined at trial.

Because this dispute is so dependent on the interpretation of the language in the bond, it is worthwhile to review certain construction precepts. No different from interpreting the terms of any contract, the goal of a court reviewing an insurance policy is to ascertain "whether, afford[ing] a fair meaning to all of the language employed by the parties in the contract and leav[ing] no provision without force and effect ... there is a reasonable basis for a difference of opinion as to the meaning of the policy" (Federal Ins. Co. v International Bus. Machs. Corp., 18 N.Y.3d 642, 646 [2012] [internal quotation marks omitted]). If so, the policy is ambiguous, and a court may consider extrinsic evidence in attempting to resolve the ambiguity (see State of New York v Home Indem. Co., 66 N.Y.2d 669, 671 [1985]). However, if a policy "has a definite and precise meaning, unattended by danger of misconception in the purport of the [agreement] itself, and concerning which there is no reasonable basis for a difference of opinion ... a court is not free to alter the contract to reflect its personal notions of fairness and equity" (White v Continental Cas. Co., 9 N.Y.3d 264, 267 [2007] [internal quotation marks omitted]).

JFI has the burden of proof as to whether the entire loss it claims is covered by the bond (Consolidated Edison Co. of N.Y. v Allstate Ins. Co., 98 N.Y.2d 208, 218 [2002]). In arguing that the undefined term "loss" unambiguously encompasses the fictitious profits, it principally relies on two provisions in the bond. The first is rider 9, which provides that JFI's application for the bond, which included the most recent statement from BLMIS, "constitutes part of this policy." The second is the section of the bond entitled "Ownership," which provides that "[t]his bond shall apply to loss of Property (1) owned by the Insured, (2) held by the Insured in any capacity, or (3) for which the Insured is legally liable." Conversely, JFI claims that the court, in finding that the term "loss" unambiguously embraces only the loss of "real" assets, improperly "supplied" terms to the policy that do not actually appear in the text of the bond, such as "actual" and "direct."

The sections of the bond which JFI affirmatively relies on are insufficient to carry its burden. Rider 9 is an especially slim reed to hang on. It is true that it technically makes the most recent BLMIS statement part of the bond. Nevertheless, this offers no guidance on how to define the term "loss." Moreover, it hardly serves as an estoppel against defendants taking the position that only "real" losses are covered. JFI's attempt to analogize to Haber v St. Paul Guardian Ins. Co. (137 F.3d 691 [2d Cir 1998]) fails. There, the court observed that it would be reasonable to infer, where a homeowner applies for a workers' compensation policy and discloses the existence of a live-in housekeeper, that the homeowner intended for the policy to cover that employee. However, in contrast to the fictitious Madoff gains at issue here, there was no question that the employee existed. The Haber court may have reached a different conclusion had the homeowner only "thought" that the employee existed.

The "Ownership" section of the bond also fails to advance JFI's position that the term "loss" covers the phantom gains. JFI claims that, if it could not have legally "owned" the fictitious profits, it was at least "legally liable" for them to the extent that it paid taxes on them. Also, it asserts, it held the profits in the "capacity" of having a Uniform Commercial Code (UCC) "security entitlement" in them. JFI is correct that, before the Madoff scheme was exposed, the taxing authorities, like most everyone else in the world, assumed the investments were legitimate and would have had a right to collect taxes on the investment "gains." However, any such authority evaporated at the same time JFI learned it had no such "gains" to "lose." Indeed, JFI has attempted to take advantage of Internal Revenue Service procedures designed to relieve taxpayers who calculated their returns in part on "phantom" income. Similarly, any protectable UCC "interest" based on the fictitious value of securities only existed for as long as the Madoff scheme remained hidden.

