PREGERSON, Circuit Judge:

Petitioners are Oregon public employees who are members of the state's Public Employee Retirement System ("PERS"). They appeal the district court's summary judgment order that held that the 1978 Title VII consent decree entered was not intended to permanently lock in the 1978 refund annuity rates for all PERS members. After reviewing the language of the consent decree and the evidence of the parties' intent in the context of the litigation as a whole, we conclude that the consent decree was not meant to create a perpetual floor on refund annuity benefits. Accordingly, we affirm the district court.

BACKGROUND

I. PERS and Refund Annuities

PERS is a statutory retirement plan for state and local government employees in Oregon. Many retired PERS members receive a monthly "service retirement allowance" composed of two parts: (1) a "refund annuity" and (2) a life pension. The amount of the refund annuity depends upon three variables: (1) life expectancy rates for persons in the retiree's age group; (2) assumed interest rates; and (3) the size of the retiree's account balance at the time of retirement. The first two variables, life expectancy and assumed interest rates, are "actuarial equivalency factors" ("AEFs") used to convert the third variable, the retiree's account balance at the time of retirement, into the monthly refund annuity. See Or.Rev.Stat. § 238A.005(2).

PERS governing statutes mandate that the refund annuity be "the actuarial equivalent of accumulated contributions, if any, by the member and interest thereon at the time of retirement." Or.Rev.Stat. § 238.300. Thus, to convert a sum of money to a refund annuity, the PERS actuary multiplies that sum of accumulated contributions by the appropriate AEF from an annuity table adopted by Oregon's Public Employee Retirement Board ("Board"), the governing authority for PERS. For example, if a retiree had accumulated $100,000 in contributions and earnings, and the appropriate AEF for that retiree was 8.32 per $1,000, the retiree would receive a monthly refund annuity payment of $832 until death.

II. Henderson I and the 1978 Consent Decree

In 1974, four female PERS members filed a lawsuit in the United States District Court for the District of Oregon against the State of Oregon and the Board (" Henderson I"). The Henderson I plaintiffs alleged that the Board's practice of using two sets of life expectancy tables, one for men and another for women, in calculating refund annuities for retired PERS members constituted sex discrimination in violation of Title VII of the Civil Rights Act of 1964. The challenged practice resulted in lower monthly refund annuities for female PERS members, who were actuarially shown to live longer than their male counterparts.

The district court issued an order holding that the use of sex-segregated life expectancy tables violated Title VII, but staying any injunctive relief pending appeal. While the Henderson I appeal was pending, the United States Supreme Court ruled in City of Los Angeles Dep't of Water and Power v. Manhart, 435 U.S. 702, 98 S.Ct. 1370, 55 L.Ed.2d 657 (1978), that Title VII's prohibitions on sex-based discrimination applied to employee benefit plans.

Following Manhart, the parties agreed to a settlement. Thus, the district court vacated its order and judgment and entered a consent decree on September 20, 1978. The consent decree provided:

1. The judgment of this Court dated January 16, 1976 is set aside and this judgment is entered in lieu of it.

2. Title VII of the Civil Rights Act of 1964 prohibits the use of sex-segregated life expectancy tables in calculating "refund annuity" retirement allowances of employee members of the Oregon Public Employee Retirement System.

3. Defendant Public Employee Retirement System is permanently enjoined and restrained from the use of sex-segregated life expectancy tables in calculating "refund annuity" retirement allowance prospectively only for members retiring effective July 1, 1978, and thereafter, shall provide a monthly "refund annuity" retirement allowance to female members retiring after that date which is identical to the "refund annuity" retirement allowance males of the same age and amount received prior to that date. Defendant shall have no obligation to recalculate "refund annuity" retirement allowances to female members already retired or retiring before July 1, 1978.

4. Plaintiffs shall have and recover against defendants costs and counsel fees in the total amount of $7,000.00.

The Henderson I consent decree remains in effect today.

III. Legislative Reforms to PERS

The Board has used gender-neutral life expectancy tables ever since the district court entered the 1978 consent decree. Additionally, from 1978 through 2003, despite changing life expectancy and interest rate assumptions, the Board did not apply any updated AEFs that would have decreased a PERS member's projected monthly refund annuity.

