REINHARDT, Circuit Judge:
Plaintiffs Richard J. Plastino, Janice Gudde Plastino and Doris L. Brown (“Plas-tino/Brown”) appeal the district court’s entry of summary judgment in favor of defendants Allan Jamieson and Eureka Savings and Loan Association (“Eureka”) on Plastino/Brown’s claims for declaratory relief and damages. We reverse and remand.
BACKGROUND
Because many of the issues raised on this appeal involve disputes over what factual inferences the evidence in the record supports, we begin by sketching only the undisputed facts. In June, 1983, Plasti-no/Brown agreed to sell realty to a joint venture known as Sunset Bay Associates (“Sunset”) for the purpose of building a residential condominium project. In exchange for the property, Plastino/Brown received approximately $2.4 million in cash and a deed of trust secured by the property in the amount of $1.4 million.
Also in June, 1983, Sunset obtained financing for the project by signing a construction loan agreement with Eureka. Eureka was the principal lender of a group of 14 lenders. At that time, Allan Jamie-son was Eureka’s Chief Lending Officer and supervised Eureka employees who oversaw the construction funding on the project. Sunset’s promissory note for $33 million was also secured by a deed of trust against the property.
There is no dispute as to the order of the recording of the deeds held by Plasti-no/Brown and Eureka. Both deeds were recorded on June 22, 1983: Eureka’s was recorded first, and Plastino/Brown’s second. There is also no dispute that both parties knew at the time of recording that Plastino/Brown were recording after Eure[1507]*1507ka for the purpose of making Eureka’s lien senior to Plastino/Brown’s lien.
The promissory note that Plasti-no/Brown received from Sunset and a rider to its deed of trust each contained a section titled “Construction Loan.” Both the deed rider and the note stated that “loan fees” on the project could not exceed two points, that the loan funds had to be used exclusively for the project, and that Eureka was not obligated to “ensure the proper application or use of Construction Loan funds.”1 The parties dispute whether the record shows that Eureka knew of these terms before the filing of the Plastino/Brown deed.
Between June, 1983 and February, 1986, Eureka disbursed approximately $31 million under the loan. The parties dispute whether all of this money was properly applied to the project or whether, as Plasti-no/Brown contend, Jamieson and Eureka diverted up to $4.7 million to other recipients. The parties also dispute whether the “loan fees” for the project exceeded the two point limit.
For reasons that are not material to the resolution of this appeal, Sunset failed to complete the project on schedule or within budget. It defaulted and filed for bankruptcy before Eureka could foreclose. In August, 1987, the Bankruptcy Court approved a sale of the property to a realty company for $23.7 million.
PROCEEDINGS BELOW
Sunset filed its bankruptcy petition on December 26, 1985. On April 25, 1986, Plastino/Brown filed an amended complaint against Eureka, Jamieson, and several other defendants.2 The complaint against Eureka included five claims for relief. Plastino/Brown sought a declaratory judgment that their lien is superior to Eureka’s; a quiet title judgment to the same effect; money damages for two claims of intentional misrepresentation and conspiracy to commit fraud; and money damages for one claim of negligent misrepresentation. The complaint against Jamie-son sought relief on the intentional misrepresentation claims.
After the parties had completed discovery, the Bankruptcy Court heard cross-motions for summary judgment by Plasti-no/Brown, Eureka, and Jamieson on January 31, 1989. Because this is a non-core bankruptcy proceeding, the bankruptcy judge prepared a report and recommendation in which he made findings of fact and conclusions of law. See 11 U.S.C. § 105 (1988); 28 U.S.C. § 1334 (1988); 28 U.S.C. §§ 157(a), (b)(2)(E) and (C) (1988). He recommended granting the defendants’ summary judgment motions and denying Plasti-no/Brown’s summary judgment motions on all of Plastino/Brown’s claims. The United States District Court for the Central District of California adopted the findings of the Bankruptcy Court. This appeal fol[1508]*1508lowed.3
STANDARD OF REVIEW AND APPLICABLE LAW
We review the grant of a summary judgment motion de novo. Kruso v. International Tel. & Tel. Corp., 872 F.2d 1416, 1421 (9th Cir.1989) cert. denied, — U.S. -, 110 S.Ct. 3217, 110 L.Ed.2d 664 (1990). In order to uphold the grant of summary judgment on each claim we must find that there is no genuine issue of material fact which precludes a finding for the appellant on that claim. See Richards v. Neilsen Freight Lines, 810 F.2d 898, 902 (9th Cir.1987).
California law provides the rules of decision for determining the parties’ substantive rights. See 11 U.S.C. § 510(a) (1988) (providing for enforcement of subordination agreements “to the same extent [as] applicable nonbankruptcy law”). Federal law governs the admissibility of evidence for purposes of summary judgment pursuant to Fed.R.Civ.P. 56(e). See American Title Ins. Co. v. Lacelaw Corp., 861 F.2d 224, 226 (9th Cir.1988) (stating that admissibility of evidence is governed by Federal Rules of Evidence even where state law provides the substantive rule of decision).
