STEPHEN POLASKY, HOLLY DOREMUS, and BRUCE RETTIG ^*

This paper analyzes the effect of alternative institutional arrangements on the conservation of endangered species and economic activity on private land. Because a landowner does not capture the full value of species conservation, her preferences on land use will not coincide with social preferences. Under current law, the landowner has incentives to invest in lowering conservation value and to deny access to regulators in order to prevent collection of information. Paying compensation corrects many of these perverse incentives. An alternative approach is to limit the ability of landowners to affect the regulatory outcome. Whether it is better to entice landowners to make socially efficient decisions by paying compensation or to limit the ability of landowners to affect outcomes through changes in the regulatory regime depends on both practical implementation difficulties and distributive justice considerations. (JEL Q2, K1)

A large fraction of species listed as threatened or endangered under the Endangered Species Act (ESA) inhabit private land, and some are almost totally dependent on private land for habitat ( The Nature Conservancy, 1993). Conflicts between economic activity and species conservation on private land have become widespread. In San Diego, Orange, and Riverside Counties in Southern California and in the Texas Hill Country near Austin, conflicts exist between preserving habitat for endangered species and development activity caused by rapidly expanding urban areas. The Pacific Northwest and the Southeast have witnessed conflicts over the choice between preserving habitat for endangered species and harvesting timber on private land. Other less publicized cases exist around the country as well ( Mann and Plummer, 1995).

These conflicts arise because the ESA allows the government to restrict private land use activities in order to protect species. In June 1995, the Supreme Court upheld federal regulations prohibiting significant adverse modification of habitat of threatened and endangered species on private land ( Babbitt v. Sweet Home Chapter of Communities for a Great Oregon, 1995). These regulations limit the ability of private landowners to engage in otherwise unobjectionable land uses such as timber harvest, crop cultivation, or construction. The current ESA does not provide compensation to landowners when such restrictions reduce the value of their property. There is pressure in Congress and elsewhere, however, to reduce the burden of environmental legislation in general, and the ESA in particular, on private landowners. One suggested reform is to require compensation for regulations that reduce property values by some threshold amount. Critics of this suggestion argue that landowners do not deserve compensation for fulfilling their duty to conserve.

This paper analyzes the effect of alternative institutional arrangements on the conservation of endangered species and economic activity on private land and discusses problems in obtaining an efficient solution when species conservation and economic activity on private land conflict. Species conservation may yield positive externalities, in which case a market solution will be inefficient. In theory, government intervention can lead to an efficient outcome. However, three potentially important elements may make this difficult: (i) incomplete information about the benefits of alternative land use decisions, (ii) the ability of a landowner to invest prior to regulatory intervention, and (iii) government failure, such as fiscal illusion. The latter two points have been extensively analyzed in the economic literature on takings ( Blume et al., 1984; Fischel, 1995; Fischel and Shapiro, 1989; Hermalin, 1995; Innes, 1995; Miceli and Segerson, 1994). Decision making under uncertainty and the incentive to acquire information prior to making a land use decision have received little attention in the takings literature to date. Hermalin (1995) considers cases with incomplete information. In his model, however, the information structure is exogenous and does not affect the outcome unless government failure occurs as well. The problems caused by incomplete and endogenous information play an important role in the species conservation context. Initially, little may be known about the conservation value of particular parcels of land. With some expense, though, information can be acquired about the biological and economic significance of a parcel. Polasky and Doremus (1995) analyze equilibrium information collection and land use decisions in a model without landowner investment.

The analysis also considers species conservation on private land with initially unknown conservation value, endogenous acquisition of information, and the potential for investment by landowners prior to a regulatory decision on land use. The current system creates perverse incentives for landowners to prevent information collection and to invest in ways that harm species. Requiring compensation to landowners for market value lost due to mandated conservation improves the efficiency of the outcome relative to the current system, but investment incentives may still yield an inefficient outcome. An alternative compensation scheme in which landowners are paid the conservation value of their property, rather than lost market value if they choose to conserve, yields an efficient outcome ( Hermalin, 1995; Polasky and Doremus, 1995). Nonetheless, practical difficulties limit its usefulness as a policy option. Changing the regulatory regime to reduce the ability of landowners to manipulate regulatory outcomes also can improve the outcome.