The use of outdated AEFs, however, created unfunded liabilities for PERS. As a result, in 2003, Oregon Governor Kulongski signed the PERS Reform and Stabilization Act of 2003 ("Reform Legislation"). The Reform Legislation required that the Board adopt updated actuarial equivalency tables every two years. Accordingly, on June 10, 2003, the Board adopted new annuity tables based on current AEFs to calculate refund annuities for PERS members retiring after July 1, 2003.

The Reform Legislation provided that its constitutionality could be challenged directly in the Oregon Supreme Court. On March 8, 2005, the Oregon Supreme Court decided Strunk v. Public Employees Retirement Bd., 338 Or. 145, 108 P.3d 1058 (Or. 2005). Strunk rejected several constitutional and contractual challenges to the Reform Legislation. Specifically, Strunk rejected challenges to the requirement that the Board adopt and use current AEFs when calculating retirement benefits. Id. at 1109-10. Strunk further determined that the Board lacks the authority to use outdated AEFs. Id. at 1110.

IV. Henderson II and III

On October 16, 2003, six PERS members challenged the Board's use of updated life expectancy tables by filing a new action in the United States District Court for the District of Oregon (" Henderson II"). In Henderson II, the PERS members moved to reopen Henderson I and to hold the Board in civil contempt for violating the terms of the consent decree. They argued that the consent decree requires the Board "to provide all PERS members, retiring thereafter, a `refund annuity' allowance based upon the actuarial life expectancy tables in effect for males on July 1, 1978." At oral argument on June 10, 2004, the district court issued a minute order denying both the motion to reopen and the motion for civil contempt. The Henderson II plaintiffs appealed that minute order to this court.

On September 16, 2004, six PERS members (including two Henderson II plaintiffs) filed another action in the United States District Court for the District of Oregon (" Henderson III"). They sought to reopen Henderson I and requested a different remedy, a "declaration that the September 20, 1978 injunction requires [the Board] to adopt and use refund annuity tables which grant benefits not less than those benefits which were in effect for male members of [PERS] on July 1, 1978." On October 25, 2004, the district court issued a minute order denying the motion to reopen and denying as moot the motion for declaratory judgment. The Henderson

III plaintiffs also appealed that minute order to this court.

Henderson II and III were consolidated on appeal. On October 20, 2006, we issued a disposition addressing the issues raised in Henderson II and Henderson III. Henderson v. Oregon, 203 Fed.Appx. 45, 49-50 (9th Cir. 2006) (unpublished memorandum disposition). We held that the district court had not abused its discretion when it denied the motion for civil contempt filed in Henderson II because "viewing the language in the Henderson consent decree in favor of [the Board] . . . the consent decree is not sufficiently specific to give notice to [the Board] that it would violate the consent decree by updating the actuarial tables." Id. at 50. We further held that the motions to reopen were properly denied because the motions were not needed to invoke the district court's continuing jurisdiction over the Henderson I consent decree. Id. at 52.

However, we also held that the district court had improperly denied as moot the motion for declaratory judgment filed in Henderson III because the district court had never actually interpreted the language of the consent decree. Id. at 50-51. Accordingly, the panel remanded the Henderson litigation to the district court for a ruling on the meaning of the consent decree.

V. Current Appeal

On remand, the parties filed cross-motions for summary judgment regarding the correct interpretation of the consent decree. On April 18, 2007, the district court issued an order denying the PERS members' summary judgment motion and granting the Board's summary judgment motion. Henderson v. Oregon, 2007 WL 1187978 (D.Or. Apr.18, 2007). The district court held that the text of the consent degree was ambiguous, id. at *5, but that extrinsic evidence of the parties' intent demonstrated that the consent decree "does not require [the Board] to use refund annuity tables which grant benefits no less than the benefits in effect for males on July 1, 1978," id. at *8. The district court also struck certain paragraphs from affidavits filed in support of the PERS members' summary judgment motion.

On May 15, 2007, the PERS members filed their timely notice of appeal.

JURISDICTION

The district court had jurisdiction under 28 U.S.C. § 1331. We have jurisdiction under 28 U.S.C. § 1291.

