Tahoe-Sierra Preservation Council, Inc. v. Tahoe Regional Planning Agency

216 F.3d 764, 2000 Daily Journal DAR 6356, 30 Envtl. L. Rep. (Envtl. Law Inst.) 20638, 2000 Cal. Daily Op. Serv. 4765, 50 ERC (BNA) 1751, 2000 U.S. App. LEXIS 13941, 2000 WL 770512

• Court: Court of Appeals for the Ninth Circuit
• Decided: June 15, 2000
• Docket: Nos. 99-15641, 99-15771
• Status: Published

Opinion

REINHARDT, Circuit Judge:

This case involves approximately 450 plaintiffs who own property in the Lake Tahoe Basin. The lead plaintiff, Tahoe-Sierra Preservation Council, Inc. (TSPC), is an association of Tahoe-area property owners. Each individual property owner has alleged, inter alia, that each of several land-use regulations enacted in the 1980’s by the Tahoe Regional Planning Agency (TRPA) constituted a “taking” of his property under the Fifth and Fourteenth Amendments. The principal question on this appeal is whether a temporary planning moratorium, enacted by TRPA to halt development while a new regional land-use plan was being devised, effected a taking of each plaintiffs property under the standard set forth in Lucas v. South Carolina Coastal Council, 505 U.S. 1003, 112 S.Ct. 2886, 120 L.Ed.2d 798 (1992). In addition, we must determine whether any taking that may have occurred following TRPA’s adoption of a regional land-use plan in 1984 was attributable to the promulgation of that plan, in light of the fact that the plan was enjoined immediately after it was enacted and was never implemented. Finally, we must decide whether the plaintiffs’ claims regarding the successor regional land-use plan enacted by TRPA in 1987 — which did become effective — are time-barred.

FACTUAL BACKGROUND

Lake Tahoe is a large alpine lake located in the northern Sierra Nevada mountains. The lake is unique, both aesthetically and ecologically, because of its size, depth, and the astounding clarity of its water. Indeed, it is one of the clearest large lakes in the world. The unusual clarity of Lake Tahoe results from the fact that it historically was “oligotrophic” — that is, very low in nutrients and lacking a steep temperature gradient that would prevent deep circulation and mixing. Since mid-century, however, the lake has been undergoing “eutrophication,” a process by which the nutrient loading in the lake increases dramatically, due to nitrogen and phosphorus (contained in soil) being washed into the lake. The excessive enrichment of the lake by these nutrients encourages the growth of algae. As algal growth in the lake increases, the lake loses its clarity and color, becoming green and opaque. In addition to destroying the water’s visual perfection, the algae also depletes its oxygen content, thereby jeopardizing the survival of fish and other lake-dwelling animal life. In short, the eutrophication of the lake is causing serious, and effectively permanent, environmental damage.

The dramatic increase in Lake Tahoe’s nutrient levels has been caused by the rapid development of environmentally sensitive land in the Lake Tahoe Basin. The land in the basin drains into the lake, and artificial disturbances of the land — the destruction of vegetation, the creation of impervious objects such as roads and houses, etc. — greatly increase the flow of nutrients into the lake.¹ Of course, the degree to which the development of a particular parcel of land in the basin increases the nutrient flow into the lake depends on the particular characteristics of that property. In general, the development of steeper land leads to more environmental damage, because steeper land is susceptible to more rapid soil erosion. Along with steepness, other land characteristics also affect the amount of damage caused to the lake by development. For example, certain areas near streams and other wetlands, known as Stream Environment Zones (SEZs), act as filters for much of the nutrient loading that runoff carries. Disturbance of SEZ lands can lead to the rapid release of these stored nutrients into the lake. In addition, disturbance of SEZ lands may prevent them from performing their natural filtering function, thereby permitting more of the nutrients contained in runoff from higher elevations to reach the lake. Accordingly, SEZ lands are considered especially sensitive to the impact of development.

In an effort to halt the increasing rate of environmental damage to Lake Tahoe, the bi-state Tahoe Regional Planning Compact was approved in 1969 by the United States Congress after being passed by the legislatures of both Nevada and California. The Compact created the Tahoe Regional Planning Agency and set goals for the preservation of the lake and the surrounding basin. Pursuant to the Compact, TRPA adopted land-use Ordinance No. 4, which, among other things, classified the land in the basin according to its susceptibility to environmental damage. Land in the Lake Tahoe Basin was divided into seven “land capability districts,” numbered 1 through 7, with 1 being the most environmentally sensitive and 7 the least. Land capability districts 1 through 3' — consisting of the steepest land in the basin — were denominated “high hazard” or “sensitive” lands. SEZ lands were classified as a special subcategory of high hazard lands and were labeled “lb” lands. Land capability districts 4 through 7 were referred to as “low hazard” or “non-sensitive” lands.

For each land classification, Ordinance No. 4 adopted recommendations as to what degree of artificial disturbance the land could safely sustain. There were numerous exceptions to the recommendations, however, and these exceptions caused significant dissatisfaction with TRPA’s regulatory scheme. This dissatisfaction, combined with evidence that the 1969 Compact was not strong enough to remedy the problems causing the decline in the basin environment, led to the amendment of the Tahoe Regional Planning Compact in 1980.² In addition to restructuring TRPA and its voting procedures, the 1980 Compact directed TRPA (1) to adopt “environmental threshold carrying capacities” within eighteen months of the date on which the Compact became effective;³ (2) to adopt a new regional plan within twelve months of the adoption of the carrying capacities; and (3) to review all projects and establish temporary restrictions on development in the basin pending the enactment of a new regional plan.

To comply with the Compact’s requirement that it temporarily restrict development pending the enactment of a new regional plan, TRPA enacted Ordinance 81-5, which became effective on August 24, 1981. Among other things, the ordinance temporarily prohibited most residential and all commercial construction on both Class 1-8 and SEZ lands. The ordinance did contain some exceptions to the development moratorium, however, which permitted TRPA to approve construction of some single family homes on Class 1, 2, and 3 lots on the Nevada side of the basin. The ordinance stated that the provisions setting forth the moratorium “shall expire upon the adoption by the agency of Amendments to the Regional Plan.”

On August 26, 1982, TRPA adopted environmental threshold carrying capacities. The agency then proceeded with the development of a new regional plan. Due to the tremendous complexity of the task, however, it soon became clear that TRPA would be unable to adopt a new regional plan within twelve months of the adoption of the carrying capacities, as required by the Compact. Concerned that it lacked the authority to issue any building permits after this date without a new regional'plan in place, TRPA adopted Resolution 83-21. The Resolution suspended all permitting activities “pending adoption of the new regional plan.” This suspension temporarily prohibited the development of all of the covered land. Although the Resolution was drafted to expire after a ninety-day period that ended on November 26, 1983, it was extended, in accordance with its design, until the new regional plan was adopted.

On April 26, 1984, thirty-two months after it had initially suspended development, TRPA adopted a new land-use plan, the 1984 Regional Plan. See Ordinance 84-1. On the day of its adoption, the State of California sued TRPA to block the implementation of the plan on the ground that it failed to establish land-use controls sufficiently stringent to protect the Lake Tahoe Basin. The next day, the League to Save Lake Tahoe followed suit and sought an injunction against the plan on the same grounds. The United States District Court for the Eastern District of California, Judge Edward J. Garcia presiding, immediately issued a temporary restraining order prohibiting TRPA from taking any action to approve building projects, and ordered TRPA to show cause why a preliminary injunction should not issue. On June 15, 1984, Judge Garcia granted a preliminary injunction. We upheld the preliminary injunction on appeal, see California ex rel. Van de Kamp v. Tahoe Reg’l Planning Agency, 766 F.2d 1308 (9th Cir.1985), and the injunction remained in force until a completely revised land-use plan— the 1987 Regional Plan — was adopted.

PROCEDURAL HISTORY

After TRPA adopted the 1984 Plan, property owners in Nevada and California filed separate lawsuits. Plaintiffs owning property on the Nevada side of the Tahoe Basin filed an action in the United States District Court for the District of Nevada, and plaintiffs owning property on the California side of the basin filed one in the United States District Court for the Eastern District of California. All of the plaintiffs sought declaratory and injunc-tive relief, as well as damages, for various violations of the Takings Clause, the Due Process Clause, the Equal Protection Clause, and the Contracts Clause. See Tahoe-Sierra Preservation Council, Inc. v. Tahoe Reg’l Planning Agency, 911 F.2d 1331, 1333-34 (9th Cir.1990) (“TSPC I”) (Nevada-side suit) (describing the claims in detail); Tahoe-Sierra Preservation Council, Inc. v. Tahoe Reg’l Planning Agency, 938 F.2d 153, 154 n. 1 (9th Cir.1991) (“TSPC II”) (California-side suit) (same).

For purposes of litigation, the plaintiffs were divided into two groups: the first was composed of those who own land in areas classified as SEZs, and the second of those who own land in Class 1, 2, and 3 areas. In addition to the land-classification division, the plaintiffs’ claims were divided into four time periods: Period I covers August 24,1981, to August 26,1983, the time during which Ordinance 81-5 was in effect; Period II covers August 27, 1983, to April 25, 1984, the time during which Resolution 83-21 was in operation; Period III covers April 26, 1984, to July 1, 1987, the period that ran from the enactment of the 1984 Plan to the enactment of the 1987 Plan; and Period IV covers July 2, 1987, to the present, the time during which the 1987 Plan has been in effect.⁴

Period

Nevada Class 1-3 plaintiffs plaintiffs SEZ plaintiffs X⁵

California Class 1-3 plaintiffs X⁶ plaintiffs

SEZ plaintiffs X

The procedural history of the California and Nevada actions, which involves three previous Ninth Circuit opinions, see TSPC I, 911 F.2d at 1331; TSPC II, 938 F.2d at 153; Tahoe Sierra Preservation Council, Inc. v. Tahoe Reg’l Planning Agency, 34 F.3d 753 (9th Cir.1994) (“TSPC III”), and several district court opinions, need not be repeated in full. We review here only the three pieces of information regarding the procedural history that are necessary for our resolution of the present appeal.