DISCUSSION
I. The Relative Priority of the Liens
Plastino/Brown argue that summary judgment was improper with respect to their first two causes of action against Eureka because, viewed in the light most favorable to Plastino/Brown, the evidence supports inferences that: (A) Eureka’s lien was senior to Plastino/Brown’s only if certain conditions in Plastino/Brown’s deed rider and note were met; (B) those conditions were violated; and (C) as a result of the violation, Eureka’s lien is not merely reduced to the extent of the violation, but is rendered entirely subordinate to Plastino/Brown’s. In order for us to reverse we must find that the evidence could support each of these inferences. We address these issues in turn.
A. Whether the Priority of Eureka’s Lien was Contingent Upon Compliance with the Terms of Plasti-no/Brown’s Deed Rider and Note
Under California law, a purchase money deed of trust “has priority over all other liens created against the purchaser, subject to the operation of the recording laws.” Cal.Civ.Code § 2898(a) (West Supp.1991). The recording laws mandate that the first recorded deed of trust on real property is superior to subsequently recorded deeds. Cal.Civ.Code § 1214 (West Supp.1991). Typically, a lender will not provide funds to a developer without assurance that his lien will be senior to any lien given to the seller. Taken together, sections 1214 and 2898(a) provide a simple mechanism to subordinate the seller’s lien to the lender’s: as happened in this case, the seller agrees to record his deed after the lender’s deed is recorded, thus making the lender’s lien superior to the seller’s. The seller willingly subordinates his lien because he knows that absent such subordination, the developer would be unable to obtain financing, and therefore unwilling to purchase the property. We shall refer to this arrangement as a “subordination by subsequent recording.”
In Middlebrook-Anderson Co. v. Southwest Savings and Loan Assoc., 18 Cal.App.3d 1023, 1029, 96 Cal.Rptr. 338, 341 (1971), the California Court of Appeals held that such an arrangement is in substance a subordination agreement, so “that the duties owed by a lender to a seller under a formal subordination agreement do not differ from the duties owed by a lender to a seller when the lender obtains priority over the seller under an agreement by the seller to record after the lender.” In that case, the seller stated a valid cause of action for violation of the provisions of a subordination by subsequent recording where he alleged that the lender had “actual knowledge of the provisions of the seller's lien in [1509]*1509general, and of the subordination therein in particular.” Id. at 1037, 96 Cal.Rptr. at 347. It is therefore clear that if Eureka had actual knowledge of the terms of the Plastino/Brown deed rider and note, the priority of Eureka’s lien would depend upon compliance with those terms.
1. Actual Knowledge
Eureka contends that there is no evidence in the record from which it could be inferred that it had actual knowledge of the terms of the Plastino/Brown deed rider and note. We disagree. First, although Jamieson did not recall discussing the actual terms of the Plastino/Brown deed rider and note, he stated in his deposition that Eureka “required a specific subordination agreement as a condition of financing.”4 The reference to a specific subordination agreement could imply that Eureka required the particular subordination agreement that ended up in the Plastino/Brown deed rider and note. This inference draws support from the fact that although the note Plastino/Brown received from Sunset affected Plastino/Brown’s rights as against Sunset, it also contained a term that protected Eureka’s interests — i.e., the disclaimer absolving Eureka of any obligation to ensure the proper application of loan funds. See supra, n. 1. Because this term benefits Eureka, it is reasonable to infer that it may have been included at the request of someone at Eureka, regardless of whether Jamieson was aware of it.5
jn addition, Plastino/Brown contend that the deposition of Thomas Ginella, president of Baywood Equities — which is one of the joint venturers comprising Sunset— provides further evidence that Eureka knew of the terms of the Plastino/Brown deed rider and note before the deed was recorded. Before assessing Ginella’s testimony, however, we must decide whether it is part of the record. Eureka contends that Ginella’s deposition testimony is inadmissible because it was neither completed nor signed. Eureka argues that it was thereby denied its opportunity for cross-examination, and further, that the unsigned deposition is unreliable.
In Hoover v. Switlik Parachute Co., 663 F.2d 964 (9th Cir.1981), we held that deposition testimony that had not been subject to cross-examination by the party opposing its introduction could be used to support (or oppose) a summary judgment motion if it met the requirements for an affidavit, i.e., if the depositions “were made on personal knowledge and set forth facts that were admissible in evidence.” Id. at 966. See Fed.R.Civ.P. 56(e). Hoover is controlling here except to the extent that it does not address the question whether an uncompleted deposition that is also not signed is admissible as an affidavit. Neither party has cited any case addressing this question, nor have we been able to find any.6 How[1510]*1510ever, the logic of Hoover implies that an unsigned deposition should be admissible where, as here, the deponent was sworn. The approach of Hoover is pragmatic, looking to see if the deposition testimony is at least as reliable as an admissible affidavit. Because there is no reason to believe that the sworn answers to questions are less reliable than an affidavit, to the extent that the content of the deposition testimony is otherwise admissible, that testimony should be admissible on summary judgment. See Fed.R.Civ.P. 56(e). We hold that it is. Accordingly^ we may consider Ginella’s deposition testimony.
