BRYSON, Circuit Judge.

This is an appeal from a judgment of the Court of Federal Claims in a takings case. Independence Park Apartments v. United States, 61 Fed.Cl. 692 (2004). This case has had an extended history and now reaches this court for the fourth time. In the immediately preceding appeal to this court, Cienega Gardens v. United States, 331 F.3d 1319 (Fed. Cir. 2003) ( "Cienega VIII"), we held that federal statutes enacted in 1988 and 1990 caused a compensable, temporary taking of certain rights that the plaintiffs had under contractual arrangements with the Department of Housing and Urban Development ("HUD"). On remand, the Court of Federal Claims awarded the plaintiffs a total of $3,388,208.43 as just compensation for the taking. The United States and the plaintiffs have both appealed from that judgment. We reverse and remand for a recalculation of damages.

I A

The plaintiffs are four owners of rental property in Los Angeles. In the 1970s they entered into agreements with HUD as part of a nationwide federal program designed to increase the supply of low-income housing by providing incentives to owners and developers of such housing. Under the agreements, HUD subsidized 40-year mortgages for the plaintiffs, and the plaintiffs agreed to abide by restrictions on the mortgaged property, such as making the property available to low-income tenants, keeping rents significantly below market rates, and not selling or encumbering the property without HUD approval. Each agreement provided that it would remain effective throughout the life of the mortgage. A rider to each agreement provided that the owner could not prepay its mortgage without HUD approval during the first 20 years of the mortgage. After 20 years, however, the owner could prepay its mortgage and thereby free itself from the restrictions in the regulatory agreement.

In the late 1980s, when a large number of the mortgages were nearing the 20-year mark, HUD became concerned that many of the owners might exercise their prepayment rights, resulting in the removal of hundreds of thousands of units from the nation's supply of low-income housing. Congress responded in 1988 and 1990 by enacting two statutes forbidding owners in the federal program from prepaying their mortgages without HUD approval. The first, the Emergency Low Income Housing Preservation Act ("ELIHPA"), Pub.L. No. 100-242, tit. II, 101 Stat. 1877 (1988), was a temporary measure. The second, the Low-Income Housing Preservation and Resident Homeownership Act of 1990 ("LIHPRHA"), Pub.L. No. 101-625, tit. VI, 104 Stat. 4249, made the restriction permanent. The two statutes also empowered HUD to enter into "use agreements" with owners in which HUD offered the owners further incentives to remain subject to the federal restrictions on their property. In 1996, Congress enacted another statute, the Housing Opportunity Extension Act of 1996 ("the HOPE Act"), Pub.L. No. 104-120, 110 Stat. 834, which effectively removed the prepayment restrictions, although it did not expressly repeal LIHPRHA. In the late 1990s, Congress stopped funding the use agreements.

B

In 1994, the four plaintiffs, along with 38 others, sued the United States in the Court of Federal Claims on a number of theories, including breach of contract and violation of the Takings Clause of the Fifth Amendment. The Court of Federal Claims granted summary judgment to the plaintiffs on their breach of contract claims, and the parties selected the four plaintiffs in the instant case as "model plaintiffs" for a trial on damages. See Cienega Gardens v. United States, 38 Fed. Cl. 64, 67 n. 3 (1997) ( "Cienega III"). The court then held a damages trial. At the conclusion of the trial, the court awarded the four plaintiffs a total of $3,061,107 as damages for the breach of contract consisting of the statutory ban on mortgage prepayment. On appeal, we vacated the award because of lack of privity of contract between the plaintiffs and the United States. Cienega Gardens v. United States, 194 F.3d 1231, 1245 (Fed. Cir. 1998) ( "Cienega IV"). After further proceedings both in the trial court and on appeal, the case reached us again in 2003, when we were presented with the question whether ELIHPA and LIHPRHA had caused the plaintiffs to suffer a compensable temporary regulatory taking under the test established by the Supreme Court in Penn Central Transportation Co. v. New York City, 438 U.S. 104, 98 S.Ct. 2646, 57 L.Ed.2d 631 (1978). Relying on the factual record already developed in this case, we concluded in Cienega VIII that the plaintiffs had suffered a regulatory taking. We then directed the Court of Federal Claims on remand to consider the breach of contract damages award as just compensation for the taking and to adjust the award as appropriate.

