concurring in the judgment:
The District of Columbia’s Historic Preservation Board imposed historic landmark status not only on an apartment building named Cathedral Mansions South but also on a substantial stretch of adjacent lawn bordering the sidewalks of Connecticut Avenue. District Intown, the owner of both, claims that as applied to the lawn the landmarking effects a taking of its property in violation of the Takings Clause of the Fifth Amendment. The majority’s disposition is — with one important exception — in general accord with the current opinions of the Supreme Court. Those decisions are of course binding. At the same time, however, it is not inappropriate to identify ways in which the prevailing analysis elevates formal concepts over economic reality and tends to strip the Clause of its potential for fulfilling the framers’ likely purposes.
The economist’s justification for the Takings Clause is that it provides a check on government’s likely tendency to waste resources by treating private property as a free good. See Richard A. Posner, Economic Analysis of Law 58 (4th ed.1992) (“The simplest economic explanation for the requirement of just compensation is that it prevents the government from overusing the taking power.”). This is just an application of the general principle that if *885a firm can externalize costs (e.g., the health costs of polluting the air), it will use more of the unpriced resource (in this example, air as a waste sink) than it would if required to pay. And it will tend to overproduce the goods or services whose production uses the. superficially “free” good — i.e., it will produce them at a level where the true value of the extra inputs exceeds the true value of the extra output. See generally Robert Cooter & Thomas Ulen, Law and Economics 45-46 (1988). As applied to government regulation, similar oversupply can be expected — here, production of regulations that impose more costs than they afford benefits, that do more harm than good.
The framers, though not articulating the purpose of the Clause in economic terms, evidently did view it as aimed at correcting the incentives of the political branches. There is evidence, for example, that James Madison saw electoral power slipping into the hands of a non-landholding majority, which in a “leveling” mode could be expected to invade landowners’ rights. See William Michael Treanor, The Original Understanding of the Takings Clause and the Political Process, 95 Colum. L.Rev. 782, 849 (1995). Late twentieth century America, of course, displays a far greater range of purposes than “leveling” for reallocation of rights. While the resulting proposals are naturally advanced in the name of the public good, many are surely driven by interest-group purposes, commonly known as “rent-seeking.” Among these proposals, at least some inflict aggregate costs considerably outweighing their aggregate benefits, paralleling the wasteful production associated with private firms’ externalization of costs. The Takings Clause serves to curb such inefficiencies. See, e.g., Richard A. Epstein, Takings: Private Property and the Power of Eminent Domain 281 (1985) (“[T]he Takings Clause is designed to control rent seeking and political faction. It is those practices, and only those practices, that it reaches.”).
A Takings Clause construction that was dedicated without qualification to preventing such government externalization would require compensation whenever regulation reduced the value of anyone’s property, however slightly. Balanced against that goal is an array of considerations. Most obvious is the cost of calculating and administering compensation, which would tend to sink many a beneficent statute. “Government hardly could go on if to some extent values incident to property could not be diminished without paying for every such change in the general law.” Lucas v. South Carolina Coastal Council, 505 U.S. 1003, 1018, 112 S.Ct. 2886, 120 L.Ed.2d 798 (1992) (quoting Pennsylvania Coal Co. v. Mahon, 260 U.S. 393, 413, 43 S.Ct. 158, 67 L.Ed. 322 (1922)). (The compensation cost itself would be only a weak countervailing factor, for most beneficent regulation would presumably generate gains large enough to pay the losers if identification and calculation were costless.) My goal here is not to pinpoint the appropriate balance between these competing considerations, much less to suggest that the correct reading is one under which all regulation materially adversely affecting a property’s value would be compensable. Rather, it is simply to note the ways in which modern interpretation of the Takings Clause, as exemplified in today’s decision, impairs its role as a disincentive to wasteful government activities.