When asked when he delivered a copy of the Plastino/Brown promissory note to Eureka, Ginella responded, “[s]ometime prior. Three months prior to the time that we closed the construction loan. A month or two after we closed.” This testimony is hardly a model of clarity. However, our function is not “to weigh the evidence,” Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 249, 106 S.Ct. 2505, 2511, 91 L.Ed.2d 202 (1986), but to view the record evidence in the light most favorable to the parties opposing summary judgment. See Nillson, Robbins, Dalgarn, Berliner, Carson & Wurst v. Louisiana Hydrolec, 854 F.2d 1538, 1542 (9th Cir.1988). Ginella’s response to when the note was delivered could be read as if a comma appeared immediately before the second “we closed.” The purport of his statement would then be as follows: “[The note was delivered] sometime prior — three months prior to the time that we closed the construction loan. A month or two after [the delivery of the note] we closed.” However one reads the statement there is some conflict between the “three months” statement and the “month or two” assertion. This apparent inconsistency can be explained if we assume that Ginella was thinking aloud. He initially states that the note was delivered “sometime” prior to the closing. He then estimates that time as “three months.” After thinking about the events some more he revises his estimate to “two or three months.” Of course there are other ways to read Ginella’s statement, but for purposes of summary judgment, we must conclude that it supports Plastino/Brown’s contention that a copy of the note was delivered to Eureka before Eureka’s deed was recorded.
Finally, Paul R. Pearlson, attorney for Plastino/Brown stated in a declaration that pursuant to formal discovery in this case, he saw a copy of the Plastino/Brown note in a Eureka file. Eureka acknowledged in an answer to an interrogatory that the file contained the copy of the note. Eureka also acknowledged that the same file contained a copy of the deed and rider with the recording stamp already on the deed. Eureka therefore points out that the deed could have been received after it was recorded. On this point, Eureka is certainly correct. However, according to Pearlson, the copy of the note did not indicate when it was received. It could have arrived with the deed, or, as Ginella apparently stated, it could have arrived before the deed was recorded. Because we are required to draw the inference most favorable to Plas-tino/Brown, we view the evidence as supporting the conclusion that Eureka received notice of the terms of Plastino/Brown’s subordination agreement before the deed was recorded, and therefore was obligated not to violate those terms.7 The grant of summary judgment cannot be sustained where, as here, a genuine issue of material fact exists concerning Eureka’s actual prior knowledge of the terms of the subordination agreement.
[1511]*1511Our determination that Plastino/Brown have introduced sufficient evidence of Eureka’s actual knowledge to defeat a summary judgment motion does not, of course, imply that Plastino/Brown will necessarily prevail on this issue at trial. After a trial, the finder of fact might well be justified in concluding the opposite. Accordingly, we rest our decision on an alternate holding as well. We hold that the district court erred in its conclusion that, under Middlebrook-Anderson, a lender must have actual knowledge of the conditions of a seller’s subordination by subsequent recording in order for the priority of its lien to be contingent upon those conditions being met. We conclude, to the contrary, for reasons we explain below, that constructive knowledge is sufficient. Thus, the presence of a genuine issue of material fact concerning Eureka’s actual knowledge of these conditions is a sufficient but not a necessary defense to summary judgment; so long as the lender has advance knowledge of the fact that the seller intends to subordinate by subsequent recording, the priority of the lender’s lien is contingent upon compliance with the terms of the subordination agreement. Accordingly, in order to defeat summary judgment, Plastino/Brown need only raise a genuine issue of material fact as to Eureka’s constructive knowledge of the terms of the subordination agreement — i.e., as to Eureka’s knowledge of the fact of subordination.
2. Constructive Knowledge or Actual Knowledge
We begin by observing that neither Middlebrook-Anderson nor any subsequent California case has addressed the question whether actual knowledge of the provisions of the seller’s lien is required for the lender’s lien to be contingent upon compliance with those conditions, or whether the lender need only know that the seller has subordinated his lien by recording second. We must therefore attempt to predict, based upon Middlebrook-Anderson and the other California cases, as well as general principles of law, what rule California would adopt.8
Eureka’s first argument is textual, calling our attention to the following language in Middlebrook-Anderson: “An implied agreement in the instant case can and, in equity, should be spelled out from [the] lender’s alleged actual knowledge of the provisions of the seller’s lien in general, and of the subordination therein in particular.” 18 Cal.App.3d at 1037, 96 Cal.Rptr. at 347 (emphasis added). Furthermore, a California court subsequently characterized the lender’s duties under Middlebrook-Anderson as arising because the lender “knew of the seller’s security interest and knew that the seller’s lien was subordinate ed only on condition that loan funds were to be used just for construction purposes.” Gluskin v. Atlantic Sav. & Loan Ass’n, 32 Cal.App.3d 307, 314, 108 Cal.Rptr. 318, 322 (1973). However, contrary to Eureka’s contention, this language states only what we already know — that in Middlebrook-Anderson the lender had actual knowledge of the terms of a subordination agreement, and that such knowledge is a sufficient condition for binding it. The California cases do not tell us whether such knowledge is a necessary condition.