Fairness and other concerns may outweigh efficiency considerations in species conservation policy or in any public policy that affects private landowners. Both economists (e.g., Fischel, 1995 , p. 11) and courts ( Pruneyard Shopping Center v. Robins, 1980) note that government intrusion on private property rights primarily raises issues about the fairness of initial allocation and redistribution of entitlements. Accordingly, the analysis also considers the distributive consequences of the various approaches.

For simplicity, consider a single land parcel that may have value for both species conservation and economic activity. These values may be incompatible, as is the case with timber harvesting and conserving species that are dependent on old growth forests. If so, the parcel may be either conserved or developed. In a typical situation, the landowner can capture the value of economic activity but cannot capture more than a fraction of the conservation value. While she may be able to obtain some return from tourism or the sale of products derived from conserved species, the market provides no mechanism for her to collect non-use values, such as species existence value. An extensive literature examines why landowners fail to receive much return for conservation and discusses possible solutions to this problem (e.g., Defenders of Wildlife, 1993; McNeely, 1988). Conservation, then, is a positive externality, and a purely market outcome will be inefficient.

If species conservation and economic activity values are known by a benevolent government whose aim is to maximize value derived from the parcel, any number of policy interventions can yield an efficient outcome. A subsidy for conservation or a tax on economic ac

tivity set at the appropriate rate can yield an efficient outcome. Alternatively, the government could simply mandate the efficient choice through direct regulation of land use.

Real conservation policy decisions are not so easy. Incomplete and endogenous information, landowner investment prior to regulatory intervention, and government failure, all make achieving an efficient outcome difficult and make the form of conservation policy important.

Incomplete and endogenous information about the potential benefits of various land uses raise two issues. First, they may make establishing reasonably objective, verifiable conservation values for particular parcels difficult or impossible. Second, landowners and regulators may not collect an efficient amount of information about conservation value.

1. Establishing a Value for Conservation. Establishing a conservation value presents several challenges for both economists and biologists. In general, conservation values might encompass ecosystem as well as species concerns, though here we focus solely on the value of species survival. As a first step, biologists must predict the impact of alternative land uses on extinction risks. Population viability analysis (PVA) examines the probability of survival of a population under various conditions ( Boyce, 1992; Soulé, 1987). The simplest approach attempts to quantify the relationship between the probability of survival for some specified length of time (e.g., 100 years) and the amount of suitable habitat available. In general, greater amounts of habitat can support a larger population that will be less subject to extinction through chance genetic or environmental events. Some recent PVAs, for the spotted owl in particular, emphasize the importance of the spatial pattern as well as the overall amount of suitable habitat ( Lamberson et al., 1992; Lande, 1988). While PVA has made great strides, it does not yet reliably predict the probability of survival for a particular species given particular land use decisions. Biologists typically lack both the empirical data and the theoretical understanding necessary to determine those probabilities with any confidence.

Perhaps an even greater challenge awaits the economist. Assuming that one can quantify the risk to species survival as a function of alternative land use patterns, the next step requires measuring the social value of species persistence. There is an extensive literature on estimating non-market values including species value. A number of studies estimate species value using contingent valuation (see Loomis and White, 1996, for a recent summary). The usefulness and reliability of these estimates is subject to debate. Respondents to contingent valuation surveys may view species survival as a moral imperative not subject to monetary valuation ( Sagoff, 1988; Stevens et al., 1991). Respondents may not be familiar with the species they are asked to value or with the concept of assigning value to species existence. In addition, respondents may have difficulty separating species value from wilderness value, habitat preservation value, or generalized support for environmental causes.