Following the remand proceedings, the Court of Federal Claims awarded the plaintiffs a total of $3,388,208.43. The government sought to argue that the plaintiffs' damages should be reduced because even if they had prepaid their mortgages after 20 years, the Los Angeles Rent Stabilization Ordinance, L.A. Mun. Code, ch. XV, §§ 151.00-151.20 ("LARSO"), would have applied to their property and limited the rent that they could have charged tenants after leaving the federal program. The Court of Federal Claims rejected that argument, holding that it was barred by this court's mandate and by the doctrine of law of the case from entertaining that contention. In the alternative, the court held that LARSO would not have reduced the amount of the award because it was expressly preempted by the preemption provision of LIHPRHA, which was codified at 12 U.S.C. § 4122(a). Independence Park, 61 Fed.Cl. at 705-06.

Two of the plaintiffs argued that they were entitled to an increase in the damages award on the basis of what they characterized as additional costs that they incurred in connection with use agreements that they entered into in an effort to mitigate their damages before the mortgage prepayment rights were restored in 1996. The court, however, rejected that argument. Id. at 700-02.

On appeal, the United States challenges (1) the trial court's refusal to reduce the damages award on the ground that, if the owners had exercised their prepayment rights and escaped federal restrictions, they would have been subject to LARSO, and (2) the trial court's selection of the "valuation date" for the temporary regulatory taking as being the end of the takings period rather than as of the beginning. The plaintiffs cross-appeal, challenging the trial court's refusal to increase the award to two of the plaintiffs based on their claimed mitigation costs.

For the reasons set forth below, we hold that the trial court erred in declining to consider the effects of the Los Angeles rent control ordinance in calculating the damages from the taking. We therefore remand for reconsideration of the quantum of damages in light of the effect of the Los Angeles rent control ordinance on the compensation due to the plaintiffs. On the cross-appeal, we remand for a recalculation of damages for the two plaintiffs that entered into use agreements before the passage of the HOPE Act.

II

We first address the question whether the Court of Federal Claims erred in holding that this court's mandate in Cienega VIII and the law of the case doctrine preclude the United States from litigating the effect of LARSO on the plaintiffs' damages award. The plaintiffs make two arguments in defense of the judgment below. First, they argue that our holding in Cienega VIII that the plaintiffs had suffered a "serious financial loss" under the second prong of the Penn Central test foreclosed the government from arguing the effect of LARSO on remand. The plaintiffs reason that the decision in Cienega VIII rested in part on the trial court's holding in Cienega III that LARSO was preempted by federal law and thus did not affect the calculation of the plaintiffs' damages. Accordingly, they contend, the preemption ruling in Cienega III became law of the case. Second, the plaintiffs argue that the government waived the right to argue the effect of LARSO on remand by failing to raise that issue before this court in Cienega VIII.

We conclude that our mandate in Cienega VIII is best understood as leaving the preemption issue open for remand. In Cienega III, the trial court awarded the plaintiffs $3,061,107 as damages for breach of contract, reasoning that the plaintiffs were unlawfully prevented from prepaying their mortgages, and that the award was an appropriate estimate of the amount of money necessary to compensate the plaintiffs for the loss of their right to prepay their mortgages. In reaching that conclusion, the court assumed that the plaintiffs would have prepaid their mortgages after 20 years and that they would have raised the rental rates on their properties to market levels after prepayment. The government countered by arguing that even if the plaintiffs had prepaid their mortgages, LARSO would have restricted their ability to raise rents, and that the damages award should be reduced accordingly. The Court of Federal Claims rejected that argument, holding that LARSO was preempted by federal law. See Cienega III, 38 Fed.Cl. at 81-85. On appeal from that judgment, we vacated the damages award based on lack of privity of contract between the United States and the plaintiffs. In our opinion in that appeal and in a subsequent appeal, we explicitly declined to address the applicability of LARSO. See Cienega IV, 194 F.3d at 1239; Cienega VI, 265 F.3d at 1247.