Eureka also contends that we should not hold it responsible for knowing what conditions were in Plastino/Brown’s deed because at the time of the subordination Eureka may have reasonably believed that it was under no obligation to discover those conditions. However, whatever rule we apply will result in some unfairness to one of the parties. The question whether constructive knowledge of the terms of a deed subordinated by subsequent recording is sufficient to bind the lender is, after all, an open question under California law. Plastino/Brown no more knew that they would be required to obtain evidence of Eureka’s actual knowledge of the terms of their deed than Eureka knew that they would be presumed to have constructive [1512]*1512knowledge. The fact that our holding has retroactive effects in this case does not provide a basis for preferring Eureka’s claims to Plastino/Brown’s.
We now turn to the question whether the California courts would bind a lender with knowledge only of the fact that the seller has entered into a subordination agreement. There are several reasons why we believe that they would. The first relates to the California policy of treating traditional subordination agreements and subor-dinations by subsequent recording identically. See Middlebrook-Anderson, 18 Cal.App.3d at 1029, 96 Cal.Rptr. at 341. Eureka acknowledges that in a traditional subordination agreement the seller’s lien is subordinated only to the extent that the conditions in the subordination agreement are met — whether or not the lender has actual prior knowledge of those conditions. See 3 H. Miller & M. Starr, Current Law of California Real Estate § 8:166, at 647 (2d ed. 1989) (“Where the new lender has actual or constructive knowledge of the terms of the subordination and the new loan does not satisfy those terms, there is no alteration of priority”) (emphasis added). To effectuate California’s policy of identical treatment for subordination agreements and subordinations by subsequent recording, we should apply the same rule here.
Eureka argues, however, that the two kinds of subordination arrangements should be treated differently. It notes that when a subordination is accomplished by agreement, the reason that actual knowledge of the seller’s terms is not required is that the lender has constructive notice of those terms from the public records. By contrast, in the case of a subordination by subsequent recording there is no public record of the terms of the seller’s trust deed or its restrictions because the seller's deed is not recorded until after the lender’s deed. Therefore, Eureka argues, it is inappropriate to bind the lender unless the lender has actual knowledge of the terms of the seller’s trust deed.
We believe that Eureka’s proffered distinction is not significant. In the traditional subordination situation, California law places a duty on the lender to consult the public records and discover the conditions upon which the seller has subordinated his lien. When a subordination is accomplished by subsequent recording, we believe that California law would place an analogous duty on the lender to consult a different source: the seller himself. Given the degree of cooperation between lender and seller that is necessary to effectuate a subordination by subsequent recording, this is a fairly simple matter. In order to coordinate the subsequent recording, the lender must have actual knowledge of the fact that the seller intends to subordinate. We may use this case as an example. Here, it is undisputed that the two deeds were recorded within minutes of each other. It is also undisputed that the reason for this was Eureka’s requirement that Plastino/Brown’s lien be subordinated to its own. A rule that presumes that under such circumstances the lender will ask to see a copy of the seller’s deed imposes only a very slight burden on the lender — and certainly no greater burden than the current California rule, which presumes that the lender will consult the public records before signing a traditional subordination agreement.
In addition to serving the California policy of identical treatment for the two kinds of subordination arrangements, conditioning the priority of the lender’s lien on compliance with the terms of the seller’s trust deed where the lender has knowledge of the fact of subordination also furthers the California policy “to protect the seller in subordination situations_” Middlebrook-Anderson, 18 Cal.App.3d at 1036, 96 Cal.Rptr. at 346. The basis for this policy is the fact that the lender will generally be a large financial institution with access to expert legal advice, while sellers are usually less sophisticated. The lender is therefore in the better position to know its obligations, and less likely to permit an important right to be sacrificed through inadvertence.
Finally, although, as we have stated, the precise question we address here has not been previously decided by the California courts, one closely analogous case provides [1513]*1513considerable guidance. In Radunich v. Basso, 235 Cal.App.2d 826, 835, 45 Cal.Rptr. 824, 830 (1965), the seller was informed by the purchaser that his lien could only be subordinated to an improvement and construction loan deed. The purchaser then borrowed money from Basso to pay part of the original purchase price. Without the seller’s knowledge, and in violation of their understanding, the purchaser then arranged to have Basso’s lien recorded before the seller’s. The court held that the seller’s lien retained priority even though it was recorded second, and even though Basso may not have had actual knowledge of the terms of the agreement between the seller and the purchaser. The court stated:
Basso knew that the loan he was making was to be used for the down payment on the property and not for improvements. This fact put him on notice and required him to investigate to determine if [the seller] agreed that the deed of trust Basso was buying should be a first lien.
Id., 235 Cal.App.2d at 835, 45 Cal.Rptr. at 830. Hence, binding the lender to the terms of the seller’s lien imposes a type of duty on the lender — a duty to investigate— that the California courts have previously approved.
Therefore, we conclude that under California law, when a seller subordinates his purchase money lien to a lender’s lien by permitting the lender to record first, and the lender knows that the seller has agreed to record his deed second for the purpose of affecting a subordination, the priority of the lender’s lien is contingent upon the fulfillment of any conditions stated in the seller’s lien, whether or not the lender has actual knowledge of those conditions at the time the deeds are recorded. On remand Plastino/Brown need only prove that Eureka knew of the existence of the subordination arrangement in order to show that the priority of Eureka’s lien was contingent on the conditions of Plastino/Brown’s deed rider and note being met.