It is interesting to note that current endangered species policy largely avoids the species valuation question. The ESA simply forbids any activity that will "harm" or otherwise "take" a listed species no matter how valuable that activity might be. There is an important exception to this restriction, though. A landowner who obtains an incidental take permit from the U.S. Fish and Wildlife Service (FWS) may engage in activities harmful to species. Furthermore, FWS must issue such a permit if the landowner shows that the proposed activity "will not appreciably reduce the survival and recovery of the species." One way to reconcile the nearly absolute prohibition of the ESA with an economic approach that compares the relative benefits of alternatives is to assign an extremely large, possibly infinite, cost to extinction. If extinction carries an extremely large cost, avoiding "harm" is nearly always the efficient solution. The central issue then is a biological determination of whether "harm" would result from development rather than an economic determination of whether development would yield positive net benefits.

While it ducks the valuation problem, the ESA's refusal to explicitly consider the costs and benefits of species conservation carries its own costs. Conservation policy that fails to take these factors into account may mandate inefficient outcomes. Further, even when expressly forbidden to do so, regulators often

compare net benefits of alternatives in hidden or disguised ways. Several observers point out examples where economic and political considerations appear to have affected decisions, such as listing, that are supposed to be based on purely biological considerations ( Bean, 1991 , p. 41; Houck, 1993 , pp. 285-296; Thomas and Verner, 1992 , p. 628). In other words, it may be impossible in practice, and undesirable in theory, to divorce conservation and land use policy from value judgments. The analysis that follows assumes that the regulator attempts to maximize net benefits; this is consistent with current law only when extinction costs are extremely large.

2. Acquiring Information about Conservation Value. As argued above, the impact of land use on conservation values initially is likely to be uncertain. Information concerning the impact of land use on species survival can reduce this uncertainty. Two potential efficiency concerns that arise with endogenous information are (i) whether the landowner will allow information gathering on her property and (ii) whether incentives exist to collect the efficient amount of information.

To block a proposed activity under the ESA, the government must demonstrate that a listed species uses the area for some essential behavior and that the proposed activity will interfere with this behavior. The government can gather some information about the presence of species on a particular parcel from satellite data, aerial surveying, or by visual inspecting neighboring parcels or roadways (off-site inspection). However, gathering the type of detailed information necessary to prove that an activity will cause harm to a listed species often requires on-site inspection of the property. Entry of government agents onto private property without the landowner's permission risks confrontation and often is prohibited by trespass laws or explicit statutory bans. For example, the Omnibus Consolidation Rescission and Appropriations Act of 1996 forbids the use of federal funds to conduct aerial biological surveys of private property without the landowner's consent. Therefore, the government often needs landowner cooperation to obtain the information it needs to regulate effectively. However, because the landowner may be hurt financially if the government obtains information about conservation value, she may have an incentive to deny regulators access to her property.

The second efficiency concern with endogenous information is whether the proper incentives to collect information exist. Imperfect information about conservation value might cause decision makers to err in either of two basic ways. Underestimating the importance of habitat to a species may allow development to proceed in spite of negative net benefits. On the other hand, overestimating the importance of habitat unnecessarily restrains development. Gathering information is beneficial because it can prevent either of these mistakes from occurring. However, gathering information is not costless. New information should be acquired if and only if the expected benefits exceed acquisition costs.

An additional issue arises in a dynamic framework with imperfect information and irreversible outcomes. Postponing a decision with irreversible consequences until more information can be acquired has a value, called option value or quasi-option value ( Arrow and Fisher, 1974). In the context of species conservation, the decision to develop may have irreversible consequences for two reasons. First, returning the land back to its original state may not be possible. Second, development may increase the probability of extinction, which forecloses future uses. Species with no known current value may turn out to be the next cure for cancer. There is a growing literature on the option value of preserving species for genetic prospecting (e.g., Mendelsohn and Balick, 1995; Polasky and Solow, 1995; Simpson et al., 1996).