When the case reached this court for the third time in Cienega VIII, we were presented with the plaintiffs' regulatory taking theory, rather than their breach of contract theory. The plaintiffs urged us to rely on the factual findings made in Cienega III to support their argument that they had suffered a financial loss sufficient to satisfy the "economic impact" prong of the Penn Central test. With respect to those plaintiffs whose properties were located in Los Angeles, the government could have responded, as it had argued in earlier stages of the litigation, that their financial loss was not as great as they asserted because any consideration of the plaintiffs' situation in the absence of LIHPRHA and ELIHPA would have to take LARSO into account. Instead, the government argued that the facts of Cienega III showed that the plaintiffs had not suffered a taking at all, because they had lost no more than 35% of the value of their properties as a result of the enactment of LIHPRHA and ELIHPA. The government noted that the Court of Federal Claims had observed in another case that "diminutions in value approaching 85 to 90 percent do not necessarily result in a regulatory taking."

We found that argument unconvincing. Although we did not address LARSO at all in Cienega VIII, we held — based on the factual findings made in Cienega III — that the plaintiffs had suffered a "serious financial loss" by being deprived of the opportunity to prepay their mortgages. See Cienega VIII, 331 F.3d at 1340-45. Based on our analysis of the Penn Central factors and the extensive record already developed in the case, we ruled in favor of the plaintiffs on their regulatory takings claims and directed that the original damages award entered in Cienega III be reinstated. Id. at 1353. The government filed a rehearing petition, but in the petition it merely reiterated arguments made in its briefs and did not mention LARSO.

Before the mandate in Cienega VIII issued, the plaintiffs filed a "motion regarding mandate" asking this court to amend the mandate to make clear that the Court of Federal Claims could alter the Cienega III award on remand to add interest. In its response, the government agreed that the court should alter the mandate. It argued, however, that the court should do so not only to permit consideration of the plaintiffs' arguments for an increase in the award, but also to permit consideration of the government's arguments that would decrease the award. In particular, the government referred to the argument that the damages should be reduced to account for the effect of the Los Angeles rent control ordinance. The government noted the Ninth Circuit's decision in Topa Equities, Ltd. v. City of Los Angeles, 342 F.3d 1065 (9th Cir. 2003), in which that court held that LARSO was not preempted by federal law.

In response to those filings, the Cienega VIII panel amended the mandate by adding the following sentence:

On remand, the trial court may adjust the original damages award reinstated by this court if it is shown by either party not to compensate accurately for the regulatory taking, and may also determine whether interest is or is not due.

The plaintiffs had asked only for the right to request interest, so the first half of the sentence necessarily responded to the government's request. The government had argued that each party should be given an opportunity on remand to argue the proper quantum of just compensation damages, "including the issues that both parties did not have an opportunity to litigate," and that is the relief that the panel granted in the first half of the sentence amending the mandate. Because it is clear from the context that the amendment to the mandate left open the question of the effect of LARSO on the amount of the plaintiffs' damages, we reject the plaintiffs' contention that the mandate rule or the law of the case doctrine bars the government from raising the LARSO issue in this appeal.

We also reject the plaintiffs' argument that the government's failure to raise the LARSO issue in its brief in Cienega VIII resulted in a waiver of that issue. The government was the appellee in Cienega VIII, and thus was in the position of defending a favorable judgment that did not rest on LARSO grounds. As appellee, the government was not required to raise all possible alternative grounds for affirmance in order to avoid waiving any of those grounds. See Laitram Corp. v. NEC Corp., 115 F.3d 947, 954 (Fed. Cir. 1997); Schering Co. v. Ill. Antibiotics Co., 89 F.3d 357, 358 (7th Cir. 1996); Crocker v. Piedmont Aviation, Inc., 49 F.3d 735, 740-41 (D.C. Cir. 1995). While the government should have made clear in that appeal that it was not abandoning its argument based on LARSO, it is perhaps understandable that it did not press that argument at that time, since the argument applied only to those plaintiffs with properties in Los Angeles and thus did not apply at all to many of the plaintiffs who were parties to that appeal. Thus, while the government's failure to allude to the LARSO issue in Cienega VIII has led to an unfortunate multiplicity of proceedings in this already protracted litigation, we conclude that the government's failure to press the LARSO issue in its brief in that appeal did not result in a binding waiver of that issue for all subsequent proceedings.