The majority applies an apparent presumption that contiguous parcels under common ownership should be treated as one parcel for purposes of the takings analysis. This presumption tends to reduce the likelihood that courts will order compensation. The larger the parcel, the greater the chance that the regulated land will retain an economically viable use. Where no such use remains, there is a “total taking” and the government can “resist compensation only if the logically antecedent inquiry into the nature of the owner’s estate shows that the proscribed use *886interests were not part of his title to begin with,” Lucas, 505 U.S. at 1027, 112 S.Ct. 2886; where an economically viable use survives regulation, the best the owner can hope for is “partial” takings analysis. Under the latter courts will.determine whether to award compensation by looking to “the economic impact of the regulation, its interference with reasonable investment backed expectations, and the character of the governmental action,” Kaiser Aetna v. United States, 444 U.S. 164, 175, 100 S.Ct. 383, 62 L.Ed.2d 332 (1979); see also Eastern Enterprises v. Apfel, 524 U.S. 498, -, 118 S.Ct. 2131, 2146, 141 L.Ed.2d 451 (1998); Lucas, 505 U.S. at 1019 n. 8, 112 S.Ct. 2886, and will generally deny compensation so long as the restriction “substantially advance[s] legitimate state interests,” Agins v. City of Tiburon, 447 U.S. 255, 260, 100 S.Ct. 2138, 65 L.Ed.2d 106 (1980); see also Dolan v. City of Tigard, 512 U.S. 374, 385, 114 S.Ct. 2309, 129 L.Ed.2d 304 (1994). New regulations will flunk this nearly vacuous test. In fact, the Supreme Court has only once found a partial taking to be compensable, and even then only a plurality applied the partial takings analysis. See Eastern Enters., 524 U.S. at -, 118 S.Ct. at 2149; see also id. at ---, 118 S.Ct. at 2154-60 (Kennedy, J.) (rejecting the plurality’s 'takings analysis and finding invalidity on other grounds).
The Supreme Court has offered several justifications for this distinction between partial and total takings. See, e.g., Lucas, 505 U.S. at 1017-18, 112 S.Ct. 2886 (suggesting that “from the landowner’s perspective,” a total taking is tantamount to a physical taking, and that from the government’s perspective the concern that an obligation to compensate for any incidental value diminution would impede effective functioning cannot apply in the “relatively rare situations” of total takings). From the perspective of ensuring that the government not engage in wasteful behavior, however, the focus on the uses of the land that remain is misplaced: “[W]hat is decisive is that which is taken, not that which is retained.” Epstein, Takings, supra, at 58. Whether the landowner is left with a limited use of the land or none at all is hardly relevant to that issue. And as the regulating government delineates the scope of regulation, the opportunity for strategic behavior is obvious.
The majority’s cursory application of the Penn Central factors further broadens the gap between the two modes of analysis, reinforcing the seemingly predetermined conclusion: in partial takings cases, the government wins. The majority states that District Intown has not shown the land “unprofitable to maintain,” Maj. Op. at 883; it is unimaginable, however, absent an extraordinary tax liability, that a parcel could retain an economically viable use yet have a net negative value. The majority goes on to say that District Intown has failed to show that the land does not “bring in a sufficient return,” id., but does not answer the all-important question: a return on what? on out-of-pocket costs? on initial purchase price? on fair market value? Moreover, the majority provides no guidance as to how “sufficient” the return must be, except to cite Penn Central, in which the Court found that a 75% diminution in value did not constitute a com-pensable taking. See id.
Similarly, in its consideration of District Intown’s “reasonable investment-backed expectations,” the majority’s analysis begs the question whether any landowner, in a world where zoning regulations are prevalent, could ever argue that a particular regulation was “unexpected.” The presumption is insurmountable: “Businesses that operate in an industry with a history of regulation have no reasonable expectation that regulation will not be strengthened to achieve established legislative needs.” Maj. Op. at 883-84. Although the 1931 Shipstead-Luce Act might have put District Intown on notice that some regulation of architectural design might be expected, it is farfetched to conclude that District Intown, merely because of its *887proximity to the zoo, should reasonably have anticipated an absolute ban on construction; the city’s counsel, under questioning at oral argument, failed to identify any uses, or even attempted uses, of the Shipstead-Luce Act to support a complete construction veto. Although the Takings Clause is meant to curb inefficient takings, such a notion of “reasonable investment-backed expectations” strips it of any constraining sense: except for a regulation of almost unimaginable abruptness, all regulation will build on prior regulation and hence be said to defeat any expectations. Thus regulation begets regulation.