B. Whether the Conditions of the Subordination Agreement Were Met
The terms of the Plastino/Brown deed rider and note made the subordination contingent on two conditions. First, loan fees could not exceed two points. Second, construction loan funds had to be used exclusively for the project. Plastino/Brown contend that both of these conditions were violated. Because we find that there is a genuine issue of material fact with respect to the loan fees, we need only consider the first condition.9
The two point limit on construction loan fees meant that loan fees could not exceed two percent of $33 million, or $660,000. Plastino/Brown contend that loan fees actually totaled at least $1.38 million. They note that Jamieson stated in his deposition that Eureka’s share of the loan fees amounted to $390,000. They then point to two June, 1983 memoranda from Eureka’s files showing that: (1) $390,000 in “loan fees” were retained by Eureka, $360,000 of which was retained on its own account and $30,000 on another account, and (2) a total of $1,380,000 in “construction loan fees” were listed for the Sunset account. Plasti-no/Brown also point to Exhibit 31, an escrow closing statement for the project listing a “loan fee” of $360,000 for Eureka and a “Loan fee to Various lenders” of $1,020,000. In its answers to interrogatories, Eureka admitted the truth of the information contained in these documents. Thus, Plastino/Brown argue that the total construction loan fees were at least $1,380,-000, and thus clearly in violation of the provisions of the deed of trust. Plasti-no/Brown contend further that brokers’ fees, finders’ fees, loan servicing fees, and permanent loan fees bring the total loan fees to $2,850,000.
Eureka argues that the escrow closing statement of Exhibit 31 was not authenticated in accordance with the rules [1514]*1514of evidence, and that its hearsay contents are therefore inadmissible against Eureka. Under Fed.R.Civ.P. 56(e), material presented on summary judgment must be admissible under the rules of evidence. Eureka is thus correct that Plastino/Brown cannot introduce hearsay to defeat a summary judgment motion. However, Plasti-no/Brown did not rely on the hearsay of Exhibit 31, but rather on the admission by Eureka that Exhibit 31 contained truthful information. An admission is, of course, admissible in evidence. American Title Ins. Co, 861 F.2d at 226. Thus, we may consider the material in Exhibit 31. That document could reasonably be read to show that loan fees totaled as much as $1.38 million, far in excess of the two point limit. Moreover, Eureka does not challenge the two documents that it prepared which support Plastino/Brown’s calculations. Those two documents also support an inference that loan fees totaled $1.38 million. Thus, there is a genuine issue of material fact with respect to whether loan fees in excess of the two point limit were charged, and we need not consider whether the brokers’ fees, finders’ fees, and loan servicing fees should be classified as loan fees.10
C. The Effect of the Failure of the Conditions of the Subordination Agreement
Eureka argues that whether or not the conditions of the subordination agreement were met is wholly irrelevant to the question of the priority of its lien. It first relies on clause (g) of the Construction Loan section of the rider, which provides in part:
Beneficiary acknowledges that the Construction Lender, in making disbursements pursuant to any Construction Loan agreement, will be under no obligation or duty, nor will the Construction Lender represent that it will ensure the proper application of Construction Loan funds. Any application or use of Construction Loan funds, for purposes other than those provided for in the Construction Loan agreement, shall not defeat, alter, or invalidate the lien of the Construction Trust Deed in whole or in part.
(Emphasis added).
It is apparent from the terms of the disclaimer that clause (g) only applies to violations of the requirement that loan funds be used for the project. It does not apply to a violation of the two point loan fee limit. Thus, the provision does not excuse Eureka from the consequences of any failure of the condition relating to excessive loan fees. Rather, it applies only to the issue we do not reach, relating to the diversion of funds.11
Relying on general principles of California contract law, Eureka next argues that even if the conditions of the Plasti-no/Brown deed rider were violated, the appropriate remedy is not to make Plasti-no/Brown’s deed senior to Eureka’s, but to reduce the amount of Eureka’s lien by an amount equal to the interest charged in excess of two percent and/or the value of any funds diverted. In other words, Eure[1515]*1515ka argues that compliance with the conditions of the deed rider is not a condition precedent to the seniority of Eureka’s lien. Eureka points out that even if all of Plasti-no/Brown’s factual allegations are true, when Eureka’s lien is reduced by the appropriate amount, it still exceeds the amount obtained from the bankruptcy trustee’s sale of the property.