III

On the merits, we hold that the Court of Federal Claims erred in ruling that LARSO does not affect the damages calculation because its relevant provisions were preempted by federal law. An assumption underlying the damages award in Cienega III was that, in the absence of ELIHPA and LIHPRHA, the plaintiffs would have prepaid their mortgages after 20 years and would have raised their rents to market levels. If LARSO would have been applicable to the plaintiffs' properties after they had prepaid their mortgages, however, they would not have been able to raise their rents freely. The question whether LARSO would have restricted the rents that the plaintiffs could have charged after prepaying their mortgages is therefore important to the calculation of damages in this case.

The Los Angeles City Council originally passed LARSO in 1979 to control rent increases in rental housing within the city. For rental housing units within its scope, LARSO restricts rent increases in two ways. First, it provides that, as long as the same tenant occupies a unit, the landlord (absent permission from the city's Rent Adjustment Commission) may raise rent only once every 12 months, and then only by an amount based on the Consumer Price Index. LARSO § 151.06(D) (1996). Second, LARSO's "vacancy decontrol" provision states that, when a unit is vacant due to a tenant either having left voluntarily or having been evicted for one of certain enumerated reasons, the landlord may freely increase the rental rate upon renting to the subsequent tenant. Id. § 151.06(C)(1) (1996). However, if the landlord either evicts a tenant for a reason not listed in the ordinance, or causes the tenant to leave "by creating an unreasonable interference with the tenant's comfort, safety, or enjoyment of the rental unit," then the vacation of the unit is deemed involuntary and the owner is not allowed to raise the rental rate upon finding a replacement tenant. Id. § 151.06(C)(2). The ordinance exempts certain types of rental housing from its reach by excluding them from the definition of "rental units." Id. § 151.02(H) (1996). Before 1990, LARSO exempted "[h]ousing accommodations which a government unit, agency or authority owns, operates, or manages, or which are specifically exempted from municipal rent regulation by state or federal law or administrative regulation." LARSO § 151.02(H) (1979). The City Council amended LARSO in 1990 to make clear that, although rental units that were part of a federal low-income housing program were not subject to the ordinance, their rents would be controlled by LARSO after the units left the federal housing program. See L.A. Mun. Ordinance No. 166,320 (Oct. 17, 1990).

A

The Court of Federal Claims held that LARSO was preempted by LIHPRHA's preemption provision, 12 U.S.C. § 4122. See 61 Fed.Cl. at 704-05. That statute expressly preempted any state or local law that "restricts or inhibits the prepayment of any mortgage described in [LIHPRHA] on eligible low income housing." The statute, however, also contained a savings clause removing from the scope of the preemption provision local laws that were "not inconsistent with the provisions of this subchapter, such as any law or regulation relating to . . . rent control, [and other examples], to the extent such law or regulation is of general applicability to both housing receiving Federal assistance and nonassisted housing."

We held in Cienega VIII that the enactment of ELIHPA and LIHPRHA effected the taking at issue in this case by depriving the plaintiffs of their "contractual and regulatory rights to post-twentieth-year prepayment and under real property law to repossess." 331 F.3d at 1337, 1353. In order to determine the amount taken, we must compare the rights the plaintiffs had before the enactment of that legislation with the rights they had afterwards. See First English Evangelical Lutheran Church v. Los Angeles County, 482 U.S. 304, 320, 107 S.Ct. 2378, 96 L.Ed.2d 250 (1987). The plaintiffs argue that the enactment of ELIHPA and LIHPRHA deprived them of valuable rights because those statutes barred them from freely prepaying their mortgages and then raising their rents to market levels. The extent of the loss they suffered as a result of those statutes depends, however, on whether, if they had been free to prepay their mortgages, LARSO would have prevented them from raising their rents to market levels. If LARSO would have been inapplicable to the plaintiffs before the enactment of ELIHPA and LIHPRHA, either because of federal preemption or for some other reason, their takings claims would be of considerable value. If LARSO would have been applicable, however, their takings claims would be much less valuable.