JFI further argues that, even if its own effort to define the term "loss" is not directly supported by the language employed in the bond, the motion court overreached in attempting to fashion its own definition. The court's interpretation was primarily based on Horowitz v American Intl. Group, Inc. (2010 WL 3825737, 2010 US Dist LEXIS 103489 [SD NY, Sept. 30, 2010, No. 09-Civ-7312 (PAC)], affd 2012 WL 3332375, 2012 US App LEXIS 17055 [2d Cir, Aug. 15, 2012, No. 10-4408-cv]). In Horowitz, the plaintiffs purchased a homeowner's policy from the defendant with coverage for "the loss of money [or] securities... resulting directly from fraud, embezzlement, or forgery" (2010 WL 3825737, *1, 2010 US Dist LEXIS 103489, *2). They invested money with BLMIS and were net winners. However, the plaintiffs filed a claim for their vanished investment "profits." The term "loss" was, like here, undefined in the policy. The court agreed with the defendant that the policy was not ambiguous on its face. It further rejected the plaintiffs' interpretation of the term "loss" as unreasonable, finding that "it is not reasonable to contend that one can lose money that never existed in the first place" (2010 WL 3825737, *6, 2010 US Dist LEXIS 103489, *19). The Horowitz court favorably cited the decision in the Madoff bankruptcy litigation in excluding the fictitious gains in creditors' claims, which was based in part on the observation that "`[i]t would be simply absurd to credit the fraud and legitimize the phantom world created by Madoff'" (2010 WL 3825737, *7, 2010 US Dist LEXIS 103489, *24, quoting In re Bernard L. Madoff Inv. Sec. LLC, 424 B.R. 122, 140 [SD NY 2010]).

We adopt the logic of the Horowitz court and hold that no reasonable interpretation of the term "loss" in the context of the bond allows for coverage of fictitious Madoff gains. We further note that Horowitz is not alone in finding that "bookkeeping or theoretical loss[es], not accompanied by actual withdrawals of cash or other such pecuniary loss," are not covered by fidelity bonds (Cincinnati Ins. Co. v Star Fin. Bank, 35 F.3d 1186, 1191 [7th Cir 1994] [internal quotation marks omitted] [endorsing position of insurance carrier for bank that bank did not stand to suffer a "loss" were it forced to reimburse payor bank for funds that insured bank received by accepting a forged check]; see also In re New Times Sec. Servs., Inc., 371 F.3d 68, 88 [2d Cir 2004] [holding that Securities Investment Protection Act (SIPA) did not have to compensate Ponzi scheme victims beyond their cash investments because "basing customer recoveries on fictitious amounts in the firm's books and records would allow customers to recover arbitrary amounts that necessarily have no relation to reality"] [internal quotation marks omitted]).

JFI attempts to distinguish Horowitz by noting that the court there only considered whether the plaintiffs had parted with "something of value" within the meaning of that policy. However, this is too narrow a reading of Horowitz. The Horowitz court clearly meant to convey that no insurance policy can be interpreted to compensate an insured for something that, unbeknownst to the parties, only appeared to exist because of someone else's fraud. JFI criticizes the Horowitz court's reliance on In re Bernard L. Madoff Inv. Sec., arguing that the Bankruptcy Court was concerned with the application of SIPA, not state insurance law. However, the distinction is meaningless. Under either scenario, it is not reasonable to claim that the revelation that an asset, once thought to exist, did not exist, constitutes a "loss," whether for the purpose of a claim under SIPA or under a fidelity bond. Further, in concluding that the bond covered only "actual" or "direct" losses, the motion court did not, as defendants argue, improperly "supply" terms to the policy that do not actually appear in the text of the bond. Nothing about the court's analysis deviated from standard interpretation of contract terms, which, after all, was the court's central role in resolving this dispute.

The recent Court of Appeals decision in Simkin v Blank (19 N.Y.3d 46 [2012]) does not compel a different result than the one reached here. There, a divorcing couple entered into a settlement agreement in which the wife received a distribution that was based on a valuation of the marital estate that included Madoff funds, which the husband retained. The husband sought to reform the agreement after it was revealed that a significant portion of the Madoff fund valuation was based on fictitious profits. The Court, in rejecting the husband's position, stated that "[t]his situation, however sympathetic, is more akin to a marital asset that unexpectedly loses value after dissolution of a marriage; the asset had value at the time of the settlement but the purported value did not remain consistent" (19 NY3d at 55). The Court's statement that the Madoff assets "had value" has no impact on this case. That observation was directly related to the court's holding that the parties did not commit a mutual mistake when they entered into the settlement agreement. It held that they did not because at the time the Madoff assets could have been redeemed in full. Here, there is no claim of mutual mistake. Rather, the analysis is strictly limited to what extent JFI's assets were protected by the bond.