First, it is important to point out that, at this stage in the litigation, the only claims of the plaintiffs that remain at issue are some of the § 1983 takings claims. The remainder of the claims were dismissed at one point or another. The following table shows the § 1983 takings claims that are before us (the claims are represented by X’s):

I Period II Period III Period IV

X X

X X

X X⁷ X

XXX

The TSPC III court’s treatment of the plaintiffs’ Period IV claims is the second relevant element of procedural history. Following remand from TSPC I and TSPC II, the Nevada and California cases were consolidated for further proceedings in the United States District Court for the District of Nevada, Judge Edward C. Reed presiding. The plaintiffs were permitted to amend their complaints and, in doing so, they added the Period IV claims — the claims regarding the 1987 Regional Plan which had been adopted while the litigation was pending. Shortly thereafter, the district court dismissed all of plaintiffs’ claims for Periods I, II, and IV, including the § 1983 claims, on the ground that they were all time-barred by a sixty-day statute of limitation provision contained in the 1980 Compact. See Tahoe Sierra Preservation Council, Inc. v. Tahoe Reg’l Planning Agency, 808 F.Supp. 1474 (D.Nev.1992) (Nevada case); Tahoe Sierra Preservation Council, Inc. v. Tahoe Reg’l Planning Agency, 808 F.Supp. 1484 (D.Nev.1992) (California case). On appeal, this court reversed in part. See TSPC III, 34 F.3d 753 (9th Cir.1994). The TSPC III court upheld the dismissal of some of the plaintiffs’ claims on statute of limitations grounds. See 34 F.3d at 755. The court held, however, that the sixty-day statute of limitations period contained in the 1980 Compact did not apply to the plaintiffs’ § 1983 claims. See id. at 756. Finding that the plaintiffs had “fail[ed] to plead affirmatively any other statute of limitations,” the court held that the defendants had forfeited all other statute of limitations defenses. Id.

The events of the most recent remand constitute the final element of the procedural history relevant to the present appeal. Following remand from TSPC III in 1995, the parties conducted a year-long settlement effort. Those efforts collapsed in early 1997. The district court then resolved the plaintiffs’ remaining claims in two published orders. The first order, issued in early 1998, resolved the plaintiffs’ Period IV claims (the claims regarding the 1987 Plan). In it, the district court held that, despite this court’s holding in TSPC III, the plaintiffs’ claims for Period IV were in fact time-barred by the relevant statute of limitations for § 1983 actions in Nevada and California. See Tahoe-Sierra Presentation Council, Inc. v. Tahoe Reg'l Planning Agency, 992 F.Supp. 1218 (D.Nev.1998). The second order, published in early 1999 following an eleven-day bench trial, resolved on the merits the issue of liability for the remaining claims.⁸ See Tahoe-Sierra Preservation Council, Inc. v. Tahoe Reg’l Planning Agency, 34 F.Supp.2d 1226 (D.Nev.1999). In the order, the district court held, among other things, that: (1) Ordinance 81-5 and Resolution 83-21 were facially invalid because they constituted a categorical taking of the plaintiffs’ property, and that TRPA is therefore liable under § 1983 for violating the plaintiffs’ Fifth Amendment rights during Period I and Period II; and (2) TRPA is not liable under § 1983 for any taking that may have occurred during Period III, because it was the court injunction prohibiting the implementation of the 1984 Plan that barred further development during that period and TRPA’s conduct was not the actionable cause of the injunction.

ISSUES ON APPEAL

The defendants appeal the district court’s holding that they are liable under § 1983 for a categorical taking during Period I and Period II. The plaintiffs cross-appeal the district court’s holding that the defendants are not liable under § 1983 for any taking that occurred during Period III, and the court’s holding that the plaintiffs’ § 1983 claims for Period IV are time-barred.⁹

DISCUSSION

I. TIME PERIODS I & II

The defendants contend that the district court erred in holding that Ordinance 81-5, which was in effect during Period I, and Resolution 83-21, which was in effect during Period II, constituted categorical takings of each plaintiffs property under the standard set forth in Lucas v. South Carolina Coastal Council, 505 U.S. 1003, 112 S.Ct. 2886, 120 L.Ed.2d 798 (1992). As discussed above, these laws were enacted to suspend development in the Lake Tahoe Basin pending the adoption of a new regional land-use plan.

A.

The Takings Clause of the Fifth Amendment prohibits the government from taking “private property ... for public use, without just compensation.” U.S. Const. amend. V. Courts have had little success in devising any set formula for determining when government regulation of private property amounts to a regulatory taking.¹⁰ Thus, the Supreme Court has repeatedly recognized that “whether a particular restriction [amounts to a taking] depends largely ‘upon the particular circumstances [of each] case’ ” — that is, on “essentially ad hoc, factual inquiries.” Penn Central Transportation Co. v. City of New York, 438 U.S. 104, 124, 98 S.Ct. 2646, 57 L.Ed.2d 631 (1978); see also Lucas, 505 U.S. at 1015, 112 S.Ct. 2886. In this regard, the Supreme Court and lower courts have indicated that most regulatory takings cases should be resolved by balancing the public and private interests at stake, with three primary factors weighing in the balance: (1) the economic impact of the regulation on the claimant, (2) the extent to which the regulation has interfered, with distinct investment-backed expectations, and (3) the character of the governmental action. See, e.g., Penn Central, 438 U.S. at 124, 98 S.Ct. 2646; Dodd v. Hood River County, 136 F.3d 1219, 1228 (9th Cir.1998); District Intown Properties Ltd. v. District of Columbia, 198 F.3d 874, 879 (D.C.Cir.1999).¹¹

The Court’s evaluation in various cases of the three factors identified in Penn Central illustrates that the content of these factors is not amenable to pat description. For example, it is clear that a regulation’s “economic effect upon the claimant” may be measured in several different ways. See Hodel v. Irving, 481 U.S. 704, 714, 107 S.Ct. 2076, 95 L.Ed.2d 668 (1987) (looking to the market value of the property); Keystone Bituminous Coal Ass’n v. DeBenedictis, 480 U.S. 470, 493-96, 107 S.Ct. 1232, 94 L.Ed.2d 472 (1987) (looking to whether the regulation makes the property owner’s business operation “commercially impracticable”); Andrus v. Allard, 444 U.S. 51, 66, 100 S.Ct. 318, 62 L.Ed.2d 210 (1979) (looking to the possibility of other economic uses besides sale, which was prohibited by the challenged regulation). Similarly, the “character of the governmental action” depends on several things, including whether the action is properly characterized as a physical occupation of the property, see Loretto v. Teleprompter Manhattan CATV Corp., 458 U.S. 419, 434-35, 102 S.Ct. 3164, 73 L.Ed.2d 868 (1982), and whether the regulation advances a legitimate public purpose, see Keystone Bituminous Coal Ass’n, 480 U.S. at 485-6, 107 S.Ct. 1232; Agins v. Tiburon, 447 U.S. 255, 260, 100 S.Ct. 2138, 65 L.Ed.2d 106 (1980).

The Supreme Court has identified two specific circumstances in which it will find a government regulation to constitute a “categorical” taking without performing an ad hoc balancing under Penn Central. The first situation involves regulations that compel a property owner to suffer a permanent physical “invasion” or “occupation” of his property. See, e.g., Loretto v. Teleprompter Manhattan CATV Corp., 458 U.S. 419, 426, 102 S.Ct. 3164, 73 L.Ed.2d 868 (1982). In that circumstance, the “character of the governmental action”— ordinarily one of the factors balanced under Penn Central — itself becomes sufficient to effect a taking. As the Court held in Loretto:

[A] “taking” may more readily be found when the interference with property can be characterized as a physical invasion by government .... [w]hen the physical intrusion reaches the extreme form of a permanent physical occupation, a taking has occurred. In such a case, “the character of the government action” not only is an important factor in resolving whether the action works a taking but also is determinative.

Loretto, 458 U.S. at 426, 102 S.Ct. 3164. The categorical rule applicable to physical invasion cases is clearly not applicable to the present case.

“The second situation in which [the Court] ha[s] found categorical treatment appropriate is where regulation denies all economically beneficial or productive use of land.” Lucas, 505 U.S. at 1015, 112 S.Ct. 2886; see also Agins, 447 U.S. at 260, 100 S.Ct. 2138. Here, too, ad hoc balancing is not required because one factor of the Penn Central test becomes dispositive. In this instance, it is the “economic impact of the regulation” that becomes determinative: the regulation constitutes a “categorical” taking when it deprives the claimant of “all economically beneficial and productive use of [his] land.” Lucas, 505 U.S. at 1015, 112 S.Ct. 2886. Cf. Dodd, 136 F.3d at 1228.¹²

With respect to Periods I and II, the only question before us is whether the rule set forth in Lucas applies- — that is, whether a categorical taking occurred because Ordinance 81-5 and Resolution 83-21 denied the plaintiffs “all economically beneficial or productive use of land.” Lucas, 505 U.S. at 1015, 112 S.Ct. 2886. Below, the district court ruled that the regulations did not constitute a taking under Penn Central ’s ad hoc approach, but that they did constitute a categorical taking under Lucas. See Tahoe-Sierra Preservation Council, 34 F.Supp.2d at 1238-45. The defendants appealed the district court’s latter holding, but the plaintiffs did not appeal the former. And even if arguments regarding the Penn Central test were fairly encompassed by the defendants’ appeal, the plaintiffs have stated explicitly on this appeal that they do not argue that the regulations constitute a taking under the ad hoc balancing approach described in Penn Central. More generally, the plaintiffs do not contest that California’s and Nevada’s objective of preserving the environmental health and aesthetic beauty of Lake Tahoe is an entirely permissible governmental goal. Nor do they dispute that the restrictions imposed on their properties are appropriate means of securing the purpose set forth in the Compact. In view of the limited nature of the plaintiffs’ arguments on appeal, we address our specific inquiry to whether the facts of this case present one of the “relatively rare situations” where “regulation denies all economically beneficial or productive use of land.” Lucas, 505 U.S. at 1018, 112 S.Ct. 2886.

Our focus is also narrowed by the fact that the plaintiffs bring only a facial challenge to Ordinance 81-5 and Resolution 83-21. In facial takings claims, our inquiry is limited to “whether the mere enactment of the [regulation] constitutes a taking.” Agins, 447 U.S. at 260, 100 S.Ct. 2138; see also Keystone Bituminous Coal Ass’n, 480 U.S. at 493, 107 S.Ct. 1232; Hodel v. Virginia Surface Mining & Reclamation Ass’n, 452 U.S. 264, 295, 101 S.Ct. 2352, 69 L.Ed.2d 1 (1981); Garneau v. City of Seattle, 147 F.3d 802, 807 (9th Cir.1998). For that reason, we look only to the regulation’s “‘general scope and dominant features,’ ” rather than to the effect of the application of the regulation in specific circumstances. Garneau, 147 F.3d at 807 (quoting Village of Euclid v. Ambler Realty Co., 272 U.S. 365, 397, 47 S.Ct. 114, 71 L.Ed. 303 (1926)); see also Agins, 447 U.S. at 260, 100 S.Ct. 2138; Keystone Bituminous Coal Ass’n, 480 U.S. at 493, 494, 107 S.Ct. 1232; Virginia Surface Mining & Reclamation Ass’n, 452 U.S. at 297, 101 S.Ct. 2352. In this connection, “since it is difficult to demonstrate that [the] ‘mere enactment’ of a piece of legislation” amounts to a taking, the Court has recognized that facial takings challenges “face an uphill battle.” Suitum v. Tahoe Reg’l Planning Agency, 520 U.S. 725, 736 n. 10, 117 S.Ct. 1659, 137 L.Ed.2d 980 (1997); see also Keystone Bituminous Coal Ass’n, 480 U.S. at 495, 107 S.Ct. 1232; Garneau, 147 F.3d at 807.