Although the parties frame the preemption issue as whether LARSO was preempted by LIHPRHA, that is not the correct inquiry. The enactment of LIHPRHA was the measure that is argued to have constituted the taking that caused the plaintiffs' losses. In order to determine the degree of the loss caused by LIHPRHA, it is necessary to look to whether LARSO would have applied to the plaintiffs in the absence of LIHPRHA; that is, whether LARSO would have prevented the plaintiffs from raising the rents in their rental properties even if they had been free to prepay their mortgages. Whether or not ELIHPA or LIHPRHA preempted LARSO at the same time that those statutes barred the plaintiffs from prepaying their mortgages is thus irrelevant. What matters is whether LARSO was preempted prior to the enactment of ELIHPA and LIHPRHA.

The following example may clarify the point. Suppose that, prior to 1988, developers were entitled to prepay their federally subsidized mortgages after 20 years, but the federal statute in effect during that period made it clear that rent control ordinances were not preempted and would apply fully to any developers who prepaid their mortgages. Suppose further, for purposes of simplicity, that the rents allowed under the local rent control ordinance were identical to the rents that developers still subject to the federal subsidized housing program were allowed to charge. Finally, suppose that with that background Congress enacted a statute that prohibited developers with HUD-subsidized mortgages from prepaying their mortgages. Under those circumstances, any developers who were subject to local rent control ordinances would not suffer any injury as a result of the prohibition on prepayment, because they would not be able to charge higher rents after repaying their mortgages than they could before the mortgages were repaid. Moreover, even if the hypothetical 1988 statute that barred prepayment of mortgages also plainly preempted all local rent control ordinances, the developers could not claim that the statute deprived them of property having substantial value. That is because the statute would have granted a right in one provision (freedom from the ordinances), while taking the benefits of that right away in another (barring prepayment of mortgages). The net effect of the statute would be to leave the developers exactly where they were before its enactment. It would be improper to hold that a takings judgment can be predicated on one portion of the statute while ignoring that the statute as a whole has left the potential plaintiffs no worse off than they were before. We therefore reject the contention that statutory preemption bars consideration of the effects of LARSO in calculating the damages award.

B

In defending the judgment below, the plaintiffs make the alternative argument that even if LARSO is not expressly preempted by statute, it "would have been conflict preempted for frustrating Congress's long-standing policy of encouraging private developers to service the nation's low-income housing needs." Their argument is essentially that when Congress established its low-income housing program in the 1950s and 1960s, the prepayment option was a valuable incentive that enticed owners and developers to participate in the program. As such, they argue, LARSO is preempted because it conflicts with that program by reducing the value of that incentive; in effect, they contend that owners who knew they would be subject to local rent control laws upon leaving the federal program would be less likely to find the promise of prepayment sufficient to entice them to participate in the federal program.

This argument is inconsistent with general principles of implied conflict preemption and with cases that have applied those principles in settings closely analogous to the setting of this case. In Freightliner Corp. v. Myrick, 514 U.S. 280, 115 S.Ct. 1483, 131 L.Ed.2d 385 (1995), the Supreme Court summarized the principles of implied conflict preemption, noting the high threshold for finding state and local laws impliedly repealed:

We have recognized that a federal statute implicitly overrides state law either when the scope of a statute indicates that Congress intended federal law to occupy a field exclusively, or when state law is in actual conflict with federal law. We have found implied conflict pre-emption where it is "impossible for a private party to comply with both state and federal requirements," or where state law "stands as an obstacle to the accomplishment and execution of the full purposes and objectives of Congress."

Id. at 287, 115 S.Ct. 1477 (citations omitted). Where the claim of implied conflict preemption is asserted "to bar state action in fields of traditional state regulation," moreover, the Court has made it clear that preemption is even less readily found. In Medtronic, Inc. v. Lohr, for example, the Court explained that "in a field which the States have traditionally occupied," courts "start with the assumption that the historic police powers of the States were not to be superseded by the Federal Act unless that was the clear and manifest purpose of Congress." 518 U.S. 470, 485, 116 S.Ct. 2240, 135 L.Ed.2d 700 (1996) (quoting Rice v. Santa Fe Elevator Corp., 331 U.S. 218, 230, 67 S.Ct. 1146, 91 L.Ed. 1447 (1947)); see also Cal. Div. of Labor Standards Enforcement v. Dillingham, N.A., Inc., 519 U.S. 316, 325, 117 S.Ct. 832, 136 L.Ed.2d 791 (1997).