Finally, JFI argues that extrinsic evidence proves that "loss" under the bond is not limited to investment interest. This evidence consists of defendants' alleged insistence on the "investment interest" limitation in the 2003 bond and the multiplication of the premium by 375% to cover the total "assets at risk." However, because we find that the term "loss" is not ambiguous, there is no reason to look outside the policy to interpret the term (see State of New York v Home Indem. Co., 66 NY2d at 671). In any event, we disagree that the evidence is sufficient to create an issue of fact whether the fictitious profits are covered by the bond. Merely because the 2003 bond stated in explicit terms that only cash outlays were covered does not, without more, lead to the conclusion that another bond issued four years later without such express language was not so limited. As for the amount of premium paid, JFI fails to present sufficient evidence for a factfinder to infer that the increase applied to the bond was predominantly related to the amount of assets being protected. To the contrary, the record reveals that, in the five years leading up to the issuance of the bond, the total amount of assets managed by outside investment advisors tripled, while the premium actually decreased over time. Further, the deposition testimony presented by JFI to support the premium theory is highly equivocal as to whether there was a direct correlation between the amount being insured and the premium, and in fact suggested that the amount was only one of several risk factors considered in calculating the premium.

Having agreed with defendants that JFI may only recover the actual cash investment it lost to Madoff's fraud, we must also conclude that plaintiffs' claim for breach of the implied covenant of good faith and fair dealing was properly dismissed, as plaintiffs did not establish that defendants had "no arguable basis" for denying coverage (Wurm v Commercial Ins. Co. of Newark, N.J., 308 A.D.2d 324, 329 [1st Dept 2003], lv denied 3 N.Y.3d 602 [2004]). We turn instead to defendants' position that the gains and losses of each individual investing entity must be seen as one whole. Defendants maintain that the bond unambiguously provides that the individual entities were to be considered together for purposes of making a claim. Again, however, we must apply the plain meaning of the relevant terminology (see White v Continental Cas. Co., 9 NY3d at 267). Defendants assert that because rider 8 to the bond lists the covered investment vehicles under the umbrella term "Complete Named Insured," any claim made on behalf of more than one of those entities must be considered as a single, aggregated claim. We disagree. Rider 8 is a definitional, as opposed to an operative, section of the bond. As such, it is impossible to conclude that it has anything to do with whether claims by the individual insureds are to be considered separately or aggregated together. Similarly, the "Joint Insured" provision is a housekeeping measure which has no bearing on how individual claims may be accounted for. The provision which defines "single loss" as "all covered loss" arising out of related acts of wrongdoing cannot be reasonably read as requiring a setoff for insured entities that did not suffer a loss at all. At the same time, however, consistency dictates that, if each net loser is seeking coverage for its own loss, each individual claim is subject to the $3 million single loss deductible.

Finally, we reject defendants' argument that the bond requires the SIPC settlement to offset the total loss accumulated by the net loser plaintiffs. The bond unambiguously provides that any "recovery" from a third party shall be applied "to the satisfaction of the Insured's loss." Since we have already rejected defendants' argument that the individual plaintiff investment entities are not one aggregate "Insured," we must interpret this clause as requiring a "recovery" to the insured entity itself. Defendants' position, however, is that, by releasing their claims against SIPC, the "net loser" plaintiffs effected a benefit for the "net winners" in the form of a release of potential clawback claims. Accordingly, the recovery provision is not implicated by the settlement. However, on this record, it cannot be determined, as a matter of law, whether the alleged $2.2 million paid by two plaintiffs who were net winners to two plaintiffs who were net losers constituted a "recovery" under the bond. It is unclear for what purpose these payments were made and whether they were intended to compensate the two net losers for their loss such that payment by defendants would constitute a double recovery.

Accordingly, the order of the Supreme Court, New York County (Richard B. Lowe, III, J.), entered July 13, 2011, which, to the extent appealed from, denied plaintiffs insureds' motion for partial summary judgment, and granted defendants insurers' cross motion for summary judgment to the extent of limiting any recovery by plaintiffs under the fidelity bonds at issue to the loss of their investment interest and dismissing plaintiffs' claim for breach of the implied covenant of good faith and fair dealing, should be affirmed, with costs. The order of the same court and Justice, entered February 29, 2012, which, to the extent appealed from, granted plaintiffs' motion for leave to renew and adhered to the original determination, and denied so much of defendants' motion for summary judgment as sought dismissal of the "net loser" plaintiffs' claims, should be modified, on the law, to apply the $3 million single loss deductible to each net loser's recovery, if any, and otherwise affirmed, with costs.

Order, Supreme Court, New York County, entered July 13, 2011, affirmed, with costs. Order, same court and Justice, entered February 29, 2012, modified, on the law, to apply the $3 million single loss deductible to each net loser's recovery, if any, and otherwise affirmed, with costs.