Proceeding on the assumption that the deed rider does not state whether fulfillment of its conditions is a necessary condition to Eureka’s lien retaining priority, we must turn to general principles of California law.12 Eureka relies principally on two cases for the proposition that violation of a condition of the subordination voids the subordination only to the extent of the violation. See United States Cold Storage of Calif. v. Great Western Savings & Loan Assoc., 165 Cal.App.3d 1214, 212 Cal.Rptr. 232 (1985); Miller v. Citizens Savings & Loan Assoc., 248 Cal.App.2d 655, 56 Cal.Rptr. 844 (1967). However, those cases involved implied conditions on the subordination. In each case the seller agreed to subordinate his lien to the lender’s lien without conditions, and the court held that to' protect the seller, it was necessary only to reduce the lender’s lien by an amount equal to the value of the improper disbursals. Where, as here, the subordination conditions in the seller’s lien are express, the California courts require substantial compliance with those provisions as a necessary condition to the lender’s lien retaining any priority over the seller’s. See Jones v. Sacramento Savings and Loan Assoc., 248 Cal.App.2d 522, 56 Cal.Rptr. 741 (1967) (holding that in order to achieve any priority, the lender’s encumbrance must substantially comply with the conditions of the seller’s subordination, but, under circumstances not present here, also allowing equitable .lien to lender against developer to avoid developer’s unjust enrichment); Collins v. Home Savings and Loan Assoc., 205 Cal.App.2d 86, 22 Cal.Rptr. 817 (1962).13
[1516]*1516As we have noted, for purposes of summary judgment, we must assume that the loan fees were in excess of twice the two point limit. Eureka does not contend that there was, nevertheless, substantial compliance with the conditions of the subordination. In any event, whether there is substantial compliance with a condition precedent to a contract is a question of fact. Cline v. Yamaga, 97 Cal.App.3d 239, 248, 158 Cal.Rptr. 598, 603 (1979). We certainly cannot conclude at this stage in the proceedings that there was substantial compliance. Thus, for present purposes, we must assume that if the terms of the subordination agreement were violated in the manner alleged, Plastino/Brown’s lien would become senior to Eureka’s. And since we have already concluded that the determination whether a violation occurred depends upon facts that are in dispute, we must reverse the grant of summary judgment with respect to Plastino/Brown’s claim for declaratory relief and its quiet title action.
II. Plastino/Brown’s Conspiracy to Defraud Claims Against Eureka and Jamieson
Plastino/Brown’s conspiracy to defraud claims against Eureka and Jamieson are based on the theory that Plastino/Brown were fraudulently induced into accepting Sunset’s note in partial payment of the purchase price for the property. According to Plastino/Brown, Jamieson and Eureka conspired with the Sunset principals to induce Plastino/Brown to subordinate their lien in exchange for a promise that Jamie-son and Eureka knew the Sunset principals did not intend to keep — the promise that all of the construction loan funds would be used for the project.
“The elements of fraud ... are (1) misrepresentation (false representation, concealment or non-disclosure); (2) knowledge of falsity (or ‘scienter’); (3) intent to defraud, i.e., to induce reliance; (4) justifiable reliance; and (5) resulting damage.” Hackethal v. National Casualty Co., 189 Cal.App.3d 1102, 1111, 234 Cal.Rptr. 853, 857 (1987). We now consider whether, viewed in the light most favorable to Plasti-no/Brown, the evidence could support their claims against Eureka and Jamieson.
A. Conspiracy to Defraud Claims Against Eureka
Plastino/Brown argue that Eureka is liable for the fraudulent misrepresentations of its alleged co-conspirators.14 Specifically, Plastino/Brown contend that Eureka conspired with the Sunset joint venturers to induce Plastino/Brown to subordinate their deed of trust without disclosing that Eureka did not intend to comply with the conditions in the deed of trust.15 Under this theory, Eureka would be liable for the misrepresentations of Sunset, even though Eureka itself made no fraudulent misrepresentations or non-disclosures. See Guild Mortgage Co. v. Heller, 193 Cal.App.3d 1505, 1514, 239 Cal.Rptr. 59, 65 (1987) (one member of civil conspiracy is liable for the fraudulent misrepresentations of another).
[1517]*1517Although an express agreement need not be shown for a plaintiff to prevail on a civil conspiracy claim, there must be at least a tacit understanding. See Cully v. Bianca, 186 Cal.App.3d 1172, 1176, 231 Cal.Rptr. 279, 281 (1986). As discussed above, see supra at 1509-11, there is evidence from which it can be inferred that Eureka knew of the terms of the subordination agreement before Plastino/Brown’s deed was recorded. Eureka could have granted prior approval to the Sunset principals’ supposed plan to offer the deed rider terms as a fraudulent inducement. Under California law, the existence of a conspiracy “may sometimes be inferred from the nature of the acts done, the relations of the parties, the interests of the alleged conspirators, and other circumstances.” Greenwood v. Mooradian, 137 Cal.App.2d 532, 538, 290 P.2d 955, 959 (1955). Plasti-no/Brown’s sale of their land was a prerequisite to the existence of the Sunset project, and therefore a prerequisite to Eureka’s opportunity to divert funds. Thus, under the Greenwood test, it might be inferred from Eureka’s interests and its close relationship with Baywood, that Eureka was privy to the alleged deception of Plastino/Brown from the very beginning. If the record supports the inference that Eureka knowingly diverted project funds, then a fact-finder could conclude that all of the circumstances support an inference of a conspiracy. Hence we now consider each of the diversions Plastino/Brown has alleged.
Plastino/Brown contend that Eureka diverted project funds in five ways: (1) that with Jamieson’s approval Eureka issued a $200,000 letter of credit ultimately traceable to the Sunset construction loan account to settle a lawsuit by Dr. Peter Delorio against Baywood Equities, one of the joint venturers in Sunset, on a claim unrelated to the Sunset project; (2) that Baywood Equities, acting pursuant to Ja-mieson’s instructions, paid $100,000 to the Perini corporation out of Sunset project funds in order to pay off an unrelated loan from Perini to Baywood Equities; (3) that Jamieson signed a $200,000 check on behalf of Eureka for “return of capital” to Riley-Bower, one of the Sunset joint-venturers; (4) that Eureka, with Jamieson knowingly participating, diverted $60,000 in project funds to Angelo Quaranta, and that these funds ultimately went to the Chairman of the Board of Directors of Eureka, Kenneth Kidwell; and (5) that Eureka diverted nearly $2 million to Baywood Equities as fraudulent “construction management fees.” We briefly consider the evidence in support of each of these alleged diversions.