This case is not one in which conflict preemption overrides the application of traditional state and local police powers. Although the plaintiffs' loan agreements and HUD's regulations gave the plaintiffs the right to prepay their mortgages, it is an unjustified leap to conclude, based on the prepayment right, that the federal housing laws preempted local rent control ordinances and made them inapplicable, for some undefined period and to some undefined extent, to properties that were previously the beneficiaries of federal support under the federal housing laws. The National Housing Act provided certain benefits and imposed certain burdens on owners of subsidized low-income housing. It did not, however, provide them with any protection against the application of a variety of state and local laws that could affect the profitability of their investments. Thus, the National Housing Act plainly did not bar or limit the application of local statutes and ordinances governing such matters as construction standards and housing conditions, even after the participants in the low-income housing program left the program. For the same reasons, the National Housing Act cannot reasonably be deemed to preempt local laws governing allowable rents after the mortgages were paid off and the properties emerged from the federal regulatory system.

Applying the principles of implied conflict preemption, the courts that have addressed claims that local rent control ordinances are preempted by the federal housing statutes have rejected those claims. In Topa Equities, Ltd. v. City of Los Angeles, 342 F.3d 1065 (9th Cir. 2003), the court rejected a claim that the National Housing Act preempted LARSO, the same rent control statute that is at issue in this case. In response to an argument identical to that made by the plaintiffs here — that LARSO would interfere with the federal scheme by depriving developers of the opportunity to realize market rates after emerging from the program — the court responded as follows:

There is no assurance in the HUD regulations, or in any other federal statute or regulation, that after 20 years an owner may raise rents to market levels. The assurance is that after 20 years an owner may prepay his federally subsidized mortgage, exit the federal program and free himself of federal regulation, including federal rent control. Nothing in the HUD regulations purports to limit states from enacting their own rent control laws of general applicability which apply equally to apartment owners who exit the federal program as well as other apartment owners. . . . Congress never indicated — in either the text or legislative history of the [National Housing Act] or in any ancillary statute — that it intended to abrogate state rent control laws.

342 F.3d at 1071-72.

Similarly, in Kargman v. Sullivan, 552 F.2d 2 (1st Cir. 1977), the First Circuit held that local rent control laws did not interfere impermissibly with federal housing laws so as to give rise to implied conflict preemption of the local laws. The court in that case stated that the "federal legislation creating a network of subsidized housing laws is superimposed on and consciously interdependent with the substructure of local law relating to housing," and that the federal law therefore did not preempt the application of local rent control ordinances. Id. at 11. This case follows a fortiori from that one, in that the local rent control regulations at issue in that case were applied to properties that were still subject to federal housing regulation, not simply to properties that had emerged from federal regulation pursuant to a mortgage prepayment. Nonetheless, the court held that in the absence of any specific, contrary rule in federal statute or regulation, the local rent control ordinance was not preempted.

The principles of Topa and Kargman compel the conclusion that the application of LARSO to the plaintiffs following the prepayment of their HUD-subsidized mortgages would not be preempted. That being the case, it is incorrect for the plaintiffs to assert that the scope of the taking resulting from the enactment of ELIHPA and LIHPRHA should be determined based on the assumption that if they had been permitted to prepay their mortgages, they would be entitled to charge market rates for their rental properties, regardless of the restrictions imposed on rent levels by LARSO.

The plaintiffs' responses to the government's argument on this point are unconvincing. First, they argue that LIHPRHA had a severability provision and that its unconstitutional provisions must be severed from its constitutional provisions for purposes of assessing its effect on the plaintiffs' rights. That argument is a nonsequitur. The question before us is not whether any particular provision of LIHPRHA, taken by itself, would be unconstitutional in the absence of compensation. Instead, the question is what is the value of the property taken by the enactment of LIHPRHA. The issue of severability has nothing to do with that question.