STANDARD OF REVIEW

We review de novo a district court's decision on cross motions for summary judgment. See Arakaki v. Hawaii 314 F.3d 1091, 1094 (9th Cir. 2002). On appeal, we must determine, viewing the evidence in the light most favorable to the nonmoving party, whether there are any genuine issues of material fact and whether the district court correctly applied the relevant substantive law. See Olsen v. Idaho State Bd. of Medicine, 363 F.3d 916, 922 (9th Cir. 2004).

We review the district court's evidentiary rulings for an abuse of discretion. Tritchler v. County of Lake, 358 F.3d 1150, 1155 (9th Cir. 2004).

DISCUSSION

I. The Consent Decree Does Not Require a Perpetual Floor on Benefits Measured by the Refund Annuity Allowance Rates in Effect for Males in 1978

Our main task on appeal is to construe the consent decree. "Because a consent decree has attributes of both a contract and a judicial act, courts use contract principles in construing it." Washington v. Penwell, 700 F.2d 570, 573 (9th Cir. 1983) (internal citation omitted), abrogated on other grounds as recognized by Jeff D. v. Kempthorne, 365 F.3d 844, 852 (9th Cir. 2004). Here, the parties entered into the consent decree in Oregon. Therefore, Oregon law on contract interpretation applies. Gates v. Gomez, 60 F.3d 525, 530 (9th Cir. 1995) ("A consent decree is construed with reference to ordinary contract principles of the state in which the decree is signed.").

Oregon courts follow three steps when interpreting contractual provisions. First, they examine "the text of the disputed provision, in the context of the document as a whole." Yogman v. Parrott, 325 Or. 358, 937 P.2d 1019, 1021 (Or. 1997). "If the provision is clear, the analysis ends." Id. However, if the contractual provision at issue is ambiguous, Oregon courts follow the second step of "examining] extrinsic evidence of the contracting parties' intent." Id. at 1022. If the second step fails to resolve the ambiguity, Oregon courts rely on "appropriate maxims of construction." Id.

A. Text of the Consent Decree

We turn first to the text of the consent decree, viewed in the context of the document as a whole. The third paragraph of the consent decree contains the disputed provision:

3. Defendant Public Employee Retirement System is permanently enjoined and restrained from the use of sex-segregated life expectancy tables in calculating "refund annuity" retirement allowance prospectively only for members retiring effective July 1, 1978, and thereafter, shall provide a monthly "refund annuity" retirement allowance to female members retiring after that date which is identical to the "refund annuity" retiremen allowance males of the same age and amount received prior to that date. Defendant shall have no obligation to re-calculate "refund annuity" retirement allowances to female members already retired or retiring before July 1, 1978.

(Emphasis added.)

The PERS members argue that the disputed provision means that, from the date of the consent decree onward, all PERS members were entitled to receive refund annuity allowances in the same amount as those received by male PERS members before 1978. Thus, the PERS members believe that the consent decree created a perpetual "floor" on benefits. They argue that the "thereafter" clause is superfluous if it is not interpreted to create a benefits floor.

By contrast, the Board argues that the disputed provision means that, from the date of the consent decree onward, all similarly situated male and female PERS members were entitled to receive identical refund annuity allowances. The Board contends that the "thereafter" clause memorializes the Board's decision to "top up" (i.e., raise) female allowances on July 1, 1978 to the male allowance levels in effect prior to that date; the clause does not demonstrate an agreement between the parties to lock in a permanent allowance rate.

As we previously stated in our 2006 memorandum disposition, given the Title VII context of this case, both parties' constructions of the disputed provision are arguably plausible. There, we explained:

The "prior to that date" language of the Henderson consent decree to a degree supports [the] argument that the consent decree literally established a floor for female employee refund annuity tables. But taken in its context, a Title VII claim of sex discrimination, in which parity on payments to men and women was the goal, the consent order's prohibition might be read only to order PERB to stop using segregated actuarial tables, and from July 1, 1978, onward, to use identical refund annuity tables for women and men.

203 Fed.Appx. at 49-50.

Because the disputed provision is susceptible to at least two plausible interpretations when examined in the context of the contract as a whole, it is legally ambiguous. Moon v. Moon, 140 Or.App. 402, 914 P.2d 1133, 1135 (Or. 1996). Therefore, we proceed to the extrinsic evidence analysis.