B.

The plaintiffs contend that, for purposes of determining whether the regulations constitute a categorical taking under Lucas, we should not treat the plaintiffs’ properties as the fee interests that they are. Instead, they argue, we should define narrowly, as a separate property interest, the temporal “slice” of each fee that covers the time span during which Ordinance 81-5 and Resolution 83-21 were in effect. It is this carved-out piece of each plaintiffs property interest, the plaintiffs assert, that has been “taken” by the regulations.

“Because our test for regulatory taking requires us to compare the value that has been taken from the property with the value that remains in the property, one of the critical questions is determining how to define the unit of property ‘whose value is to furnish the denominator of the fraction.’ ”¹³ Keystone Bituminous Coal Ass’n, 480 U.S. at 497, 107 S.Ct. 1232 (quoting Frank I. Michelman, Property, Utility, and Fairness: Comments on the Ethical Foundations of “Just Compensation” Law, 80 Harv. L.Rev. 1165, 1192 (1967)). In other words, for purposes of determining whether a “taking” of the plaintiffs’ “property” has occurred, the proper inquiry is what constitutes the relevant “property”? Is it the fee interest that must be “taken,” or is it some lesser unit of property? Property interests may have many different dimensions. For example, the dimensions of a property interest may include a physical dimension (which describes the size and shape of the property in question), a functional dimension (which describes the extent to which an owner may use or dispose of the property in question), and a temporal dimension (which describes the duration of the property interest). At base, the plaintiffs’ argument is that we should conceptually sever each plaintiffs fee interest into discrete segments in at least one of these dimensions — the temporal one — and treat each of those segments as separate and distinct property interests for purposes of takings analysis. Under this theory, they argue that there was a categorical taking of one of those temporal segments.

While Supreme Court precedent has not over the years been entirely uniform in its treatment of the conceptual severance question, compare Pennsylvania Coal Co. v. Mahon, 260 U.S. 393, 43 S.Ct. 158, 67 L.Ed. 322 (1922) (employing conceptual severance) with Keystone Bituminous Coal Ass’n v. DeBenedictis, 480 U.S. 470, 107 S.Ct. 1232, 94 L.Ed.2d 472 (1987) (rejecting conceptual severance in the identical context), most modern case law rejects the invitation of property holders to engage in conceptual severance, except in cases of physical invasion or occupation. Several cases illustrate the Court’s refusal to employ this concept in other types of circumstances. In Penn Central Transportation Company v. City of New York, the Penn Central Transportation Company entered into a contract for the construction and lease of an office building above its Grand Central Terminal in New York City. 438 U.S. at 116, 98 S.Ct. 2646. When the City denied two alternate building plans on the ground that they would destroy the architectural appeal of the historic landmark, Penn Central filed suit, claiming that the rejection of the building plans constituted a taking. Id. at 117-18, 98 S.Ct. 2646. In affirming the denial of its takings claim, the Court explicitly rejected Penn Central’s proposal to consider the airspace above the Terminal as a property interest separate from the rest of the Terminal site. The Court explained:

“Taking” jurisprudence does not divide a single parcel into discrete segments and attempt to determine whether rights in a particular segment have been entirely abrogated. In deciding whether a particular governmental action has effected a taking, this Court focuses rather both on the character of the action and on the nature and extent of the interference with rights in the parcel as a whole-here, the city tax block designated as the “landmark site.”

Id. (emphasis added); see also MacLeod v. County of Santa Clara, 749 F.2d 541, 547 (9th Cir.1984).

The Court also refused to employ conceptual severance in Keystone Bituminous Coal Association v. DeBenedictis, which considered the effect of Pennsylvania’s Bituminous Mine Subsidence and Land Conservation Act on the property rights of mining companies. As implemented, the Act generally required 50% of the coal beneath certain protected structures to be kept in place as a means of providing surface support. Id. at 476-77, 107 S.Ct. 1232. The petitioners, who had purchased both mining rights and waivers for any surface damage caused by mining, argued that the Act constituted a taking. Id. at 478-79, 107 S.Ct. 1232. In particular, the petitioners argued that the Act appropriated the portion of coal that they were required to leave in the ground. Id. at 497, 107 S.Ct. 1232. They also argued that the Act entirely destroyed the value of each petitioner’s “support estate,” which is recognized under Pennsylvania law as a separate interest in land. Id. In essence, the petitioners argued that there had been a categorical taking of these two distinct property interests.

In holding that the regulation of the petitioners’ mining rights did not amount to a taking, the Supreme Court refused to consider the coal that the Act required the petitioners to leave in place as a separate property interest; rather, the Court emphasized, takings jurisprudence must consider the “parcel as a whole.” Id. at 497-99, 107 S.Ct. 1232. In addition, the Court described Pennsylvania’s recognition of a support estate as a “legalistic distinction[ ]” that could not convert the support estate into a separate property interest for takings law purposes. Id. at 500-02, 107 S.Ct. 1232.

The Court’s general rule against conceptual severance is not limited to the spatial dimension of property rights. In Andrus v. Allard, the Court applied its general rule in a more functional dimension, to the “bundle” of rights that make up what we think of as “property.” In that case, federal regulation prohibited the sale of eagle feathers. Id. at 53-54, 100 S.Ct. 318. The plaintiffs argued that the denial of the right to sell their feathers constituted a taking. Id. at 64 & n. 21, 100 S.Ct. 318. The Court rejected the takings claim, and, in so doing, rejected the conceptual disaggregation of property rights: “[Wjhere an owner possesses a full ‘bundle’ of property rights, the destruction of one ‘strand’ of the bundle is not a taking because the aggregate must be viewed in its entirety.” Andrus, 444 U.S. at 65-66, 100 S.Ct. 318 (citing Penn Central, 488 U.S. at 130-31, 98 S.Ct. 2646).¹⁴

In fact, the Supreme Court has already once rejected conceptual severance in the temporal dimension of property rights. In Agins v. Tiburon, the city of Tiburon had instituted condemnation proceedings against the plaintiffs’ property, but abandoned the proceedings a year later. The plaintiffs brought suit, contending, inter alia, “that the city’s aborted attempt to acquire the land through eminent domain had destroyed the use of the land during the pendency of the condemnation proceedings.” Agins, 447 U.S. at 258 n. 3, 100 S.Ct. 2138. The Supreme Court rejected the plaintiffs’ claim, holding:

The State Supreme Court correctly rejected the contention that the municipality’s good faith planning activities, which did not result in the successful prosecution of an eminent domain claim, so burdened the appellants’ enjoyment of their property so as to constitute a taking. Even if the appellants’ ability to sell their property was limited during the pendency of the condemnation proceeding, the appellants were free to sell or develop their property when the proceedings ended. Mere fluctuations in value during the process of governmental decisionmaking, absent extraordinary delay, are “incidents of ownership. They cannot be considered as a ‘taking’ in the constitutional sense.”

447 U.S. at 263 n. 9, 100 S.Ct. 2138 (citations omitted) (emphasis added) (quoting Danforth v. United States, 308 U.S. 271, 285, 60 S.Ct. 231, 84 L.Ed. 240 (1939)). In rejecting the takings claim, the Court relied only on the fact that the plaintiffs were able to sell or develop their property after the city abandoned its condemnation claim. By relying on the temporary nature of the restriction, the Court rejected the invitation to carve out, as a separate property interest, a temporal “slice” of the parcel that existed for the time period during which the condemnation proceedings were in progress. For, had the Court considered the plaintiffs’ rights in their property during that time period as a separate interest, the plaintiffs’ ability to sell or develop their property after the time period ended would have been irrelevant to the Court’s takings analysis. Only the ability to sell or develop the property during the condemnation period would have mattered.

Agins’s rejection of conceptual severance in the temporal dimension is consistent with the Court’s rejection of other forms of conceptual severance in Penn Central and Andrus. It would make little sense to accept temporal severance and reject spatial or functional severance. A planning regulation that prevents the development of a parcel for a temporary period of time is conceptually no different than a land-use restriction that permanently denies all use on a discrete portion of property, or that permanently restricts a type of use across all of the parcel. See Margaret Jane Radin, The Liberal Conception of Property: Cross Currents in the Jurisprudence of Takings, 88 Co-lum.L.Rev. 1667, 1674-78 (1988). Each of these three types of regulation will have an impact on the parcel’s value, because each will affect an aspect of the owner’s “use” of the property — by restricting when the “use” may occur, where the “use” may occur, or how the “use” may occur. Prior to Agins, the Court had already rejected takings challenges to regulations eliminating all “use” on a portion of the property, and to regulations restricting the type of “use” across the breadth of the property. See Penn Central, 438 U.S. at 130-31, 98 S.Ct. 2646; Keystone Bituminous Coal Ass’n, 480 U.S. at 498-99, 107 S.Ct. 1232; Village of Euclid v. Ambler Realty Co., 272 U.S. 365, 384, 397, 47 S.Ct. 114, 71 L.Ed. 303 (1926) (75% diminution in value caused by zoning law); see also William C. Haas & Co. v. City & County of San Francisco, 605 F.2d 1117, 1120 (9th Cir.1979) (value reduced from $2,000,000 to $100,000). In those cases, the Court “uniformly rejected] the proposition that diminution in property value, standing alone, can establish a ‘taking.’ ” Penn Central, 438 U.S. at 131, 98 S.Ct. 2646; see also Concrete Pipe and Products, Inc. v. Construction Laborers Pension Trust, 508 , 113 S.Ct. 2264, 124 L.Ed.2d 539 (1993). There is no plausible basis on which to distinguish a similar diminution in value that results from a temporary suspension of development.

To not reject the concept of temporal severance, we would risk converting every temporary planning moratorium into a categorical taking. See Penn Central, 438 U.S. at 130, 98 S.Ct. 2646; Stern v. Halligan, 158 F.3d 729, 734 (3rd Cir.1998); Michelman, supra, at 1193. Such a result would run contrary to the Court’s explanation that it is “relatively rare” that government “regulation denies all economically beneficial or productive use of land.” Lucas, 505 U.S. at 1015, 112 S.Ct. 2886.