Likewise unconvincing is the plaintiffs' contention that LARSO "frustrated the developers' ability to exercise the prepayment option HUD provided to induce them to participate in the federal low-income housing programs." Significantly, HUD's regulations (which, along with the mortgage-subsidy agreements, were the source of the prepayment right) preempted local rent control laws as applied to federally regulated properties, but only during the period that the properties were subject to federal regulation. See 24 C.F.R. §§ 246.20, 246.21. The federal government has never suggested that either the regulations or the mortgage subsidy program in general preempt local rent control laws as applied to properties that have emerged from federal regulation.

The local rent control ordinance at issue in this case did not bar or otherwise restrict prepayment of the HUD-subsidized mortgages. Instead, it merely affected the degree to which prepayment would have beneficial downstream financial consequences for the owners. That kind of indirect effect on the prepayment right is not enough to serve as a basis for preemption under the strict standards outlined above, as the First and Ninth Circuits have held. In this respect, a rent control ordinance is not different in kind from other state or local laws that may affect the profitability of rental properties, such as zoning laws, building codes, housing standards, or tax provisions. Each may affect the profitability of the rental property and thus the decision to prepay the mortgage, but none would be preempted by the National Housing Act.

Finally, the plaintiffs argue briefly that if they had been allowed to prepay their mortgages, HUD would likely have provided the tenants with vouchers that would have enabled the owners to raise the rents to market levels, with HUD making up the difference between the rental rates allowed by the local rent control ordinance and the market rates. Under such a regime, the plaintiffs argue, the properties would have remained exempt from LARSO. The trial court, however, did not adopt that theory of liability and made no findings with respect to that issue. It therefore does not provide an alternative basis for us to uphold the damages award. Our decision today does not bar the trial court from considering that argument on remand.

Accordingly, we reverse the judgment to the extent that it was based on rejection of the government's argument regarding the effect of LARSO on the damages award. Because we have reversed the judgment on that ground, it is unnecessary for us to address the government's challenge to specific aspects of the trial court's methodology in calculating the amount of the award.

IV

In their cross-appeal, the plaintiffs appeal from the trial court's decision not to add to the damages award certain expenses that they characterize as reasonable costs to cure or mitigate the losses caused by the taking. As noted above, Congress effectively restored the prepayment rights for participants in the HUD mortgage subsidy program in 1996 when it enacted the HOPE Act. Prior to the enactment of the HOPE Act, however, HUD gave the participants the option of entering into "use agreements," under which they would receive rent subsidies from the federal government in exchange for agreeing to be bound by certain terms of the regulatory agreements signed in connection with the original mortgages, regardless of whether they prepaid their mortgages. Independence Park, 61 Fed. Cl. at 697. Two of the plaintiffs in this case, Sherman Park Apartments and St. Andrews Gardens, entered into use agreements in 1995. When Congress subsequently enacted the HOPE Act, the restored prepayment rights were of little value to Sherman Park and St. Andrews because they were bound by the rent restrictions in the use agreements they had entered into in 1995.

The trial court calculated damages for Sherman Park and St. Andrews by determining what they lost by not being able to charge market-level rents during the period beginning on the 20-year anniversary of their original mortgages and ending on the date they entered into their use agreements. Independence Park, 61 Fed.Cl. at 702. Sherman Park and St. Andrews argued to the trial court that they were entitled to additional damages for the period following the effective date of their use agreements, based on the difference in the position they were in as a consequence of the use agreements and the position they would have been in if they had prepaid their mortgages at the 20-year mark.

The trial court rejected that argument for two reasons. First, some of the plaintiffs' arguments were based on allegations that the government had not fully performed its obligations under the use agreements. Independence Park, 61 Fed.Cl. at 700-01. To that extent, the trial court properly rejected the plaintiffs' arguments as stating a possible claim for breach of contract. To remedy a contract breach, as the court held, the plaintiffs would need to amend their complaint or file a separate action; that claim could not be remedied by increasing the award of damages for the taking. Id. at 701.