B. Extrinsic Evidence of the Parties' Intent

The PERS members make two main arguments that the extrinsic evidence of the parties' intent supports their interpretation of the consent decree. We address those arguments in turn.

1. Retroactive Relief

The PERS members first argue that the consent decree embodied a "quid pro quo" exchange of their rights to retroactive relief under Title VII in return for a permanent benefits floor. In support of this argument, they point to the last sentence of paragraph three of the consent decree:

Defendant shall have no obligation to recalculate "refund annuity" retirement allowances to female members already retired or retiring before July 1, 1978.

The PERS members maintain that this sentence was included in the consent decree to memorialize the parties' exchange of rights.

The Board argues that no such exchange of rights was contemplated by this sentence. Instead, the sentence was merely intended to clarify the scope of relief. The Board also argues that the original Henderson litigants did not have any such rights to retroactive relief, and, even if they did, the Board would not have opted to create a permanent benefits floor to avoid paying retroactive relief.

To understand these arguments, we turn to the state of Title VII case law in 1978. The parties entered into the consent decree after the Supreme Court issued its decision in Manhart. The Court held for the first time in Manhart that Title VII's prohibitions on sex-based discrimination applied to employee benefit plans. Nevertheless, the Court reversed the district court's award of retroactive relief to the retirees, reasoning that "retroactive liability could be devastating for a pension fund" because the "harm would fall in large part on innocent third parties." 435 U.S. at 723, 98 S.Ct. 1370.

When viewed from hindsight, the timing and holding of Manhart suggest that the original Henderson plaintiffs would not have believed that they were entitled to retroactive relief. But in 1978, Manhart was new law, and Title VII case law on retroactivity was not uniform. Moreover, it was not entirely clear that Manhart's holding on retroactivity applied to the facts of the Henderson litigation. Thus, it is theoretically possible that the parties were concerned about retroactive relief when drafting the 1978 consent decree. However, this does not necessarily mean that any quid pro quo arrangement with respect to a permanent refund annuity allowance rate was reached.

Therefore, we look to the evidence in the record. The PERS members contend that the assumptions of the Board and its intent on retroactivity are revealed by the briefs filed in Henderson I, the district court's pre- Manhart decision in Henderson I, the understandings of the parties as expressed by minutes from the Board's 1978 meeting about the consent decree, the affidavit of the 1978 PERS Director, the affidavit of the 1978 PERS Assistant Director, and the Affidavit of PERS counsel.

Our review of this evidence, however, reveals that it does not support the PERS members' contention. There is nothing in the Henderson I briefs, the district court's pre- Manhart decision, or the various affidavits that indicates the parties believed that the PERS members were waiving any rights they may have had to retroactive relief in exchange for a permanent floor on refund annuity rates. By contrast, the minutes from the 1978 meeting actually support the Board's position. At that meeting, the Board discussed whether to give female PERS members the same benefits as male PERS members, or whether to adopt an average between the male and female rates. The Board's attorney "questioned the legality of such a reduction in benefits" for male members. Then, "[t]o resolve the matter the Board directed [the attorney] to attempt a negotiated consent judgment with the plaintiffs, agree to increase the female annuity factor to that of the male rate effective July 1, 1978 and thereafter." Later in that same meeting, the Board acted to top up female benefits by passing a motion that "the refund annuity benefit for females be equal to that received by the males." This language makes it clear that the reason the Board chose to top-up female benefits was its concern over a reduction in male benefits, and not a concern about retroactive damages. This language also indicates that the Board's goal was to equalize male and female rates from 1978 onward. Tellingly, the minutes say nothing about locking in the 1978 rates.

Finally, as the district court recognized, "[e]ven if the settlement released retroactive liability in exchange for topping up the tables for female members, that does not mean that the newly topped-up benefit was to act as a floor in perpetuity." The Board had two options available to desegregate its refund annuity allowance system: (1) it could create blended rates, or (2) it could "top up" the female rates. "Topping up" the female rates was the best option for the PERS members. Thus, even if PERS members did consciously abandon any rights to retroactive relief, they may have decided to do so in return for the "topping up" option, not the creation of a permanent benefits floor.

