Physical and Regulatory Takings

One Distinction Too Many

Richard A. Epstein *

At this moment, it looks as though the law of eminent domain takings is in a quiet phase, as the Supreme Court has not recently taken any major case that examines the foundations of the field. One apparently settled area of takings jurisprudence deals with rent control, where the Court provides only scant protection to landlords who claim that their property has been taken when states and local governments pass local laws that restrict their right to evict tenants at the expiration of their leases. For example, a recent challenge to New York’s rent control law received a polite dust-off in the Second Circuit in Harmon v. Kimmel. 1 Harmon v. Markus, 412 F. App’x 420 (2d Cir. 2011). The head of New York City’s Rent Guidelines Board declined to answer Harmon’s petition for certiorari in the United States Supreme Court. Much to everyone’s surprise, the Supreme Court requested an answer from the New York City defendants, which is now scheduled to be filed by March 5, 2012. 2 The Supreme Court docket for the case, now under the name Harmon v. Kimmel, is available at (last visited Feb. 13, 2012). Perhaps this surprising development is attributable in part to the extensive and sympathetic coverage that Harmon’s plight has received in the press. 3 See, e.g., Anemona Hartocollis, A Landlord’s Uphill Fight to Ease Rent Restrictions, N.Y. Times (Dec. 19, 2011),; Nicole Gelinas, A Man’s Home is the Government’s Castle, N.Y. Post (Jan. 15, 2012), opedcolumnists/man_home_is_the_government_castle_lAFJ8CnFxUel212XI32t5M; Editorial, Take This Case, N.Y. Daily News (Dec. 8, 2011),; Richard A. Epstein, Rent Control Hits the Supreme Court, Wall St. J. (Jan. 4, 2012)

Whether the Supreme Court grants certiorari or not, this case highlights the simple proposition that most important doctrinal decisions are made with reference to cases that have simple and recurrent fact patterns that raise major questions of principle. There are two key subtexts of the rent control cases. First, how do the Supreme Court’s takings decisions deal with divided interests in land? Second, how does the Court deal with the now-unquestioned distinction between physical and regulatory takings? This brief article addresses both of these ever-timely issues.

Private Law: Divided We Gain

One desirable maxim in legal theory is that every distinction in legal doctrine has to pay its own way. This rebuttable presumption against the proliferation of legal categories rests on a simple proposition: it is costly to draw any line, physical or conceptual. This maxim asks only that government officials identify the benefits from drawing any line. That burden is easily met by the common device of good fences. These demarcate property that allows us to know who has the right to exclude others, which in turn allows that owner to plant crops or build houses and factories. In the public sphere, lane lines and dividers allow traffic to run freely and safely on public roads. In cooperative ventures, clear lines make it possible to play baseball, basketball, football, hockey, or soccer. Was the player in bounds when he caught the pass? Did the ball or puck cross the goal line? By and large, boundary lines generate great benefits at low administrative cost. Their widespread persistence needs no social justification.

The benefits of any mature property rights system are not, however, exhausted by creating sharp boundary lines. Single owners often divide their own property voluntarily by making partial transfers to third parties. In most cases, these transactions work hard to draw lines that approximate the clarity of boundary lines between strangers. A lease sets a temporal boundary between landlord and tenant, but only when the gain from splitting that asset in two exceeds the cost of creating two interests in “the” property. Voluntary transfers also allow for the creation of mineral rights and air rights, mortgages, restrictive covenants, and easements. Sound public policy both encourages and facilitates consensual property divisions in property rights, which in turn make leases, mortgages, cotenancies, easements and covenants possible. A good recording system gives notice of these changes in title in ways that avoid frauds and direct potential buyers and lenders to the right persons with whom to deal. As the net gains from further divisions diminish, this benign process draws naturally to a close, wholly without state intervention.

Public Law: United We Fall

Unfortunately, modern takings law is in vast disarray because the Supreme Court deals incorrectly with divided interests under the Takings Clause of the Fifth Amendment, which reads: “nor shall private property be taken for public use, without just compensation.” 4 U.S. Const. amend. V. The Supreme Court’s regnant distinction in this area is between physical and regulatory takings. In a physical taking, the government, or some private party authorized by the government, occupies private land in whole or in part. In the case of a per se physical taking, the government must pay the landowner full compensation for the value of the land occupied. Regulatory takings, in contrast, leave landowners in possession, but subject them to restrictions on the ability to use, develop, or dispose of the land. Under current law, regulatory takings are only compensable when the government cannot show some social justification, broadly conceived, for its imposition. 5 See, e.g., Lingle v. Chevron U.S.A. Inc., 544 U.S. 528 (2005).