More important, the widespread invalidation of temporary planning moratoria would deprive state and local governments of an important land-use planning tool with a well-established tradition. Land-use planning is necessarily a complex, time-consuming undertaking for a community, especially in a situation as unique as this. In several ways, temporary development moratoria promote effective planning. First, by preserving the status quo during the planning process, temporary moratoria ensure that a community’s problems are not exacerbated during the time it takes to formulate a regulatory scheme. See Elizabeth A. Garvin & Martin L. Leitner, Drafting Interim Development Ordinances: Creating Time to Plan, Land Use Law and Zoning Digest, June 1996, at 3, 3; Schafer v. City of New Orleans, 743 F.2d 1086, 1090 (5th Cir.1984). Relatedly, temporary development moratoria prevent developers and landowners from racing to carry out development that is destructive of the community’s interests before a new plan goes into effect. Such a race-to-development would permit property owners to evade the land-use plan and undermine its goals. See id.; Miller v. Board of Public Works, 195 Cal. 477, 234 P. 381, 388 (1925).¹⁵ Finally, the breathing room provided by temporary moratoria helps ensure that the planning process is responsive to the property owners and citizens who will be affected by the resulting land-use regulations. See Robert H. Freilich, Interim Development Controls: Essential Tools for Implementing Flexible Planning and Zoning, 49 J. Urb. Law 65, 79 (1971). Absent the pressure of trying to out-speed developers who are attempting to circumvent the planning goals, the “planning and implementation process may be permitted to run its full and natural course with widespread citizen input and involvement, public debate, and full consideration of all issues and points of view.” Garvin and Leitner, supra, at 3. Given the importance and long-standing use temporary morato-ria, courts should be exceedingly reluctant to adopt rulings that would threaten the survival of this crucial planning mechanism.

In opposition to the overwhelming legal and logical support for not conceptually severing fee interests into small temporal pieces, the plaintiffs argue (and the district court below decided) that the Court’s decision in First English Evangelical Lutheran Church v. County of Los Angeles, 482 U.S. 304, 107 S.Ct. 2378, 96 L.Ed.2d 250 (1987), compels such conceptual severance of the property rights in this case. The plaintiffs contend that First English holds that conceptual severance of the temporal dimension of property interests is generally required. This argument is flatly incorrect.

First English is not even a case about what constitutes a taking. In First English, property owners challenged an ordinance that prevented the development of property located on a flood plain; they sought “damages for the uncompensated taking of all use” of the property. See id. at 309, 107 S.Ct. 2378. The California Court of Appeal rejected their claims on the ground that, regardless of whether a taking occurred, the claimants could not recover damages during the period running from the time of enactment of the ordinance to the time when it was finally declared unconstitutional. See id. Because the question presented to the Supreme Court related only to the remedy available once a taking had been proven, the Court stated explicitly that it was not addressing whether the ordinance constituted a taking. See id, at 313, 107 S.Ct. 2378 (“We accordingly have no occasion to decide whether the ordinance at issue actually denied appellant all use of its property. ... ”). According to the Court, it was resolving only the question whether, once a taking is established, “abandonment [of that taking] by the government requires payment of compensation for the period of time during which regulations” that constitute a taking are in effect. Id. at 318, 107 S.Ct. 2378; see also id. at 311, 107 S.Ct. 2378 (noting that the “disposition of the case [by the California court on the assumption that a taking occurred] isolates the remedial question for our consideration”); Michelman, supra, at 1617 & n. 81.

It is true that First English holds that, when a taking has occurred, the government must compensate property owners, even if the taking is “temporary.” Contrary to the plaintiffs’ suggestion, however, the Court’s holding in First English was not that temporary moratoria are “temporary takings.” In fact, the opposite is true. The First English Court very carefully defined “ ‘temporary’ regulatory takings [as] those regulatory takings which are ultimately invalidated by the courts.” ¹⁶ 482 U.S. at 310, 107 S.Ct. 2378. What is “temporary,” according to the Court’s definition, is not the regulation; rather, what is “temporary” is the taking, which is rendered temporary only when an ordinance that effects a taking is struck down by a court. In other words, a permanent regulation leads to a “temporary ” taking when a court invalidates the ordinance after the taking. See id. at 319, 107 S.Ct. 2378 (“Invalidation of the ordinance or its successor ordinance after this period of time, though converting the taking into a ‘temporary’ one, is not a sufficient remedy to meet the demands of the Just Compensation Clause.”); id. at 317, 107 S.Ct. 2378 (discussing the fact “that the government may elect to abandon its intrusion or discontinue regulations,” and thereby turn what would otherwise be a “permanent taking” into a “ ‘temporary’ taking”). The Court’s definition, therefore, does not comprehend temporary moratoria, which from the outset are designed to last for only a limited period of time. In short, we reject the plaintiffs’ contentions that First English applies to temporary moratoria and that it works a radical change to takings law by requiring that property interests be carved up into finite temporal segments.¹⁷

In addition to misinterpreting First English, the district court erred in concluding that United States v. General Motors Corp., 323 U.S. 373, 65 S.Ct. 357, 89 L.Ed. 311 (1945), and United States v. Petty Motor Co., 327 U.S. 372, 66 S.Ct. 596, 90 L.Ed. 729 (1946), support conceptual severance in the present case. In these cases, the federal government used its power of eminent domain to condemn leaseholds during World War II. Both cases were concerned with the level of compensation owed to the ousted lessees. The fact that just compensation was required in these cases, however, has no bearing on the question before us. Both cases involved the physical occupation by the federal government of the property at issue and thus fall within the categorical rule established by the Court for cases involving physical invasions. As we have explained, physical occupations and appropriations have always received markedly different analytic treatment than other regulatory takings. The Supreme Court has clearly held that the physical invasion or possession by the government of a fragment of a larger parcel constitutes a per se taking — in other words, conceptual severance is the norm. See, e.g., Loretto v. Teleprompter Manhattan CATV Corp., 458 U.S. 419, 426, 102 S.Ct. 3164, 73 L.Ed.2d 868 (1982). Just as General Motors and Petty Motor require compensation for the government’s condemnation of a leasehold, Loretto requires compensation for the permanent physical appropriation of a tiny fragment of an apartment building. See Loretto, 458 U.S. at 426, 102 S.Ct. 3164. The conceptual severance that is the norm in these physical appropriation cases, however, does not extend to other regulatory takings claims.¹⁸

In short, we reject the plaintiffs’ suggestion that we engage in conceptual severance. The relevant property interests in the present case are the whole parcels of property that the plaintiffs own.

C.

Having determined that the property interest at stake is just what one would expect it to be — the plaintiffs’ fee interests — we must evaluate whether Ordinance 81-5 and Resolution 83-21 effected a categorical taking of each plaintiffs property.¹⁹ For purposes of this analysis, two features of these provisions are relevant. First, the provisions effectively placed a moratorium on the development of the plaintiffs’ property.²⁰ The second relevant feature of the provisions is that the moratorium they effected was intended to be temporary — the regulations were designed to institute a temporary moratorium that would remain in effect only until a new regional land-use plan could be adopted.²¹ See Union Oil Co. v. Morton, 512 F.2d 743, 751 (9th Cir.1975) (describing a regulation as temporary if its “termination is conditioned [on] the occurrence of certain future events”).

To determine whether the temporary moratorium instituted by TRPA’s regulations denies “all economically beneficial or productive use” of the plaintiffs’ land, we must first consider the meaning of the phrase “economically beneficial or productive use.” The phrase’s precise meaning is elusive, and has not been clarified by the Supreme Court. See, e.g., Lake Nacimiento Ranch Co. v. County of San Luis Obispo, 841 F.2d 872, 877 (9th Cir.1988). The central confusion over its meaning centers on the relationship between the “use” of property and its “value.”²² Clearly, the economic value of property provides strong evidence of the availability of “economically beneficial or productive uses” of that property.²³ Nevertheless, there are instances in which certain kinds of “value” may be poor measures of the existence of such uses. In any event, we need not resolve the sticky issues surrounding the meaning and proof of the existence of “economically beneficial or productive uses,” because it is clear from the “general scope and dominant features” of Ordinance 81-5 and Resolution 83-21 that the temporary moratorium imposed by these regulations did not deprive the plaintiffs’ land in the Lake Tahoe Basin of either all of its “value” or all of its “use.”²⁴ Garneau, 147 F.3d at 807.

First, as amici Cities and Counties of California note, basic principles of economics show that the moratorium did not render the plaintiffs’ property valueless.²⁵ See generally Lucas, 505 U.S. at 1020, 112 S.Ct. 2886 (assuming, as the basis for its decision, that the regulation at issue rendered Lucas’s two beachfront lots “valueless”). The moratorium was temporary— it was designed to and did dissolve upon the adoption of a new regional plan. Given that the ordinance and resolution banned development for only a limited period, these regulations preserved the bulk of the future developmental use of the property. This future use had a substantial present value.²⁶

Of course, were a temporary moratorium designed to be in force so long as to eliminate all present value of a property’s future use, we might be compelled to conclude that a categorical taking had occurred. We doubt, however, that a true temporary moratorium would ever be designed to last for so long a period. Certainly, the moratorium at issue here was not. The temporary moratorium was designed to suspend development only until a new regional land-use plan could be formulated — a process that the 1980 Compact intended would take thirty months. While the completion of the regional plan actually took forty months (which led to the temporary moratorium remaining in effect for eight months longer than expected), the moratorium still was in effect for only thirty-two months.²⁷

Moreover, there is no evidence that owners or purchasers of property in the basin anticipated that the temporary moratorium would continue indefinitely. Nor would they have had reason to: the district court found that TRPA worked diligently to complete the regional plan as quickly as possible. See Tahoe-Sierra Preservation Council, 34 F.Supp.2d at 1250-51. Thus, while the temporary moratorium surely had a negative impact on property values in the basin, we cannot conclude that the interim suspension of development wiped out the value of the plaintiffs’ properties.

Furthermore, the temporary moratorium did not deprive the plaintiffs of all “use” of their property. The “use” of the plaintiffs’ property runs from the present to the future. (This is a simple corollary of our earlier conclusion that the plaintiffs’ property interests may not be temporally severed.)²⁸ By instituting a temporary development moratorium, TRPA denied the plaintiffs only a small portion of this future stream; the thirty-two months during which the moratorium was in effect represents a small fraction of the useful life of the Tahoe properties.²⁹

Because the temporary development moratorium enacted by TRPA did not deprive the plaintiffs of all of the value or use of their property, we hold that it did not effect a categorical taking.³⁰ Indeed, given the above analysis, it is equally clear that the district court was correct to conclude that the moratorium did not constitute a taking under the Penn Central test. See Tahoe-Sierra Preservation Council, 34 F.Supp.2d at 1240-42. Thus, while the district court was correct as to this latter point, we reverse its holding that a categorical taking occurred. In reaching this conclusion, we preserve the ability of local governments to do what they have done for many years — to engage in orderly, reasonable land-use planning through a considered and deliberative process. To do otherwise would turn the Takings Clause into a weapon to be used indiscriminately to penalize local communities for attempting to protect the public interest.