The landowner's ability to invest prior to the development-or-conservation land use decision also may affect the efficiency of conservation policy. Landowners may use such investments strategically to influence the regulatory decision. Blume et al. (1984) show that landowners who are assured that the government will compensate them if it condemns their property may make socially wasteful investments. Species conservation policy may offer similar opportunities for strategic investment. Besides investing to increase market value, a landowner intentionally could harm an endangered species, perhaps eliminating it

from the property, in order to avoid future regulation. Although this would violate the ESA, potentially subjecting the landowner to civil and criminal sanction, the government has limited ability to monitor and prevent such anticonservation investment ( Polasky and Doremus, 1995). Species thought to be at risk but not yet listed present an even more tempting opportunity for investment in destruction of conservation values. Landowners who can successfully eliminate such species from their property can escape potential regulatory restriction without risking legal sanctions.

There are, of course, limits to landowners' ability to invest prior to regulatory oversight. Most significant land development activities require regulatory approval (e.g., a building permit). In areas known to harbor a listed species, regulators can screen such investments for harmful impacts. However, in areas where the regulators are not yet aware of the presence of listed species, landowners may have the ability to avoid regulation through strategic investment.

Economic models often assume costless regulatory oversight by a benevolent government, but this is not necessarily an accurate representation of reality. In simple economic models, transfer payments, such as compensating landowners for property value reduction due to environmental regulation, have no direct efficiency consequences. Efficient (firstbest) outcomes to regulatory problems can be achieved by making payments to the regulated parties that align their returns with social returns. However, the recent literature on regulation stresses that a real resource cost often is associated with government transfer payments. The government must raise money through some form of taxation, which creates deadweight loss. Citing work by Ballard et al. (1985) and Hausman and Poterba (1987), Laffont and Tirole (1993, p. 38) state that a "reasonable mean estimate" of the cost of government funds in the U.S. is 0.3, meaning that each $1 of government funds raised actually costs society $1.30. When government funds are costly, distribution becomes an issue even for a government strictly concerned with efficiency.

A different set of problems arises when the assumption of a benevolent government acting in society's best interest is relaxed. For a variety of reasons ranging from ideological conviction to agency growth or personal gain, regulators may have goals other than social welfare. Furthermore, government regulators typically have considerable leeway in implementing regulations. Legislation often is drafted imprecisely. Legislators have limited ability to oversee regulatory actions, and voters in turn have limited means of monitoring the actions of elected officials. Regulators, then, may have both the incentive and the ability to pursue their own agendas.

A well-known form of non-benevolence is fiscal illusion. Regulators (or voters) suffer from fiscal illusion when they do not give full weight to the costs imposed by regulation. Requiring regulators to compensate landowners for regulatory impacts can cure fiscal illusion by forcing regulators to take account of the full costs of regulation ( Blume et al., 1984). On the other hand, capture of agency personnel by regulated parties may cause over-sensitivity to regulatory costs imposed on those parties ( Stigler, 1971).

The analysis here develops a three stage model incorporating landowner investment, information collection, and land use decisions. In the first stage, the landowner may invest to increase the development value or to decrease the conservation value of the property. Investment activities either are not regulated by the government or are not adequately monitored if regulated.

Initially, neither the regulator nor the landowner knows the conservation value. The regulator and the landowner observe the probability distribution function for the conservation value, which depends on the level of investment at the first stage. In the second stage, either the regulator or the landowner can commission a biological survey. On-site inspection yields perfect information about the conservation value. On-site inspection requires the landowner's consent. Off-site inspection, which costs more than on-site inspection and yields less than perfect information, can occur without the landowner's permission.

In the final stage, a land use decision is made either to conserve the land or to allow development. In parts A-C, the regulator's objective is to maximize the expected social value of the property. With complete information, the regulator will mandate conservation if the conservation value is at least as large as the development value. With incomplete information, the regulator will mandate conservation if the expected conservation value is at least as large as the development value. The landowner's objective is to maximize her expected value from the property. The landowner would prefer to develop if the development value exceeds the fraction of the (expected) conservation value that accrues to her, which is assumed to be the case throughout.