Thus, under current takings law, a physical occupation with trivial economic consequences gets full compensation. In contrast, major regulatory initiatives rarely require a penny in compensation for millions of dollars in economic losses. The distinction has been defended on the ground that the Court’s cases have consistently offered higher protection to physical takings given the historical importance of protection against occupation. It is also, as Justice Thurgood Marshall wrote, that a physical taking “is perhaps the most serious form of invasion of an owner’s property interests. To borrow a metaphor, the government does not simply take a single ‘strand’ from the ‘bundle’ of property rights: it chops through the bundle, taking a slice of every strand.” 6 See Loretto v. Teleprompter Manhattan CATV Corp., 458 U.S. 419, 435 (1982) (citation omitted).

This metaphor notwithstanding, the arguments are not sufficient to defend the distinction. Quite simply, the distinction between physical and regulatory takings does not pay its own way, which becomes more evident when we ask two key questions: Why draw this line? How should we draw this line?

The per se protection for physical takings guards against the political risk that greedy neighbors will use the political process to strip their neighbors of their property by brute force. Some people exhibit a stout moral resistance to political confiscation by majority vote. Even those strong social expectations, taken by themselves, rarely offer sufficient protection to isolated property owners. Placing the government under a price system (through the payment of just compensation) increases the odds that takings will only occur when their social gains exceed their social costs. If governments always acted with good motives and full knowledge, the protection would hardly be required. But self-interested actors rarely show that measure of self-restraint—so socially, it is worth bearing the heavy costs of running a compensation system.

The identical forces of self-interest are at work with restrictions on use and development of land. Why then assume that the process will usually work well? These restrictions reduce land values to owners by amounts that are likely to exceed the net gains to outsiders. Once again, a compensation requirement can reduce the political risks of allowing the taking of partial interests in property without any compensation at all. Why then do we not lump regulations with occupations, given the parallel political risks? Do we want the owner who is entitled to build a shopping mall on the east side of the road to work overtime to block the construction of a competitive shopping mall on the west side of the road? Nor are the social losses from these maneuvers reduced if all incumbent owners on the east side of the road work in concert to block like construction by all the owners on the west side of the road. Quite the opposite: the aggregation of winners and losers only magnifies the political risks. The categorical distinction between physical and regulatory takings ignores all the relevant similarities, and it necessarily increases the cost of administering the system, especially in the puzzling cases close to the line.

Ideally, the uniform theory of takings plays out the same way for government occupations and government regulations. To see why, consider two pairs of polar opposite cases. In the first, the same large-lot restriction is imposed on all landowners within a class, but that restriction only benefits individuals outside the regulated class. The generalization of this regulation only compounds the mistake. What Justice Holmes rightly called “average reciprocity of advantage” only occurs in the second cases when everyone subject to a restriction benefits from it, as with many exterior design restrictions. 7 See Pa. Coal Co. v. Mahon, 260 U.S. 393, 415 (1922). In this context, when the losses and gains are totaled, if each person is a gainer, then the constitutional command for just compensation is satisfied. Similarly, albeit less frequently, when part of the land is taken equally from many owners to construct a road that benefits all, the retained land of each is also worth more than it was before. The analysis of regulation thus applies to occupation. But if that occupation is skewed in one direction, it does not. The correct analysis of any takings of divided interests in land looks at both the benefit and the cost side for all physical and regulatory takings.

Drawing the Line in the Wrong Place

Against this backdrop, any effort to distinguish between the two cases leads to doctrinal confusion, as the following brief account of rent control, support easements, zoning, and landmark preservation statutes shows.

In Block v. Hirsh, 8 256 U.S. 135 (1921). the Supreme Court rejected a takings challenge to the rent control scheme put in place in Washington, D.C. just after the end of World War I. In a five-to-four decision, Justice Holmes held that the ordinance was a temporary wartime measure that might not withstand scrutiny if implemented on a permanent basis. “The only matter that seems to us open to debate is whether the statute goes too far.” 9 Id. at 156. “Too far” does not explain how far is too far. When a clear line is desperately needed, a weak test of degree is offered instead, and it does not explain why any rent control statute endures in times of prolonged peace. Indeed, the New York rent stabilization law is renewed like clockwork, 10 N.Y. State Tenants & Neighbors Coal., Tenants & Neighbors: Fact Sheet on Rent Law Renewal, Int’l Union of Tenants (Dec. 2010),

so that all the middle cases on the continuum survive quite clearly, even though in theory these matters of degree should only go to the extent of the taking, not to its existence.

The wartime rationale had worn thin by 1993, when in Yee v. City of Escondido, Justice Sandra Day O’Connor, writing for seven Justices, noted that a rent control ordinance “does not effect a physical taking in the first place.” 11 503 U.S. 519, 532 (1992). Why? “Because they voluntarily open their property to occupation by others, petitioners cannot assert a per se right to compensation based on their inability to exclude particular individuals.” 12 Id. at 531. Clearly, the line between physical and regulatory takings now departs from its original meaning. Rent control statutes don’t restrict how landlords use their property. Rather, they keep the tenants in possession of the landlord’s property indefinitely. Justice O’Connor’s short description thus ignores the temporal dimension of property. The landlord consented to admit the tenant only for the duration of the lease. Why the continued occupation of the land after the expiration is not a physical taking is left to the imagination. Instead of getting the fee simple on the lease’s expiration that is worth $X, the landlord first gets a below-market rent, coupled with worthless rights to evict the tenant for nonpayment (which almost never happens) or to convert the property to some other use over, of course, determined resistance before the local zoning board. The new bundle of rights is worth far less than the old one, which is why tenants protect them with their lives. The entire rent control system rests on the most rickety of foundations.