II. TIME PERIOD III

The plaintiffs raise two issues on cross-appeal, the first of which concerns the district court’s holding regarding Period III — the period after the passage of the 1984 Plan and before the passage of the 1987 Plan.³¹ The district court held that TRPA’s adoption of the 1984 Plan was not the actionable cause of any deprivation of the plaintiffs’ Fifth and Fourteenth Amendment rights that occurred during this period. According to the court, the injunction issued by Judge Garcia against TRPA, in a separate suit brought by the state of California and the League to Save Lake Tahoe, prevented the adoption of the 1984 Plan from being either the “but for” or the proximate cause of any taking that may have occurred during Period III. For this reason, the court held that, as to that period, TRPA could not be held liable under § 1983.³² On cross-appeal, the plaintiffs argue that the district court erred in concluding that TRPA’s actions were hot the legal cause of any taking during the period in question.

In a section 1983 action, the plaintiff must demonstrate that the defendant’s conduct was the actionable cause of the claimed injury. See, e.g., Arnold v. IBM Corp., 637 F.2d 1350, 1355 (9th Cir.1981). To meet this causation requirement, the plaintiff must establish both causation-in-fact and proximate causation. See Van Ort v. Estate of Stanewich, 92 F.3d 831, 837 (9th Cir.1996); Arnold, 637 F.2d at 1355; Hoffman v. Halden, 268 F.2d 280, 295 (9th Cir.1959), overruled in part on other grounds by Cohen v. Norris, 300 F.2d 24 (9th Cir.1962). The parties agree that this requirement applies to the plaintiffs’ § 1983 claims, which allege a violation of the Takings Clause, as much as it would to any other § 1983 claim.³³ Like other factual determinations, causation-in-fact is reviewed for clear error. In addition, we review findings of proximate cause for clear error, even though they present mixed questions of law and fact. See Exxon Co. v. Sofec, Inc., 54 F.3d 570, 576 (9th Cir.1995) (holding that, although it “is an exception to the general rule that mixed questions of law and fact are reviewed de novo;” issues of proximate cause are reviewed under the clearly erroneous standard); George v. City of Long Beach, 973 F.2d 706, 709 (9th Cir.1992); Britton v. Price, 950 F.2d 602, 604 (9th Cir.1991).³⁴

The plaintiffs argue on two grounds that the district court clearly erred when it concluded that TRPA’s conduct was not the actionable cause of the claimed taking of the plaintiffs’ property from 1984 to 1987. First, they argue that the injunction is irrelevant to any eonsider-ation of “cause,” because the 1984 Plan was actually implemented and itself constituted a regulatory taking.³⁵ We disagree. Contrary to the plaintiffs’ contention, the injunction issued by Judge Garcia effectively prohibited the implementation of the 1984 Plan.³⁶ The order that granted the injunction read in pertinent part:

That, until conclusion of trial of this matter, the defendant TRPA, its agents, servants, employees, and all other persons acting under the authority of and in concert with TRPA, hereby are enjoined and restrained from taking any action to approve any project, as defined in to Tahoe Regional Planning Compact, Public Law 96-551, 94 Stat. 3233 (1980), or to approve the construction of any man-made development within the agency’s jurisdiction, including the any of the 2,343 developments referred to in the Court’s June 15, 1984 opinion, and including the acceptance of applications for such agency approval....

As the district court’s order makes clear, the injunction prohibited TRPA from taking “any action” to approve any project, and even prohibited it from accepting permit applications. Without the ability to accept, process, or grant applications, it is impossible to see how TRPA could have implemented the 1984 Plan, the purpose of which was to regulate the granting of new permits.³⁷ Consequently, as this court pointed out in TSPC II, “th[e] plan never went into effect.”³⁸ 938 F.2d at 155. Thus, the 1984 Plan itself could not have constituted a taking.

In the alternative, the plaintiffs argue that, even if the 1984 Plan was not itself implemented, TRPA’s adoption of the plan effectuated a taking by causing the injunction to issue. Specifically, they contend that (1) the adoption of the 1984 Plan was both a “but for” and the proximate cause of the issuance of the injunction, and (2) the injunction constituted a taking of the plaintiffs’ property.³⁹ For at least two reasons, we reject the plaintiffs’ alternative argument.

First, the district court did not clearly err in holding that TRPA reasonably did not foresee that the 1984 Plan would be enjoined. It is well-established that foreseeability analysis is an appropriate part of proximate cause determinations in § 1983 actions. See, e.g., Arnold, 637 F.2d at 1355; Johnson, 588 F.2d at 743-44. Cf. Van Ort, 92 F.3d at 837 (noting that unforeseen, intervening causes break the chain of proximate causation in § 1983 actions). See generally Palsgraf v. Long Island R. Co., 248 N.Y. 339, 162 N.E. 99 (1928). Employing such an analysis, the plaintiffs argue that TRPA is legally responsible for the effects of the injunction because TRPA should have reasonably foreseen that it would be sued as a result of its adopting the 1984 Plan. Even if a lawsuit was foreseeable, however, it is the foreseeability of the injunction, not the foreseeability of a lawsuit, that is relevant to proximate causation here. The district court concluded that TRPA reasonably did not foresee that an injunction would issue. See Tahoe-Sierra Preservation Council, 34 F.Supp.2d at 1248. The record supports this determination: TRPA sought to adopt a plan that would comply with the Compact, and it had been advised by counsel that its legislative discretion to adopt a plan would be given great deference by the courts; accordingly, it had no reason to believe that it would lose a lawsuit or that the 1984 Plan would be enjoined. Rather than contest this fact, the plaintiffs make the bald assertion that TRPA secretly wanted an injunction against all construction in the basin. However, they point to no evidence in the record to support this contention.

Although TRPA could not have foreseen, prior to adopting the 1984 Plan, that an injunction would issue, the fact that the district court did issue an injunction reveals a second reason why TRPA may not be held legally responsible for any injury that occurred during Period III: in retrospect, there was little TRPA could have done to prevent the injunction. The reason is that “the ‘wrongdoing’ plaintiffs claim [that TRPA committed] was not the ‘wrongdoing’ that triggered the injunction.” TSPC I, 911 F.2d at 1434 (B. Fletcher, J., concurring). The wrongdoing about which the plaintiffs complain was TRPA’s adoption, in 1984, of a regional plan that was overly restrictive of development. As Judge Fletcher noted in TSPC I, however, “the court granted the injunction based on California’s and the League’s claim that the Plan was not strict enough: it would permit construction of single-family residence in excess of compact limits and inflict environmental damage, and the Plan also did not abide by the Compact’s stricter procedural requirements for project approval.” Id. (emphasis added). Thus, had TRPA avoided its alleged wrongdoing — by adopting a more lenient plan that met with the plaintiffs’ approval — it is even more certain that an injunction would have issued. In short, the lack of a causal connection between the alleged wrongdoing and the purported harm compels the conclusion that TRPA may not be held liable for the effects of the injunction.

For the foregoing reasons, we affirm the district court’s holding that TRPA’s actions were not the actionable cause of any taking that may have occurred during Period III, and that the defendants therefore may not be held liable as to that period under § 1983.

III. TIME PERIOD IV

The second issue that the plaintiffs raise on cross-appeal concerns the district court’s resolution of the claims regarding the 1987 Plan (the Period IV claims). The plaintiffs first added these claims when they amended their complaints in 1991. At that time, the defendants moved to dismiss the plaintiffs’ claims for all four time periods on the ground that the claims were barred by the sixty-day statute of limitations included in the 1980 Compact.⁴⁰

The district court agreed as to the claims regarding Periods I, II, and IV, and dismissed those claims.⁴¹ On appeal, we reversed as to the § 1983 claims. We held:

TSPC ... brought one claim under 42 U.S.C. § 1983 for the violation of its civil rights by the actions of TRPA in adopting the 1981 Ordinance, the 1983 Resolution, the 1984 Plan and the 1987 Plan.... As to § 1983 it is established law that there is a single statute of limitation to be applied.... Obviously, the single state statute of limitations to be applied in all § 1983 actions cannot be the special 60-day period provided by the Compact. The defendants did not plead any other statute of limitations except the 60-day one. Failing to plead affirmatively any other statute of limitations, they cannot now rely on any other. No part of TSPC’s § 1983 claim is time-barred.

TSPC III, 34 F.3d at 756, as amended by 42 F.3d 1306 (citations omitted).

On remand, the defendants asserted in their answers that the plaintiffs’ § 1983 claims regarding Period IV, while not barred by the sixty-day statute of limitations, were barred by the applicable statutes of limitations — a one-year limit for § 1983 actions in California, and a two-year limit for § 1983 actions in Nevada.⁴² See Tahoe-Sierra Preservation Council, 992 F.Supp. at 1221. The district court agreed on the merits. In addition, the court held that the law of the case doctrine did not preclude it from holding the plaintiffs’ Period IV claims time-barred, even though the TSPC III court had concluded otherwise. The district court concluded that it was not compelled to follow TSPC III because we had clearly erred in that case when we held that the plaintiffs had forfeited the right to rely on the correct statute of limitations period by failing to “plead” that statute “affirmatively.” Accordingly, it dismissed the Period IV § 1983 claims. On cross-appeal, the plaintiffs argue that the district court had no authority to deviate from the law of the case set forth in TSPC III. Rather, they contend, only another panel of this court may decline to follow the law of the case established in a disposition of this court.

As an initial matter, we note that the plaintiffs’ argument is largely academic. Even were the district court bound absolutely by the holding in TSPC III, we would still have discretion, under the circumstances present here, to decline to follow the law of that case. In addition to being academic, however, the plaintiffs’ argument may well be incorrect. In United States v. Cuddy, 147 F.3d 1111 (9th Cir.1998), we held that the district court did not abuse its discretion when it declined to follow, on remand, a holding we had issued on appeal. See id. at 1114-15. We concluded that the district court had acted within its discretion because it was correct in determining that our holding was clearly erroneous. See id. In addition to our holding in Cuddy, we have suggested in dicta on several occasions that district courts may, in certain circumstances, decline to follow law of the case set forth by this court. In fact, one of our standard formulations of the law of the case doctrine includes subordinate, as well as coordinate, courts: “Under the doctrine of law of the case, ‘a court is generally precluded from reconsidering an issue that has already been decided by the same court, or a higher cowt in the identical case.’ [However,] a court may have discretion to depart from the law of the case where [certain features are present.]” Rebel Oil Co. v. Atlantic Richfield Co., 146 F.3d 1088, 1093 (9th Cir.1998) (emphasis added) (quoting United States v. Alexander, 106 F.3d 874, 876 (9th Cir.1997)); see also Moore v. Jas. H. Matthews & Co., 682 F.2d 830, 833 (9th Cir.1982). But cf. Thomas v. Bible, 983 F.2d 152, 155 (9th Cir.1993) (suggesting, in dicta, that the “level or levels of the court or courts involved” has some bearing on a court’s discretion to deviate from law of the case).