Under current law, the regulator does not compensate the landowner when mandating conservation on her property. In the absence of information, the landowner can develop when the development value exceeds the expected conservation value. In this situation, information can only hurt the landowner. Information may show that the conservation value exceeds the development value, leading the regulator to forbid development. As long as a chance exists that conservation will be mandated with information, a landowner would find it in her best interest not to cooperate with information collection. Therefore, the landowner will not permit on-site inspection. The regulator may then choose to collect information via off-site inspection. When the development value exceeds the expected conservation value, information collection may be inefficient for two reasons. First, the regulator may choose not to collect information even though the benefits exceed the inspection costs. Second, the regulator may choose off-site inspection even though on-site inspection is superior.

Suppose that the expected conservation value is at least as large as the development value. If no information is collected, the regulator will mandate conservation. The landowner gains from information collection because it may show that the actual conservation value is less than the development value, in which case development can occur. The landowner's private expected benefit from information collection exceeds the social benefit because the landowner does not include the external conservation value as an opportunity cost of development. Therefore, the landowner may choose to collect information even though it is inefficient to do so.

Now consider the investment stage. Conditions exist in which the landowner will invest inefficiently. The incentive to over-invest either to increase the development value or to decrease the conservation value occurs for several reasons. By investing, the landowner might flip the inequality between the expected conservation and development values. Having the development value exceed the expected conservation value is advantageous to the landowner because when information is not collected, development is allowed. Also, by blocking on-site inspection, the landowner may prevent the regulator from collecting information, or even if the regulator pursues offsite inspection, that pursuit may be unsuccessful. In addition, when the inequality goes in this direction, the landowner assures herself that the regulator, not the landowner, pays for any information collection cost. Finally, if information is collected, investment increases the probability that the development value will exceed the actual conservation value so that development will be allowed. These results suggest that it may be rational for a landowner to invest in lowering the conservation value (i.e., to pursue a strategy of "shoot, shovel, and shut up") even though that strategy will never produce net social gains.

In sum, the current institutional structure may lead to inefficient information collection as well as inefficient patterns of investment. A number of commentators point out the perverse investment incentives (e.g., Defenders of Wildlife, 1993; Stroup, 1995). Because the landowner can capture the full development value but only a fraction of the conservation value, she will manipulate decisions under her control, such as whether to invest or cooperate with information collection.

Suppose the current legal regime was changed to require compensation equal to the decrease in fair market value whenever the regulator prohibited development. Compensation removes the monetary incentive for the

landowner to oppose conservation because both development and conservation give the same payoff. Therefore, the landowner will not have an incentive to prevent on-site inspection of the property. Of course, the landowner will not have any incentive to pay for a survey either. The regulator, however, may want to pay for a survey. With on-site inspection, the regulator will make an efficient information collection decision. Paying compensation corrects for distortions at this stage.

Further, this compensation scheme will reduce the perverse investment incentives created by the current system. Specifically, under this scheme the landowner will not have an incentive to invest so as to influence the regulatory decision. Since the landowner expects the same payoff with conservation or development, she has no incentive to manipulate these values. In particular, the landowner will not invest to lower conservation value by destroying habitat or species. However, the landowner still has an incentive to invest inefficiently in increasing development value. Tying landowner payment to market value even when conservation is efficient gives the landowner an incentive to undertake socially wasteful investment in order to increase payments from the regulator ( Blume et al., 1984).

A slightly different compensation scheme can correct the inefficiencies of market based compensation. Compensation for the portion of conservation value not already captured by the landowner aligns the landowner's payoff with social payoffs, giving the landowner the correct incentive to make efficient investment, information collection, and land use decisions. Hermalin (1995) and Polasky and Doremus (1995) propose this approach, which also is discussed in the regulation literature ( Laffont and Tirole, 1993).