The same failure to appreciate divided interest also surfaced with mineral interests. In 1922, one year after Block, Justice Holmes again appealed to the “too far” language in Pennsylvania Coal Co. v. Mahon. 13 260 U.S. 393, 415 (1922). For a detailed account of the facts, see William A. Fischel, Regulatory Takings: Law, Economics, and Politics 13-63 (1995). In that case, the coal company had retained a support estate when it conveyed the surface estate to owners who took the risk of subsidence from the company’s mining operation. Justice Holmes held that the state of Pennsylvania owed compensation to the holders of these mineral interests when it forced the mineral owner to convey the support interest to the surface owner free of charge. 14 See generally Pa. Coal Co., 260 U.S. at 393. The ad hoc “too far” language explains nothing, for the statutory scheme challenged in Pennsylvania Coal constituted confiscation of a recognized partial interest in land for which there was, of course, no return consideration. All surface owners benefited, and all mineral owners were hurt. It is a matter of indifference whether the case involves government occupation or regulation. If the mineral rights holder had to leave coal in place, is there a physical taking if the government never mines the coal itself? Is it a regulatory taking, governed by wholly different principles, if the mineral owner can take what coal he wishes, so long as it still shores up the foundations? The proper approach ignores these refinements by requiring in both cases compensation equal to the diminution in value of the mineral estate.

The same error was repeated in the zoning context in the 1926 decision of Village of Euclid v. Ambler Realty Co., which sustained a zoning decision that had been held unconstitutional by the district court. 15 272 U.S. 365 (1926). Here, the partial interest was a restrictive covenant on land use, in which the local government placed sharp restrictions on the unified use of a single sixty-eight-acre plot of land which the plaintiff alleged reduced its value, perhaps by as much as eighty percent. The truth of that allegation has been disputed, but the case was decided on the assumption that it was true, as it is in fact true for many modern zoning schemes. As a matter of private law, these restrictive covenants, like easements, are servitudes, which private parties must acquire by purchase. But if these covenants are embodied in a zoning ordinance, they count as mere restrictions on use, which are rarely compensable. But again, in Euclid, all the restrictions were on the one isolated owner, for the benefit of some subclass of neighbors. Euclid did not sport a set of reciprocal covenants (such as those found in many planned unit developments) that often generate in-kind compensation to all owners. Euclid’s false step has marred zoning law to the current day. The want of a price system through the just compensation system leads to systematic overregulation relative to the common good.

The same conceptual misstep was repeated fifty-two years later in Penn Central Transportation Co. v. New York City. 16 438 U.S. 104 (1978). Justice Brennan upheld a landmark preservation ordinance that prevented the construction of a new Breuer tower over Grand Central Terminal. Justice Brennan’s technique let the city prevail when the relevant interests were balanced, while suppressing that air rights may be used, sold, mortgaged, or leased under state law. Justice Brennan articulated the strong distinction between physical and regulatory takings without asking whether the loss of these air rights was a physical taking if the city did not use these air rights, but only blocked their use by the former owner. The correct approach compensates for the diminution in market value no matter which side of the line the case is on. Nonetheless, Justice Brennan undercut the economic protection of divided interest by asking only if the “parcel as a whole” had retained sufficient benefit. 17 See id. at 130-31. He thus plunged the law into needless complexity when property is subdivided before a regulation is imposed, if the regulation wipes out one party’s interest but leaves the other intact. His rule first creates genuine inequities and lets the government reduce air rights to zero or very little value, resulting in excessive overclaiming as local taxpayers force the brunt of the economic losses upon the landowner.


The judicial application of takings law to these four different partial interests in land thus destroys the social value created by private transactions that create multiple interests in land. The unprincipled line between occupation and regulation is then quickly manipulated to put rent control, mineral rights, and air rights in the wrong category, where the weak level of protection against regulatory takings encourages excessive government activity. The entire package lets complex legal rules generate the high administrative costs needed to run an indefensible and wasteful system. There are no partial measures that can fix this level of disarray. There is no intellectual warrant for making the categorical distinction between physical and regulatory takings, so that distinction should be abolished. A unified framework should be applied to both cases, where in each case the key question is whether the compensation afforded equals or exceeds the value of the property interest taken. The greatest virtue of this distinction lies not in how it resolves individual cases before the courts. Rather, it lies in blocking the adoption of multiple, mischievous initiatives that should not have been enacted into law in the first place. But in the interim, much work remains to be done. A much-needed first step down that road depends on the Supreme Court granting certiorari in Harmon v. Kimmel.