While our case law appears to support the authority of a district court to deviate from the law of the case in appropriate circumstances, we do not decide whether such circumstances existed in this case. Instead, we reach the same decision as the district court on the alternate ground that, regardless of the district court’s authority to decline to follow the law of the case established by this court in TSPC III, we have the discretion to do so and we exercise that discretion here. The law of the case doctrine provides that a panel of this court has discretion to depart from the law of the case established by the same panel, or another, where: “(1) the decision is clearly erroneous and its enforcement would work a manifest injustice, (2) intervening controlling authority makes reconsideration appropriate, or (3) substantially different evidence was adduced at a subsequent trial.” Jeffries v. Wood, 114 F.3d 1484, 1489 (9th Cir.1997) (en banc) (footnote and internal quotes omitted).⁴³ The holding in TSPC III falls squarely into the first category: the decision was “clearly erroneous and its enforcement would work a manifest injustice.” Id.

In order to understand why the holding in TSPC III is clearly erroneous, one must first understand what it is that TSPC III held. In concluding that the defendants had forfeited their right to argue at any point in the. litigation that the plaintiffs’ § 1983 claims were time-barred, the TSPC III court stated: “The defendants did not -plead any other statute of limitations except the 60-day one. Failing to plead affirmatively any other statute of limitations, they cannot now rely on any other.” TSPC III, 34 F.3d at 756 (emphasis added). On appeal, the plaintiffs argue that the court used the term “plead” in its nontechnical sense, to mean simply “argue” or “contend.” We think it evident, however, that the court used the term in its conventional legal sense, to mean “assert in a pleading.” Both courts and lawyers know that the term “plead,” when used in a legal context, has that meaning. When a court uses the term in an opinion, therefore, we presume, in the absence of evidence to the contrary, that the court meant to employ the standard legal meaning. Here, there is no indication to the contrary. Quite the opposite: the adverb that the TSPC III court used in conjunction with the term “plead” bolsters our reading. In rejecting the affirmative defense, the court did not merely state that the defendants had failed to “plead” the proper limitation period; instead, it stated that the defendants had failed to “plead affirmatively” the proper period. This more specific term — “plead affirmatively” — makes it even clearer to us that the TSPC III court used the term “plead” in its legal rather than popular sense. After all, the statute of limitations defense at issue in TSPC III is an affirmative defense — a type of defense that, in legal parlance, must be “plead affirmatively.” For these reasons, we conclude that TSPC III held that the defendants forfeited their right to argue another limitations period by failing to assert that period in a responsive pleading.

Given our conclusion about what TSPC III held, it follows that TSPC III’s holding was clearly erroneous. At the time that we heard the appeal in TSPC III, none of the defendants had ever filed an answer, or any other type of responsive pleading. See TSPC, 992 F.Supp. 1218, 1223 (D.Nev.1998). Instead, in response to both the original complaints and the amended complaints, the defendants had filed motions to dismiss. A motion to dismiss is not a pleading. See Fed.R.Civ.P. 7(a) (designating “pleadings”); Miles v. Department of Army, 881 F.2d 777, 781 (9th Cir.1989) (“[A] motion to dismiss the complaint is not a responsive pleading.”). Thus, it was not until the remand following TSPC III, when the defendants filed answers to the plaintiffs’ complaints, that the defendants finally “pled.”⁴⁴ And in those first responsive pleadings, the defendants pled the pertinent statutes of limitations for the § 1988 claims. Accordingly, TSPC III clearly erred in holding that the defendants failed to “plead affirmatively” those statutes.

In addition to the clear factual error, TSPC Ill’s bare legal holding — that the defendants forfeited the correct statute of limitations defense — is clearly wrong. The defendants raised the correct limitations periods in their answers. The inclusion of the defense in an answer is sufficient to preserve the defense. See Magana v. Northern Mariana Islands, 107 F.3d 1436, 1446 (9th Cir.1997); Fed. R. Civ. Pro. 8(c). In fact, we have permitted defendants to raise affirmative defenses for the first time well after filing an answer, see Magana, 107 F.3d at 1446, and have also permitted district courts to raise an affirmative defense sua sponte, so long as the defense has not been affirmatively waived, see Levald, Inc. v. City of Palm Desert, 998 F.2d 680, 686-87 (9th Cir.1993).

Although it is true that, prior to filing their answers, the defendants had asserted, by way of a motion to dismiss, a statute of limitations defense premised on a mistaken choice of law, the defendants’ mistake does not in any way constitute a waiver of all other limitations periods. See Zotos v. Lindbergh Sch. Dist., 121 F.3d 356, 361 (8th Cir.1997) (rejecting argument that defendant waived application of 90-day limitation period by relying only on a two-year limitation period in its summary judgment motion); Yoder v. Honeywell Inc., 104 F.3d 1215, 1224 n. 3 (10th Cir.1997) (“We hold that [the defendant] did not waive the statute of limitations defense simply by pleading the defense based on the wrong choice of law.”) Cf. Daingerfield, Island Protective Soc’y v. Babbitt, 40 F.3d 442, 444-45 (D.C.Cir.1994).⁴⁵

In addition to being clearly erroneous, TSPC III’s holding would result in a manifest injustice were we to follow it. Jef-fries held that a court may find that an erroneous decision works a manifest injustice if “the challenged decision ... involve[s] a significant inequity or the extin-guishment of a right....” Jeffries, 114 F.3d at 1492. In the present case, the clear error in TSPC III would lead to a significant inequity. Were we to follow TSPC III’s erroneous holding, the defendants would lose their opportunity to present this dispositive defense — a defense that fully vindicates their right to be free from a trial and an adverse damage award. Of course, depriving a defendant of a statute of limitations defense might not always be significantly inequitable. Here, however, the defendants did nothing that would have misled the plaintiffs or constituted a forfeiture of that defense. In marked contrast, the plaintiffs failed to bring their claims regarding the 1987 Plan until more than four years after the plan became effective, which is more than two years after Nevada’s statute of limitations had run, and more than three years after California’s. These circumstances, combined with the fact that the defense is both timely and dispositive, convince us that the requisite inequity exists in the present case.

Because the time bar holding in TSPC III is clearly erroneous and would be manifestly unjust if enforced, we decline to follow it. On the merits, the district court held the plaintiffs’ claims time-barred, and the plaintiffs affirmatively decline to argue on appeal that the district court’s resolution of that question is incorrect. Even when asked by this court at oral argument, the plaintiffs stressed that they disputed only the district court’s authority to deviate from the law of the case set forth in a disposition of this court. Their decision not to argue the merits of the time-bar issue is not surprising, given that they filed the Period IV claims years after the relevant statutes of limitations had run. Because the plaintiffs offer no argument on the merits of the timeliness of their Period IV claims, we affirm the district court’s dismissal of those claims.

AFFIRMED in part and REVERSED in part. REMANDED for entry of judgment in favor of defendants in accordance with this disposition.

1 . The destruction of vegetation increases the transport of nutrients to the lake in part because vegetation prevents soil erosion. Similarly, the construction of impervious surfaces increases the flow of nutrients to the lake in part because such surfaces prevent rain and snowmelt from seeping into the ground. This increases the surface flow of water, which hastens erosion.

2 . The amended Compact took effect on December 19, 1980.

3 . The 1980 Compact defines “environmental threshold carrying capacity” as "an environmental standard necessary to maintain a significant scenic, recreational, educational, scientific or natural value of the region or to maintain public health and safety in the region. Such standards shall include but not be limited to standards for air quality, water quality, soil quality, soil conservation, vegetation preservation and noise.” Art. II(i), 94 Stat. 3235.

4 . Obviously, the plaintiffs’ initial complaints, filed in 1984, did not include claims regarding the 1987 Regional Plan. Following remand from TSPC I and TSPC II, however, the plaintiffs in both actions were permitted to amend their complaints. In their amended complaints, the plaintiffs added claims regarding the 1987 Regional Plan.

5 . In TSPC I, a panel of this court affirmed the dismissal of the Nevada-side, Class 1-3 plaintiffs' Period I claims, concluding that the claims were unripe because the plaintiffs could have applied to TRPA for construction permits under the limited exception contained in Ordinance 81-5. See TSPC I, 911 F.2d at 1339; see also Tahoe-Sierra Preservation Council, Inc. v. Tahoe Reg'l Planning Agency, 638 F.Supp. 126 (D.Nev.1986). Because this exception was not available for SEZ-classified lands, however, the panel held that the Nevada-side, SEZ plaintiffs’ Period I claims were ripe. See id. at 1339. In their 1991 amended complaint, the Nevada plaintiffs attempted to reinstate the Class 1-3, Period I claims on the ground that it would have been "futile” for them to seek permits under the limited exception to Ordinance 81-5. The district court rejected this argument, holding that TSPC I controlled, and that the plaintiffs should have made their futility argument to the Ninth Circuit in TSPC I. See Tahoe-Sierra Preservation Council, Inc. v. TRPA, 808 F.Supp. 1474, 1478 (D.Nev.1992). It therefore dismissed those claims once again. See id. It appears from TSPC III that the plaintiffs did not appeal the district court’s ruling. See TSPC III, 34 F.3d at 755 (appeal only of statute of limitations and Period III causation issues).

6 . The TSPC III court did not dismiss the California-side, Class 1-3 plaintiffs’ Period I claims as unripe, because the exception contained in Ordinance 81-5 for case-by-case approval of development, which was available to the Nevada-side Class 1-3 plaintiffs, was not available to the California plaintiffs. See TSPC II, 938 F.2d at 156; supra note 5.

7 . In TSPC I, a panel of this court affirmed the dismissal of the Nevada-side plaintiffs’ claims for Period III. There was no majority rationale, however, for the affirmance of the dismissal. See TSPC I, 911 F.2d 1336-39 (per curiam) (affirming the dismissal on the ground that the claims were unripe); id. at 1343-44 (B.Fletcher, J., concurring) (arguing that the dismissal should be affirmed on the ground that the claims were meritless); id. at 1344-47 (Kozinski, J., dissenting in part) (arguing that the claims were ripe); see also Tahoe-Sierra Preservation Council, 638 F.Supp. at 131-33 (dismissing the claims as unripe). In TSPC II, when confronted with the indistinguishable Period III claims of the California-side plaintiffs, this court reversed the district court’s dismissal of those claims as unripe. See TSPC II, 938 F.2d at 157 (holding that the claims were ripe, but remanding rather than reaching the merits). While TSPC II reached the opposite result from TSPC I, the Nevada-side claims were not before TSPC II. Consequently, TSPC II did not affect the TSPC I court's judgment against these claims.