Second, the court determined that it was precluded from considering the remainder of the plaintiffs' arguments because of our reasoning in Cienega VIII, and in particular our statement that "this case involves the economic effects of ELIHPA and LIHPRHA during a period of up to eight years." Independence Park, 61 Fed.Cl. at 701 (quoting Cienega VIII, 331 F.3d at 1327); see also Cienega VIII, 331 F.3d at 1341 n. 34 ("The statutes were in effect for eight years, from 1988 . . . to 1996 . . . but the Model Plaintiffs were each affected for six years (at most) because the mortgages signed in the early 1970's would not have entered their second twenty years (and hence their prepayment periods) until the early 1990's."). The court explained that it understood the plaintiffs' arguments as seeking to expand the scope of the taking beyond the scope of the remand that we ordered in Cienega VIII.

While the trial court's interpretation of our opinion in Cienega VIII is not unreasonable, at no point in that case did we discuss the 1995 use agreements or their effect on the plaintiffs' legal rights in this case. The statements on which the trial court relied consisted mainly of characterizations of the facts as applied generally to all 42 plaintiffs. We did not focus on the specific situation of Sherman Park and St. Andrews, nor did we suggest that their entry into use agreements before the enactment of the HOPE Act had no legal effect. The mandate in Cienega VIII therefore does not foreclose the plaintiffs' cross-appeal. Accordingly, we turn to the merits of the parties' arguments regarding the legal effect of the 1995 use agreements.

The government's position is that the taking initiated by ELIHPA and LIHPRHA ended for Sherman Park and St. Andrews when their owners signed those use agreements. This is so, the government argues, because the use agreements represented a prospective settlement of any takings claims that St. Andrews and Sherman Park had against the government. The government emphasizes that the Court of Federal Claims found that Sherman Park and St. Andrews did not enter into the use agreements as a result of duress, and that the two plaintiffs signed the agreements before any court had determined that ELIHPA and LIHPRHA constituted a taking. As such, the government argues, the use agreements are fairly construed as a form of settlement agreement, under which Sherman Park and St. Andrews voluntarily chose to accept the incentives that the government offered under the use agreements instead of pursuing their claims in court and bearing the risk of being denied any compensation. We disagree with that characterization of the use agreements. As the plaintiffs point out, there is no indication in the use agreements that Sherman Park and St. Andrews were relinquishing their right to sue for damages caused by ELIHPA and LIHPRHA. Moreover, Sherman Park and St. Andrews have been parties to this litigation since 1994, and at no time has the government sought to remove them from this litigation on the grounds that they had settled their claims.

Rather than constituting a prospective settlement of the owners' claims against the government, the use agreements are better characterized as an offer by the government of something of value to offset the prepayment rights taken by statute. In 1988, the government took the plaintiffs' right to prepay their mortgages after 20 years. In 1995, the government gave the plaintiffs something of value in the form of the use agreements, which were designed to mitigate the financial impact of the taking. Once the plaintiffs had signed the use agreements, however, the passage of the HOPE Act was irrelevant to them. This situation is analogous to a physical taking in which the government appropriates a plaintiff's property at the outset and then takes steps to mitigate the financial impact of the taking, rather than returning the property to the plaintiff. In such a case, the proper analysis is to treat the initial taking as permanent and to calculate the damages for the taking by starting with the amount that the plaintiff lost as a result of the initial taking and subtracting from that sum the amount by which the plaintiff was made better off by the steps taken by the government to offset the impact of the taking. See, e.g., Shelden v. United States, 7 F.3d 1022, 1031 (Fed. Cir. 1993).

In the present case, the enactment of ELIHPA and LIHPRHA took the plaintiffs' prepayment rights, as we held in Cienega VIII, and thereby deprived the plaintiffs of any benefit that they might have realized from prepayment of their mortgages in the final 20 years of their mortgages. The use agreements were designed to reduce the financial burden imposed by the statutory prohibition against mortgage prepayment. Because one of the terms of the use agreements was that the owners would not prepay their mortgages (something that they were not entitled to do at the time they entered into the use agreements), the effect of the use agreements was to ensure that the taking of the prepayment right would remain in effect. Therefore, the calculation of damages should be adjusted in the case of Sherman Park and St. Andrews to treat the ban on prepayment as lasting as long as the use agreements provided for, with the amount of the damages adjusted to account for any benefits Sherman Park and St. Andrews obtained as a result of the use agreements.

Each party shall bear its own costs for this appeal.

REVERSED and REMANDED.