Despite its simplicity and intuitive appeal to economists, policymakers have not seriously considered this approach for two reasons, one practical and one philosophical. On a practical level, basing compensation on the portion of conservation value not captured by the landowner requires an objective method for measuring that value. Quantifying total conservation value, much less the portion that is external to the landowner, is a heroic task. The cost of determining this value, the difficulty of getting both the regulator and the landowner to agree, and the expense of litigation if the parties fail to agree, make it impractical as a measure of compensation. Philosophically, this compensation scheme also may strike some as unjust because it can result in large payments to landowners for fulfilling what may be perceived as a social duty to conserve.

Current law regulating actions on private lands to protect species is likely to lead to inefficient outcomes for several reasons. The landowner may have an incentive to prevent the regulator from collecting biological information. Further, the landowner may have an incentive to take actions to lower the conservation value or to inflate the development value of her land beyond the efficient level.

Inefficient outcomes occur when (i) the landowner's payoffs differ from the social payoffs and (ii) the landowner has the ability to manipulate the outcome. Compensation for reductions in market value corrects some inefficiencies of the current system but fails to fully align landowner and social payoffs. Basing compensation on uncaptured conservation value aligns landowner and social payoffs, yielding an efficient solution. However, it presents perhaps insurmountable practical difficulties.

Rather than trying to align payoffs, an alternative approach is to limit the ability of landowners to manipulate the regulatory outcome. Changing current law to allow regulators the choice to conduct an on-site inspection yields an efficient solution at the information collection stage. This change alone, however, could increase investment inefficiency by increasing landowner incentives to alter development or conservation values prior to the regulatory decision. In order to effectively limit the ability of landowners to manipulate outcomes, regulatory decisions must be based on factors that are difficult for landowners to manipulate. In the species conservation context, regulators might assess the land's potential suitability for listed species rather than the current presence or absence of those species. Suitability might be judged on the basis of soil type, topography, and climate conditions -- fac

tors largely outside landowners' control and often determinable through off-site inspection.

A different set of concerns arises with government failure. Compensation schemes necessitate increased use of government funds. Multiplying the expected compensation bill by 0.3, assuming the cost of government funds cited in section II is roughly correct, yields the direct resource cost of requiring compensation. This amount may be quite large. On the other hand, as pointed out in the literature on takings, compensation can overcome fiscal illusion.

The distribution of costs and benefits between landowners and taxpayers varies widely in different regulatory regimes. At one extreme, the burden of species conservation falls entirely on landowners unlucky enough to host rare species. Alternatively, taxpayers could bear the full burden of species conservation. In light of the broad range of possible distribution outcomes, policymakers should consider the distributive justice implications of their choices. They should ask what proportion of species protection costs landowners and taxpayers should bear, and whether those costsharing rules should apply equally to all landowners or should vary with individual circumstances.

Fair allocation of species conservation costs depends upon the initial assignment of rights and responsibilities. If development is a right inherent in land ownership, taxpayers should pay for its withdrawal by the government. On the other hand, if the continued existence of species is a public right, landowners who fail to fulfill their duty to conserve should compensate the public.

This distinction between the rights and responsibilities of landowners is familiar in the law of takings. Courts historically have employed a harm-benefit test to decide who should bear the costs of land use restrictions ( Pennsylvania Coal Company v. Mahon, 1922). This test rests on the assumption that land ownership does not carry a right to harm one's neighbors or the public at large through nuisance activities. Rather, the landowner has a legal duty to refrain from such activities. Accordingly, the government need not compensate for prohibiting harmful actions. By contrast, the public cannot demand that a landowner use her land in such a way as to confer a public benefit. To get such a benefit, the public must buy it through compensation.

The harm-benefit test appears equitable because it imposes costs on the party responsible for causing the harm. It suffers from a serious weakness, however: distinguishing harms from benefits often is difficult ( Coase, 1960; Sax, 1964). This weakness led the Supreme Court, in Lucas v. South Carolina Coastal Council ( 1992 ), to reject the harm-benefit test as a constitutional touchstone for determining whether compensation is required.