1. The $200,000 letter of credit to settle Delorio’s lawsuit against Baywood.
The parties do not dispute that a $200,-000 letter of credit was issued to Baywood. Eureka contends that the funds were not drawn on Baywood’s construction loan account,16 and thus do not constitute a diversion. Eureka’s contention is supported by the declaration of Thomas Ginella. Plasti-no/Brown rely primarily on the uncompleted deposition of Baywood President James Zetz and the declaration of Dennis Alfaro, the former controller of Baywood. Alfaro states in his declaration that Eureka loan officer Elliot Nosaka told him that Jamie-son had instructed Nosaka to cover up a diversion of funds from the construction loan account. Although contradicted by other evidence, if admissible this statement would raise a genuine issue of material fact concerning the diversion.
Eureka objects that Nosaka’s statement to Alfaro is hearsay. However, since at the time Nosaka made the statement to Alfaro, Nosaka was Eureka’s “agent or servant [discussing] a matter within the scope of the agency or employment,” it is not hearsay. Fed.R.Evid. 801(d)(2)(D). Eureka also objects to the use of the Zetz deposition because it was not completed. However, as with the Gi-nella deposition discussed above, see supra at 1509-10, the deposition is admissible.
[1518]*15182. The $100,000 to the Perini corporation.
The principal allegation of Eureka’s knowledge that the payment of project funds to Perini was for non-project activities consists of the uncompleted Zetz deposition. Eureka objects that Zetz’s recital of statements made by Baywood representatives are inadmissible hearsay and restates its objection to the use of an uncompleted deposition. As noted above, the statements are admissible. Although a fact-finder would certainly be entitled to find that the substantial documentary and testimonial evidence offered by Eureka implies that neither it nor its agents had knowledge of the fraudulent payment,17 there is a genuine issue of material fact.
3. The $200,000 “return of capital” to Riley-Bower.
The evidence of a wrongful “return of capital” is a $200,000 check to Riley-Bower with that phrase written in the memo portion. The check appears to bear Allan Jamieson’s signature, although the parties dispute whether this is the case. In any event, the check was not authenticated, and therefore is inadmissible. Thus, this claimed diversion finds no support in the record.
4. The $60,000for Kenneth Kidwell via Angelo Quaranta.
Plastino/Brown’s allegation that project funds were diverted to Kenneth Kidwell is supported by the Zetz deposition, in which Zetz contends that he was told of the diversion by Ginella. Plastino/Brown argue that this is not hearsay because it is the statement of a co-conspirator; Ginella represented alleged co-conspirator Bay-wood Equities. The statement does not fall under the rule, however, since it was not made “in furtherance of the conspiracy,” but after the fact. As with all of Plastino/Brown’s charges, the alleged diversion to Kidwell is disputed by Jamie-son’s deposition. It is also called into question by Nosaka’s deposition testimony that it was not unusual for him to deliver checks to Kidwell when instructed to do so by his superiors. Thus, excluding inadmissible evidence, there is no genuine issue here.
5.The approximately $2 million to Baywood Equities as fraudulent “construction management fees.”
These allegations also depend upon the Zetz and Ginella depositions. For the reasons given above, that evidence is admissible, and though contested by other admissible evidence supporting Eureka’s view of the transactions, leaves a genuine issue of material fact.
Viewing the admissible evidence in the light most favorable to Plastino/Brown, we find that the record could support an inference of three diversions: the letter of credit; the payment to Perini; and the construction management fees.18 These diversions in turn could support an inference that Eureka was engaged in a conspiracy to defraud Plastino/Brown. Consequently, the district court should not have granted summary judgment on these claims.
B. Plastino/Brown’s Conspiracy to Defraud Claim Against Jamieson
The evidence offered by Plastino/Brown to show that Jamieson took part in the diversion of funds is essentially the same evidence offered against Eureka. Most of Plastino/Brown’s allegations regarding diversions rely on the Zetz deposition testimony or the Alfaro declaration containing [1519]*1519the statements made by Nosaka to Zetz and Alfaro. Jamieson argues that none of this evidence is admissible against him.
Preliminarily, the fact that the Zetz deposition was not completed does not render it inadmissible. See supra at 1509-10. Plastino/Brown offer two theories for admitting Zetz’ and Alfaro’s statement about what Nosaka told them. First, Plastino/Brown argue that the statements are admissible under the co-conspirator rule. Fed.R.Evid. 801(d)(2)(E). Under Bourjaily v. United States, 483 U.S. 171, 107 S.Ct. 2775, 97 L.Ed.2d 144 (1987), it is permissible for the court to look to the proffered hearsay itself to determine if there is a conspiracy. However, Rule 801(d)(2)(E), by its terms, requires that the statement be made “in furtherance of the conspiracy.” Since all of the statements made by Nosaka to Zetz and Alfaro were revelations of past conduct, which if anything had the effect of frustrating the conspiracy, they do not qualify under the rule.