Following remand from TSPC I and TSPC II, the California and Nevada cases were consolidated, and the plaintiffs were permitted to file amended complaints. In their amended complaint, the Nevada plaintiffs attempted to reinstate their Period III claims on the ground that TSPC II was the law of the case. The district court rejected this argument, and dismissed those claims once again. The district court held that TSPC I's ruling that the claims be dismissed was also law of the case, and that the district court would follow TSPC I's judgment in spite of the inconsistency with TSPC II's reasoning. See Tahoe Sierra Preservation Council, Inc. v. TRPA, 808 F.Supp. 1474, 1478 (D.Nev.1992). It appears from TSPC III that the plaintiffs did not appeal the district court's ruling. See TSPC III, 34 F.3d at 755.

To the extent that any of the rulings in TSPC I and TSPC II are inconsistent, our decision herein renders that fact inconsequential. Even had the Nevada plaintiffs' Period III claims not been dismissed, the plaintiffs would have lost as to those claims on this appeal for the same reasons applicable to the California plaintiffs.

8 . In accord with the above chart of claims, the district court stated that only the following claims remained to be addressed at the trial: the Period I § 1983 claims of all the California plaintiffs and the Nevada SEZ plaintiffs; the Period II § 1983 claims of all the plaintiffs; and the Period III claims of the California plaintiffs.

9 . The District Court certified all of these issues for interlocutory appeal pursuant to 28 U.S.C. § 1292(b), and we granted the petition for permission to appeal.

10 .For much of the nation’s history, it was generally thought that the Takings Clause reached only the direct appropriation, or the functional equivalent of a "practical ouster of [the owner’s] possession.” Transportation Co. v. Chicago, 99 U.S. 635, 642, 25 L.Ed. 336 (1879). In Pennsylvania Coal Co. v. Mahon, 260 U.S. 393, 43 S.Ct. 158, 67 L.Ed. 322 (1922), however, the Supreme Court recognized that government regulation of property could "go[ ] too far” and amount to a taking of that property. See id. at 415, 43 S.Ct. 158.

11 . In the special context of exactions — land-use decisions conditioning approval of development on the dedication of property to public use — the Supreme Court has developed a different test to determine whether a regulatory taking has occurred. See Dolan v. City of Tigard, 512 U.S. 374, 114 S.Ct. 2309, 129 L.Ed.2d 304 (1994); Nollan v. California Coastal Comm’n, 483 U.S. 825, 107 S.Ct. 3141, 97 L.Ed.2d 677 (1987). The Court has held that the Nollan/Dolan test is inapposite to regulatory takings cases outside the context of excessive exactions. See City of Monterey v. Del Monte Dunes at Monterey, Ltd., 526 U.S. 687, 703, 119 S.Ct. 1624, 143 L.Ed.2d 882 (1999) (overruling in part Del Monte Dunes at Monterey, Ltd. v. City of Monterey, 95 F.3d 1422 (9th Cir.1996)).

12 . As do other courts, we use the term "categorical taking” throughout this opinion as a shorthand reference to government action that is determined to be a "taking” under the categorical approach set forth in Lucas (where the "economic impact of the regulation” is determinative). See, e.g., District Intown Properties, 198 F.3d at 882-83; Dodd, 136 F.3d at 1228. In other words, as with the categorical approach set forth in Loretto, it is the approach that is categorical, not the "taking.”

13 . The problem of defining the relevant property interest at stake is commonly referred to as either the "denominator problem” or the problem of "conceptual severance.”

14 . Although Andrus demonstrates that individual strands of the property-rights bundle should generally not be treated as separate property interests, the Supreme Court has held that the total abrogation of the "right to exclude” may in certain circumstances constitute a taking because it "will result in an actual physical invasion of the privately owned [property].” Kaiser Aetna v. United States, 444 U.S. at 179-80, 100 S.Ct. 383. Cf. Irving, 481 U.S. at 716, 107 S.Ct. 2076 (holding, under the special circumstances of that case, that the complete abrogation of the plaintiffs’ rights of descent and devise was "similarf ]” to the denial of the right to exclude invalidated in Kaiser Aetna, and thus resulted in a taking).

15 . In fact, there is evidence that such a race-to-develop occurred in the Lake Tahoe Basin in the years preceding the adoption of the 1980 Compact. See Tahoe-Sierra Preservation Council, 34 F.Supp.2d at 1241.

16 . The Court was careful to include quotation marks around the word "temporary" whenever it referred to a "temporary” taking, in order to make clear that it was using the concept in the specific sense in which it had defined it.

17 . In concluding that First English created a categorical rule in favor of temporal severance, the district court relied heavily on the Court's statement in First English that it "do[es] not deal with the quite different questions that would arise in the case of normal delays in obtaining building permits, changes in zoning ordinances, variances, and the like which are not before us.” 482 U.S. at 321, 107 S.Ct. 2378. The district court held that the statement amounted to an enumeration of the acceptable forms of planning-related temporary development prohibitions, and that this statement implied the existence of a categorical rule prohibiting temporary development delays caused by regulatory actions other than those listed in the Court's statement. We do not think, however, that the Court intended to create a new categorical rule concerning what constitutes a taking (one that would directly conflict with the principles applied in previous cases) in a passing remark about what the Court "do[es] not deal with.” This is especially so in light of the fact that, in First English, the Court repeatedly stated that the question whether a taking occurred was not before it.

Even were we to accord the Court’s dictum some weight, it appears to support, rather than undermine, our position. The list suggests that planning activities that temporarily prohibit development do not constitute takings. There is no suggestion that the list is exclusive. In fact, reading it to be exclusive would require us to draw a constitutional distinction between different conventional planning tools that result in similar development delays. We think such a distinction is untenable. Moreover, even were we to treat the list as exclusive, it appears to encompass temporary planning moratoria. The list allows for "normal delays in obtaining ... changes in zoning ordinances.” 482 U.S. at 321, 107 S.Ct. 2378. The reason that changes in zoning ordinances generally lead to delays and development, however, is that local governments often enact temporary development moratoria during the period in which the new ordinance is being formulated. See supra at 777. Thus, even the list itself appears to contemplate temporary development morato-ria.

18 . In fact, in a World War II era case involving a regulation that prohibited the owners of gold mines from using those mines, rather than permitting the government to take possession of the property, the Court found that no taking occurred. See United States v. Central Eureka Mining Co., 357 U.S. 155, 168-69, 78 S.Ct. 1097, 2 L.Ed.2d 1228 (1958).

19 . Although TRPA first passed Ordinance 81-5 and later adopted Resolution 83-21, each regulation was designed to terminate upon the occurrence of the same event — TRPA’s adoption of an amended regional plan. Accordingly, we treat the regulations as creating a single temporary moratorium that was designed to and did run from the date on which Ordinance 81-5 became effective to the date on which TRPA adopted the 1984 Plan.

20 . This is actually an overstatement. As the district court noted, both regulations may have permitted certain limited development, particularly on the Class 1-3 land. See Tahoe-Sierra Preservation Council, 34 F.Supp.2d at 1243-44. The district court concluded that these development possibilities were negligible, and did not affect the "takings” calculus. While we have some doubts about the district court's conclusion, the correctness of that conclusion is irrelevant to our analysis. Accordingly, we will assume arguendo that the moratorium prevented all development in the period during which it was in effect.

21 . We acknowledge that, given that moratoria are, by definition, temporary, it is redundant to refer to a moratorium as a "temporary moratorium.” See Webster's Third New International Dictionary 1469 (1976) (defining "moratorium” as a "waiting period set by some authority: a delay officially required or granted”). Nevertheless, because the parties, the district court, and this court have all repeatedly referred to the moratorium at issue in the present case as a "temporary moratorium,” and because it emphasizes the temporary nature of the development ban, we use that designation here. See, e.g., TSPC I, 911 F.2d at 1339; TSPC II, 938 F.2d at 155; Tahoe-Sierra Preservation Council, 34 F.Supp.2d at 1248.

22 . See William C. Haas & Co., 605 F.2d at 1120-21 (noting the close relationship between a land-use regulation’s effect on the available "use” of a property and the effect on its "value”); MacLeod, 749 F.2d at 546 (implicitly noting the same).

23 . Many cases treat the "use” and "value” interchangeably, or speak only of the effect of a regulation on the property's value. See, e.g., TSPC I, 911 F.2d at 1335 (suggesting that a "taking” is measured by "diminution of the value of plaintiffs' land”); Lucas, 505 U.S. at 1017, 112 S.Ct. 2886 ("What is land but the profits thereof?”); Virginia Surface Mining & Reclamation Ass’n, 452 U.S. at 295, 101 S.Ct. 2352 (the "ad hoc, factual inquiries” required under Agins's "viable used” prong "must be conducted with respect to ... particular estimates of economic impact and ultimate valuation relevant in the unique circumstances”); MacLeod, 749 F.2d at 549 ("The determination of whether or not an economically viable use of property remains ... turns upon the economic impact of the regulation and, particularly, the extent to which the regulations or prohibition has interfered with the distinct investment-backed expectations of the claimant, "(citation omitted)); Carson Harbor Village Ltd. v. City of Carson, 37 F.3d 468, 476 (9th Cir.1994), overruled on other grounds by WMX Technologies v. Miller, 104 F.3d 1133, 1136 (9th Cir.1997) (en banc) ("In the takings context, the basis of a facial challenge is that the very enactment of the statute has reduced the value of the property or has effected a transfer of a property interest.”); Keystone Bituminous Coal Ass’n, 480 U.S. at 493, 500-501, 107 S.Ct. 1232; Suitum, 520 U.S. at 748, 117 S.Ct. 1659 (O’Connor, J., concurring); Lucas, 505 U.S. at 1009, 1010, 1017-18 & n. 8, 1020 & n. 9, 1026 & n. 13, 112 S.Ct. 2886; Agins, 447 U.S. at 262, 100 S.Ct. 2138; Penn Coal, 260 U.S. at 413, 43 S.Ct. 158; Concrete Pipe, 508 U.S. 602, 644, 113 S.Ct. 2264, 124 L.Ed.2d 539 (1993).