Although the presumption in favor of looking at the parcel as a whole, and in turn the increased reliance on the partial takings mode of analysis, is at odds with the underlying principle of the Takings Clause, it is perhaps the best construction of the Supreme Court’s limited guidance. The Court has never squarely addressed the question of how courts should define the relevant geographic parcel of land, also known as “horizontal severance.” Marc R. Lisker, Regulatory Takings and the Denominator Problem, 27 Rutgers L.J. 663, 705 (1996). In Nectow v. City of Cambridge, 277 U.S. 183, 48 S.Ct. 447, 72 L.Ed. 842 (1928), the Court considered whether the city council had effectuated a taking of plaintiffs land by zoning as “residential” a 100-foot strip on plaintiffs 140,000 square foot parcel. Although the Court appeared to treat the relevant parcel as encompassing only the fractional strip, this was in no respect relevant to the Court’s decision. In Penn Central Transportation Co. v. City of New York, 438 U.S. 104, 98 S.Ct. 2646, 57 L.Ed.2d 631 (1978), the Court applied a very weak form of horizontal severance, focusing exclusively on the landmarked building itself without treating the owner’s neighboring — but not adjacent — property as part of the greater parcel, as had the New York Court of Appeals. See Penn Central Transportation Co. v. City of New York, 42 N.Y.2d 324, 397 N.Y.S.2d 914, 366 N.E.2d 1271, 1276-77 (N.Y.1977). But Penn Central tells little, as the properties were not all contiguous, had been put to different uses, and had never been treated as a unified whole by the owners or the City.
Penn Central’s handling of “vertical severance,” however, is informative, if only by analogy. Using language seemingly broad enough to encompass horizontal severance, the Court made clear that it would not consider the air rights above Grand Central separately from the land rights: “ ‘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.” Penn Central, 438 U.S. at 130, 98 S.Ct. 2646; see also Keystone Bituminous Coal Ass’n v. DeBenedictis, 480 U.S. 470, 496-502, 107 S.Ct. 1232, 94 L.Ed.2d 472 (1987) (refusing to regard either coal that statute required miners to leave in place (about 2% of total coal), or the “support estate,” as distinct property for ascertaining, whether statute denied owners all economically viable uses).
The Court has expressed similar reluctance to engage in “conceptual severance” more generally (i.e., the treatment of any specific property right as a single unit). In Andrus v. Allard, 444 U.S. 51, 100 S.Ct. 318, 62 L.Ed.2d 210 (1979), the Court refused to treat extinction of the right to sell any part of a lawfully killed bald eagle as a total taking. See id. at 65-66, 100 S.Ct. 318 (“At least where 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.”). The Court arguably evidenced a retreat from this strong position in Hodel v. Irving, 481 U.S. 704, 717-18, 107 S.Ct. 2076, 95 L.Ed.2d 668 (1987), in which it found a taking in legislation that “completely abolished” certain landowners’ rights to dispose of their property by descent or devise, even though they retained complete rights to possess and to make inter vivos transfers. The Court has not, however, reached agree*888ment on the scope of this retreat. Compare id. at 719, 107 S.Ct. 2076 (Scalia, J., concurring) (saying the decision “effectively limits Allard to its facts”), with id. at 718, 107 S.Ct. 2076 (Brennan, J., concurring) (saying that the case was “unusual” and thus had no impact on Allard). Overall, I think the majority is correct in its implicit understanding that the Supreme Court is reluctant to carve a landowner’s parcel into smaller units for which compensation might be more likely.