Plastino/Brown’s second theory is that Nosaka’s statements are admissible against Jamieson because Nosaka, as an officer of Eureka working under Jamie-son’s supervision, was Jamieson’s “agent or servant [discussing matters] within the scope of the agency or employment....” Fed.R.Evid. 801(d)(2)(D). Nosaka was not employed by Jamieson, but by Eureka. Thus, while it is unquestionably true that Nosaka’s statements are admissible against Eureka, whether they are admissible against Jamieson is less clear. The Eighth Circuit has held that the mere fact that one employee of an enterprise (in that case the government) is subject to the supervision of someone else in the hierarchy, does not render the employee his supervisor’s agent. See Ellis v. Kneifl, 834 F.2d 128, 131 (8th Cir.1987) (holding that clerk is not judge’s agent within the meaning of the rule). However, our prior cases suggest that Nosaka’s statements are admissible against Jamieson. Under circumstances where an employee’s superior has instructed the employee to act in furtherance of a fraudulent scheme, statements made by the employee relating to that scheme are admissible against the superior under Fed.R.Evid. 801(d)(2)(D). See United States v. Gibson, 690 F.2d 697, 701 (9th Cir.1982), cert. denied, 460 U.S. 1046, 103 S.Ct. 1446, 75 L.Ed.2d 801 (1983) (alternative holding). Significantly, the statements need only concern matters within the scope of the agency; they need not be made within the scope of the agency. Hoptowit v. Ray, 682 F.2d 1237, 1262 (9th Cir.1982).
Since Nosaka was allegedly instructed by his superior Jamieson to cover up a diversion of funds — clearly an instruction to act in furtherance of a fraudulent scheme — his statements to Zetz and Alfaro about the instruction are admissible against Jamieson. Thus, all of the evidence admissible against Eureka to show a diversion is also admissible against Jamie-son. In addition to that evidence, Plastino/Brown have also pointed to portions of Zetz’ deposition which show that Jamieson instructed Perini to draft a phony bill to be paid out of project funds. Taken together, all of this evidence could support an inference that Jamieson knowingly participated in several diversions of funds.
The fact that Jamieson participated in diversions is not by itself enough to establish that he conspired with the Bay-wood principals to defraud Plastino/Brown. While there is evidence from which it may be inferred that Eureka as an institution knew the terms of the Plastino/Brown subordination before their deed was filed, there is no evidence that Jamieson himself had such knowledge. Nevertheless, the California courts have stated that one may be held liable for a civil conspiracy where without prior knowledge he ratifies the acts done for his benefit by a wrongdoer. See Cully v. Bianca, 186 Cal.App.3d 1172, 1176, 231 Cal.Rptr. 279, 281 (citing Sindell v. Abbott Laboratories, 26 Cal.3d 588, 604, 163 Cal.Rptr. 132, 140, 607 P.2d 924, 932 (1980)). It is undisputed that Jamieson learned of the existence of the Plastino/Brown deed rider before the alleged diversions took place. As we have noted, the deed rider placed a duty on Eureka and its agents — including Jamieson — not to participate knowingly in the violation of any of [1520]*1520its terms. See supra, n. 12. From the allegation that Jamieson violated that duty it could be inferred that he ratified the initial fraud of the Baywood principals.19 Thus, summary judgment was improper on Plastino/Brown’s claim that Jamieson is jointly liable as a conspirator.
III. Plastino/Brown’s Motion for Summary Judgment Against Eureka
Plastino/Brown contend that they are not only entitled to a reversal of the district court’s grant of summary judgment for Eureka and Jamieson, but that summary judgment should be entered for Plasti-no/Brown. Because there are material issues of fact that must be resolved to determine which party should prevail on each of Plastino/Brown’s claims, they are not entitled to summary judgment.
IV. Sanctions and Attorneys’ Fees
Plastino/Brown request attorneys’ fees as a sanction for what they term the “vexatious conduct” of Eureka in filing an affidavit that Eureka allegedly knew to be false, and in refusing to withdraw the affidavit once its supposed flaws were exposed. Plastino/Brown refer to the affidavit of Deborah Imamura, which they contend was inconsistent with the answer to an interrogatory that Eureka had previously submitted. However, we find that Ms. Imamura’s affidavit was consistent with that answer. Thus, it was not improper for Eureka to submit (and not to withdraw) the affidavit.
Not to be outdone, Eureka seeks double costs and attorneys’ fees against Plastino/Brown under Fed.R.App.P. 38 and 28 U.S.C. § 1912 for Plastino/Brown’s “frivolous appeal.” Since Eureka is not a “prevailing party” it is not entitled to costs under 28 U.S.C. § 1912. Needless to say, Plastino/Brown’s meritorious appeal is hardly frivolous. In fact, it is Eureka’s request that comes perilously close to warranting sanctions.
CONCLUSION
Because there are genuine issues of material fact concerning each of the claims presented on this appeal, we reverse the district court’s entry of summary judgment in both No. 89-55705 and No. 89-55707.
REVERSED AND REMANDED.