In short, we find that, even viewing the evidence on retroactive relief in the light most favorable to the PERS members, the parties did not intend to create a permanent floor on benefits by entering into the consent decree.

2. Course of Conduct

The PERS members also argue that the Board's course of conduct after 1978 supports their interpretation of the consent decree. Again, we disagree.

It is undisputed that, from the date of the consent decree until the Reform Legislation was enacted in 2003, the Board maintained refund annuity rates at or above those in effect for males on July 1, 1978, at least for those members who joined PERS prior to 1999. The PERS members contend that the Board's practice of maintaining annuity rates at those levels suggests that the Board believed that lowering rates would violate the consent decree. However, as discussed above, the minutes from the Board meeting suggest that the Board's refusal to lower annuity rates was attributed to a fear that male PERS members might have vested property rights in the 1978 rates — not to a fear that lowering the rates would violate the consent decree. Moreover, as the Oregon Supreme Court explained in Strunk, the Board adopted administrative rules in the 1990s to maintain the rates at or above the 1978 levels "in response to legal advice that it had received regarding maintaining the fund's status as a qualified governmental retirement plan and trust under the Internal Revenue Code." Strunk, 108 P.3d at 1106. Finally, there is no contemporaneous documentary evidence indicating that, when the Board modified the rates in the years following entry of the consent decree, it was concerned about a benefits floor deriving from the consent decree.

In sum, we find that the Board's continued use of the 1978 male annuity refund rates cannot be attributed to its fear of violating the consent decree. To the contrary, the evidence suggests that the Board did not dip below the 1978 rates due to other, unrelated legal concerns.

C. Conclusion

For the foregoing reasons, we hold that the extrinsic evidence of the parties' intent demonstrates that the consent decree was intended only as a remedy for sex discrimination in the calculation of refund annuity payments.

We note that the overall context of this case plays an important role in our decision. This case began as a Title VII law-suit. The Henderson I plaintiffs filed the suit to end PERB's practice of using sex-segregated refund annuity rates — not to establish that all PERS members were entitled to a minimum level of benefits. It is highly unlikely that the parties intended to create a permanent floor on refund annuity rates when they stipulated to the consent decree. It is much more likely that they decided to equalize the rates from 1978 onward, and tried to draft the consent decree to that effect.

Accordingly, we affirm the district court's summary judgment order on the meaning of the consent decree.

II. The District Court Did Not Abuse Its Discretion When It Struck Certain Paragraphs From the PERS Members' Summary Judgment Affidavits

Finally, we address the PERS members' two challenges to the district court's evidentiary rulings on summary judgment. To reverse on the basis of an erroneous evidentiary ruling, the court must conclude not only that the district court abused its discretion, but also that the error was prejudicial. Tritchler v. County of Lake, 358 F.3d 1150, 1155 (9th Cir. 2004).

First, the PERS members challenge the district court's refusal to consider paragraphs two through four, and parts of paragraph seven, of the affidavit of former PERS attorney William Hoelscher. However, as the district court correctly noted, paragraphs two through four of the affidavit, which described the holdings of Henderson I and Manhart, merely stated legal conclusions. Moreover, the language stricken from paragraph seven contained improper speculation about the judge's understanding of the case. Thus, the district court did not abuse its discretion in deciding that these paragraphs were not admissible summary judgment evidence.

Second, the PERS members challenge the district court's decision to strike paragraphs four, six, and eight of the virtually identical affidavits submitted by the former PERS Director and Assistant Director. These paragraphs described the negotiations that occurred at the 1978 Board meeting. The district court struck the paragraphs because they lacked any facts demonstrating that the PERS Director or Assistant Director were actually at that meeting. The affiants' assertions about a meeting which they apparently did not attend and about which they had no personal knowledge are not the proper subject of an affidavit. Thus, the district court also did not abuse its discretion in striking these paragraphs.

CONCLUSION

The consent decree was intended to ensure that male and female PERS members who had the same balances in their PERS accounts at the time of retirement would receive identical refund annuities from the date of the consent decree onward. The consent decree was not meant to lock in the 1978 refund annuity rates. Therefore, we affirm the judgment of the district court.

AFFIRMED.