But the factors that the majority applies in making the decision, drawn from decisions of the Federal Circuit and Claims Court and characterized by the majority as “eminently sound,” Maj. Op. at 880, strike me as uninformative and largely irrelevant. The factors considered are: (1) whether the neighboring parcels are contiguous, (2) whether they were acquired simultaneously, (3) whether they have been treated as a single unit, and (4) the extent to which the restricted lot benefits the neighboring lot. Maj. Op. at 879-80.
The first factor, contiguity, is clearly necessary but in no way sufficient. The next two factors — simultaneity of acquisition and unity of use — are more troublesome. Both elevate history — either the historical purchase or the historical use— over the real-world present relationship between the tracts. Compare Laura M. Schleich, Takings: The Fifth Amendment, Government Regulation, and the Problem of the Relevant Parcel, 8 J. Land Use & Envtl. L. 381 (1993) (proposing that courts look to the “moment of regulation” when defining the relevant parcel). The majority’s focus on the property’s use prior to regulation tells us nothing about the value-producing opportunities foreclosed at the time of regulation. “It is, of course, irrelevant that [the government] interfered with or destroyed property rights that [plaintiff] had not yet physically used. The Fifth Amendment must be applied with ‘reference to the uses for which the property is suitable, having regard to the existing business or wants of the community, or such as may be reasonably expected in the immediate future.’ ” Penn Central, 438 U.S. at 143 n. 6, 98 S.Ct. 2646 (Rehnquist, J., dissenting, quoting Boom v. Patterson, 98 U.S. (8 Otto) 403, 408, 25 L.Ed. 206 (1878)).
The majority mentions but brushes aside a fourth factor — the extent to which the regulated parcel benefits the neighboring lot. Maj. Op. at 880. Yet this appears the most relevant. The more a burdened tract in its regulated use benefits contiguous property, the less likely that the regulation has a net negative impact. In the extreme case a property interest may be worthless except in conjunction with another. Thus in Keystone Bituminous Coal Ass’n, the Court pointed out that the “support estate” had “value only insofar as it protects or enhances the value of the estate with which it is associated [i.e., the mineral estate],” 480 U.S. at 501, 107 S.Ct. 1232, and therefore refused to treat the “support estate” as a separate interest at all. Similarly, small parcels of land, either in the interior or around the edges of greater parcels, commonly are valuable only when they combine with the greater parcel to create a more valuable whole; for regulation of the exterior (such as setback requirements), then, it makes sense to measure the impact in conjunction with the “primary” parcel. Looking to the property owner’s benefit from these internal synergies parallels use of “average reciprocity of advantage,” Pennsylvania Coal Co. v. Mahon, 260 U.S. 393, 415, 43 S.Ct. 158, 67 L.Ed. 322 (1922), which considers the benefit that each burdened owner — as in ordinary zoning or historic districting — receives from the similar restriction of his neighbors.
Of course there will be some synergy between almost any two neighboring parcels under common ownership, since unified ownership creates options for the sole owner that multiple landowners could achieve only by contracting. But synergy is a matter of degree, and mere contiguity *889should not be enough. One commentator proposes a rather demanding synergy test, arguing that the regulated tract should be considered as its own parcel so long as not all of its value derives from synergies with neighboring land; in such cases, the parcel would have an independent economically viable use, which if destroyed by regulation would be compensable under Lucas. See John E. Fee, Comment, Unearthing the Denominator in Regulatory Taking Claims, 61 U. Chi. L.Rev. 1535, 1557-58 (1994). One need not go so far to see the skimpiness of the synergy here.