24 . The parties disagree over what kind of evidence either party may introduce about the effect of the land-use regulations on specific pieces of property. Before trial, the parties stipulated that, because the action consists only of a facial challenge, “the parties will not introduce evidence regarding the specific factual situations of individual plaintiffs.” In spite of this stipulation, the defendants argue on appeal that, even though the plaintiffs brought only a facial challenge, the plaintiffs cannot succeed in proving that a taking occurred without introducing evidence regarding the specific economic impact of the regulations on their individual properties. The defendants’ argument that individualized evidence is required is in tension with the pretrial stipulation. It is also in tension with the statement, made in several cases, that a facial taking challenge "present[s] no concrete controversy concerning ... [the legal provision’s] effect on specific parcels of land.” Virginia Surface Mining & Reclamation Ass’n, 452 U.S. at 295, 101 S.Ct. 2352. This tension need not be resolved, however, because reference to the general features of the land-use regulations at issue is sufficient to demonstrate that those regulations do not deprive the plaintiffs' property of all use or value.

25 . In resolving the question of the remaining value of the plaintiffs’ property, the district court relied heavily on Del Monte Dunes at Monterey, Ltd. v. City of Monterey, 95 F.3d 1422 (9th Cir.1996). Because the general features of the regulation at issue in the present case make clear that the plaintiffs’ property retains value, Del Monte Dunes's discussion of proof of value is irrelevant to our determination. We should note, however, that the district court misread Del Monte Dunes in at least two ways. First, the district court ignored the fact that the Del Monte Dunes court was reviewing a jury’s finding that a taking occurred, and that the Del Monte Dunes court thus considered only whether there was sufficient evidence to support the jury's finding. See 95 F.3d at 1433. Second, the district court interpreted Del Monte Dunes to hold that a "competitive market” is necessary to preclude a finding that a taking occurred. However, Del Monte Dunes does not stand for so strong a proposition: the Del Monte Dunes court held only that, as a matter of law, the presence of a single buyer was not sufficient to require the court to upset the jury’s finding that a taking had occurred. See id. at 1432-33. The only mention Del Monte Dunes makes of a "competitive market” is its suggestion that the absence of such a market is a piece of evidence that may tend to support a finding of an absence of value. See id. at 1433.

26 .This economic reality is precisely what differentiates a permanent ban on development, even if subsequently invalidated, from a temporary one. Basic economic theory demonstrates that the present value of land depends on the potential for future use. Accordingly, when a permanent development ban (like the one at issue in Lucas) is enacted, the value of the affected land plummets, on account of the fact that the ban bars all future development of the property. In contrast, when a temporary ban is enacted, both the owner of the affected land and any prospective purchasers know that, at a specific point in the future, the moratorium will no longer prohibit the development of the land at issue.

27 . We note that the average time between land purchase and development may be as long as 25 years. See Tahoe-Sierra Preservation Council, 34 F.Supp.2d at 1240.

28 . This corollary also demonstrates that, had we engaged in conceptual severance, we would have read into the Takings Clause a requirement that the government never interfere with a property owner's wish to put his property to immediate use.

29 . Although we need not rely on it, we note that, even during the period of time during which the moratorium was in effect, there were cognizable "uses” to which the basin property could have been put. Although we assume for purposes of our disposition that no development of the property was permitted, see supra note 19, "use” includes much more than development. For example, Lucas treats the sale of property as an “economically productive use.” See Lucas, 505 U.S. at 1028, 112 S.Ct. 2886. In addition, we have held that "[h]olding property for investment purposes can be a 'use' of property.” MacLeod, 749 F.2d at 546 n. 7.

30 . Note that, in so holding, we do not overturn any factual findings made by the district court, only its legal conclusions. Although the district court found that the plaintiffs were denied "all economically beneficial or productive use” of their property, it did so on the basis of its legal conclusion that only substantial developmental uses available during the moratorium period were relevant.

31 . This portion of the appeal concerns only the California-side plaintiffs. See supra note 7.

32 . Because the district court concluded that TRPA could not be held liable for any taking that occurred during the 1984-87 period, it did not determine (1) whether the developmental moratorium effected by the injunction actually amounted to a taking, or (2) whether the 1984 Plan, had it been implemented, would have amounted to a taking. We, too, do not decide these questions.

33 . It is true that there is little discussion of a "causation” requirement in any of the case law involving regulatory takings. But cf. Penn Central, 438 U.S. at 124, 98 S.Ct. 2646 (noting, in passing, the proximate cause requirement). However, this is due to nothing more than the fact that, in most regulatory takings cases, there is no doubt whatsoever about whether the government’s action was the cause of the alleged taking.

34 . The plaintiffs suggest that, in the absence of sufficient guidance from § 1983 cases or other federal constitutional case law, we should look to state tort law for guidance as to the meaning of proximate cause. The Supreme Court "ha[s] repeatedly noted that 42 U.S.C. § 1983 creates a species of tort liability, and ha[s] interpreted the statute in light of the background of tort liability.” City of Monterey v. Del Monte Dunes at Monterey, Ltd., 526 U.S. 687, 709, 119 S.Ct. 1624, 143 L.Ed.2d 882 (1999) (citations and internal quotation marks omitted). Nevertheless, there are meaningful differences between § 1983 liability and state tort liability, and these differences may in certain circumstances diminish the relevance of, or warrant departure from, the principles of tort liability. See Monell v. Department of Social Servs., 436 U.S. 658, 693-94, 98 S.Ct. 2018, 56 L.Ed.2d 611 (1978) (holding that the tort law doctrine of respondent superior does not apply to § 1983 suits brought against local governments). In the present case, we need not rely on any specific state tort case law, because it is clear that under forseeability analysis— which is widely accepted by both federal and state case law as a conventional method of assessing proximate cause — TRPA is not legally responsible for any taking that occurred during Period III.

35 . Although the plaintiffs did not raise this argument until their cross-appeal reply brief, we exercise our discretion to address it on the merits.

36 . Because we conclude that the injunction prohibited the implementation of the 1984 Plan, we need not decide whether the plan, had it been implemented, would have constituted a taking of the plaintiffs’ property. See supra note 32.

37 . The plaintiffs suggest that the 1984 Plan somehow effected a taking in the few days that passed between the time of the plan's adoption and the adoption of the TRO. As Judge B. Fletcher noted in TSPC I, however, it would have been impossible for TRPA to implement the plan during these few days, or for the plaintiffs to "complete[ ] all the necessary prerequisites to building, including getting a TRPA permit.” TSPC I, 911 F.2d at 1344 (B. Fletcher, J., concurring).

38 . The plaintiffs also argue that, even if the plan did not go into effect, the injunction did not “prevent TRPA from providing the Plaintiffs with Just Compensation.” The failure to provide such compensation, they assert, is itself an actionable wrong. The plaintiffs’ argument skips over the question of liability. If TRPA is not legally responsible for any taking that occurred during Period III, then TRPA has no obligation to provide just compensation to the plaintiffs.

39 .Although it is a logically necessary component of their argument, the plaintiffs do not directly support their implicit assertion that the court injunction resulted in a taking of the plaintiffs’ property. Their general argument that a taking occurs when a legislative or executive body implements a temporary building moratorium provides indirect support for this position, but the plaintiffs never explain why a temporary moratorium ordered by a court should be treated in the same fashion. In fact, there are good reasons to believe that a government action that would otherwise constitute a taking should be treated differently when it is ordered by a court. Moreover, even if a court order could qualify as a taking, it is not clear who, if anyone, would be held liable for that taking under § 1983. Obviously, however, it would ordinarily not be the party that has no choice but to follow the court order.

40 . Article VI(j)(4) of the 1980 Compact reads as follows:

A legal action arising out of the adoption or amendment of the regional plan or of any ordinance or regulation of the agency ... shall be commenced within 60 days after final action by the agency.

41 . The district court dismissed the Period III claims on the ground that the 1984 Plan caused no harm because it had been enjoined. See TSPC III, 34 F.3d at 755; see also supra text accompanying notes 30-38 (affirming the district court’s holding regarding the Period III claims).

42 . Because § 1983 does not contain a statute of limitations, federal courts apply the forum state's statute of limitations for personal injury claims. See Wilson v. Garcia, 471 U.S. 261, 276, 105 S.Ct. 1938, 85 L.Ed.2d 254 (1985). In California, the applicable statute of limitations is one year. See Johnson v. California, 207 F.3d 650, 653 (9th Cir.2000). In Nevada, the applicable statute of limitations is two years. See Perez v. Seevers, 869 F.2d 425, 426 (9th Cir.1989).

43 . For some time, there have existed in the Ninth Circuit two different formulations of the set of circumstances in which a court may decline to follow the law of the case. The first formulation, set forth in a line of cases that includes Jeffries, states that a court may depart from the law of the case if "the [previous] decision is clearly erroneous and its enforcement would work a manifest injustice.” Jeffries, 114 F.3d at 1489 (emphasis added); see also Leslie Salt Co. v. United States, 55 F.3d 1388, 1392 (9th Cir.1995). In contrast, the second formulation states that a court may decline to follow the law of the case if "the first decision was clearly erroneous” or "a manifest injustice would otherwise result.” Cuddy, 147 F.3d at 1114; see also United States v. Alexander, 106 F.3d 874, 876 (9th Cir.1997); Russell v. Commissioner of Internal Revenue, 678 F.2d 782, 785 (9th Cir.1982). While one might have expected that Jeffries, being an en banc decision, would control all future cases, at least three post-Jeffries cases cite the disjunctive formulation, which is inconsistent with Jeffries. See United States v. Scrivner, 189 F.3d 825, 827 (9th Cir.1999); Cuddy, 147 F.3d at 1114; Disimone v. Browner, 121 F.3d 1262, 1266 (9th Cir.1997).

Recently, we noted this conflict — which appears to have gone unrecognized in previous case law — but did not resolve it. Mendenhall v. NTSB, 213 F.3d 464(9th Cir.2000). As in Mendenhall, we need not resolve the conflict in the present case. For, as the following discussion demonstrates, there are grounds here to depart from the law of the case set forth in TSPC III under the more stringent Jeffries standard as well as under the more popular disjunctive standard. (Of course, any case that meets the Jeffries standard a fortiori meets the disjunctive standard.)

44 . Of course, il is understandable that the panel in TSPC III might not have realized that, ten years into the litigation, the defendants had not yet actually answered the complaints.

45 . The plaintiffs suggest that if a defendant raises a particular time bar in a motion to dismiss, it must raise all potential time bars in that motion. This argument is clearly incorrect. Some time bars may be resolved on such a motion and others may not. For example, a time bar that turns on the resolution of factual disputes may only be resolved at a later stage of the litigation. It would make little sense to require defendants with multiple potential time bar defenses, some of which may be subject to resolution on a motion to dismiss and some of which may not, to raise all of those defenses at the very outset of the proceeding.