To be sure, Cathedral Mansions is more than several contiguous parcels. According to the decision of the Historic Preservation Review Board, “The buildings are sited imaginatively to provide the greatest possible integration of living space with well-landscaped open space.” Joint Appendix (“J.A.”) 320. (Passersby who observe the rather bare lawn will have to reach their own judgments on the adjective “well-landscaped.”) Integration there doubtless is — almost any lawn around a building will manifest a degree of integration. But there is no explicit showing that these synergies depend on the entire lawn remaining undeveloped. The proposed townhouses would cover only the portion of the lawn abutting Connecticut Avenue, still leaving the interior portion, approximately half the lawn, undeveloped. Common sense would suggest that at some distance from the building marginal synergies created by extra lawn space become slight, and thus that the part of the lawn beyond that line should be treated as its own parcel for takings purposes. Further, although District rent-control law evidently allows the owner to earn a return on the tax-assessed value of land in a single tract with a rent-controlled building (here the owner could apparently recover that status by undoing the formalities of subdivision), that value is likely to be only a tiny fraction of the value absent the historic land-marking.
In fact, it may well be completely different synergies — ones between the lawn and adjacent Connecticut Avenue — that have driven the landmarking decision. The Board observed that the lawn “contributes significantly to the unique open space character of Connecticut Avenue.” J.A. 320. A cynic might suspect that the alleged relationship between the lawn and the Cathedral Mansions apartments is little more than a cloak by which the citizens of Upper Northwest Washington have secured some parkland on the cheap. Parks are good, but the Fifth Amendment says that taking them is not.
Of course, there is another synergy between the two parcels and adjacent Connecticut Avenue, namely the historical value that inheres in the preservation of a building as it was initially constructed (i.e., with an expansive lawn beside it). Uncompensated landmark preservation seems to rest on this synergy. The Court in Penn Central embraced the view that “the preservation of landmarks benefits all New York citizens and all structures, both economically and by improving the quality of life in the city as a whole.” 438 U.S. at 134, 98 S.Ct. 2646. This broad language seems to redefine “reciprocity of advantage” in such a way that no government act could ever require compensation, as the afflicted owner would be a member of the taking polity and thus in receipt of offsetting advantages, artificially presumed to be adequate.
Apart from obliterating takings law, such a view has peculiarly perverse effects in the realm of historic preservation. Although such laws try to preserve for society the positive externalities created by buildings like Cathedral Mansions, inflicting the entire cost on the creator of the landmark (or his successor in interest) is bound to discourage investment in first-class design. Moreover, while insurance markets can achieve the risk-spreading (or anti-“demoralization”) goals that some attribute to the Takings Clause, compare Posner, Economic Analysis of Law, supra, *890at 58, they cannot offset non-compensation’s disincentive to good design. Historic landmark preservation, after all, is imposed selectively on those who went out of their way to secure architectural distinction. The higher the quality, the higher the premium for takings insurance; the disincentive is inescapable.
Having found that the lawn and apartment parcels should be treated as a unit, the majority nevertheless considers whether compensation would be due even if the lawn were analyzed separately; in doing so, it gratuitously takes an even harsher stance against compensation than does present law. The majority finds that District Intown has failed to offer evidence that the regulation denies it “economically viable use of [the] land,” Lucas, 505 U.S. at 1016, 112 S.Ct. 2886, even though the Mayor’s own agent found that “any construction that destroyed the lawn would be incompatible with the lawn’s status as an historic landmark.” Maj. Op. at 882. Thus, so long as the lawn is untouched, “economically viable” uses are permissible. It is hard to imagine what “economically viable” use that constraint leaves, unless the majority means that the very barest thread of value, yielded by some thoroughly bucolic use, is enough to defeat a total takings claim. By this standard, no regulation can ever effect a total taking, and at best will be tested only under the far weaker partial takings rubric.
The prevailing Federal Circuit-Claims Court method of defining the relevant parcel, followed by the panel here, focuses on marginal issues and largely overlooks the more critical concern of synergies; the focus on the landowner’s historical, rather than proposed, use further skews the analysis. But the Supreme Court’s general approach seems to militate in favor of looking to the parcel as a whole. Similarly, although resting uncompensated landmark preservation on the idea of reciprocal advantage stretches the concept into meaninglessness, and the denial of compensation discourages ex ante what it hopes to foster ex post, the current cases give these arguments little purchase. Accordingly, I concur in the majority’s decision to affirm.