Species conservation vividly illustrates the difficulties of separating harms from benefits because society has not reached a moral consensus on the treatment of non-human species. Most people do not regard the kinds of land uses likely to destroy species habitat as intrinsically wrong. Yet, many would regard extinction as a harm to the public, whether for intrinsic or utilitarian reasons. Species conservation and economic development of land are "perfectly innocent and independently desirable uses" ( Sax, 1964 , p. 49) that sometimes conflict. Without a firmer moral consensus, these conflicts cannot be resolved by the harm/benefit test.

The Supreme Court in Lucas adopted state law nuisance standards as an alternative to the harm/benefit test. However, state law typically defines a nuisance as an activity unreasonably harmful to the enjoyment of conflicting private or public rights. Looking to nuisance law does not, therefore, avoid the harm-benefit conundrum.

Policymakers also may worry about the intensity of impacts on particular landowners. Allowing the government to impose speciesprotection requirements without compensation could deprive some landowners of much of the value they expect to realize from their land, which may constitute a large proportion of their total wealth. For example, timber harvest restrictions imposed to protect the northern spotted owl could preclude a large timber company from harvesting two-thirds of its exten

sive landholdings. The same restrictions might entirely prevent a retired couple from harvesting their 40-acre woodlot. The timber company would lose millions of dollars. Nonetheless, it would still have a substantial amount of harvestable timber. The overall value of that timber might even rise if harvest restrictions increase timber prices. By contrast, the retired couple will lose much less in absolute terms, but the loss may represent nearly all of their wealth.

The Supreme Court has been considerably more sympathetic to the claims of persons like the retired couple, whose property loses virtually all economic value because of government regulation. Government must compensate for such impacts unless the nuisance exception explained above applies ( Lucas v. South Carolina Coastal Council, 1992 , pp. 1,027-1,030). In recent years, the government agencies also have been sympathetic to the claims of small landowners (see U.S. Department of Interior, Fish and Wildlife Service, 1995a, 1995b). It is difficult, however, to defend a rule requiring compensation only if all economically beneficial use is denied. Even the Supreme Court has noted the apparent arbitrariness of automatically awarding compensation to a landowner who has lost all economic value but not to a neighbor whose property value has been reduced 95% (See Lucas v. South Carolina Coastal Council, 1992 , p. 1,016, n. 7).

Furthermore, whether a regulation denies all economically viable use presents an intractable definitional problem. The extent of diminution cannot be determined without defining the relevant property. With respect to timber harvest restrictions, the relevant property could be the specific trees the landowner can no longer harvest, the land parcel as mapped in the assessor's office, all the owner's holdings in the area, all the owner's assets, or some other slice of the owner's pie. No clear basis exists for defining the relevant property interest ( Sax, 1964; McLaughlin, 1995).

Yet, for all its logical difficulties, this test also persists. Like the harm-benefit distinction, it appeals to an intuitive sense of justice. While a government regulation that imposes relatively small financial costs on a large number of people may not be universally loved, most will regard it as fair. A regulation imposing heavy costs on a select group seems less fair. It presents the appearance of the tyranny of the majority. Even if such a regulation is efficient in the sense that total gains outweigh total losses, sacrificing virtually all the resources of a few persons to achieve relatively small gains for a large number may seem unjust.

Land use restrictions may deprive landowners of the opportunity to realize anticipated value from their property. Economic use of real property often requires long-term investment. Owners contemplating such investments understandably desire long-term security. They seek assurances that the legal rules will not change before their investment bears fruit ( Freyfogle, 1995 , p. 127). When a shift in public values precludes realization of landowner expectations, policymakers must decide whether the public or the landowner should bear the costs of the change.

The Supreme Court considers a property owner's "investment-backed expectations" an important factor in determining whether a regulation requires compensation ( Penn Central Transportation Company v. City of New York, 1978). The inclusion of expectations reflects the common view that persons who made investment decisions guided by the old rules should not suffer when the rules change. Allowing the government to change the rules without offering compensation seems particularly unfair if the government had actively encouraged behavior that is now prohibited.

Of course, expectations are not legal entitlements. Perhaps landowners should anticipate changes in land use regulations, which have never been static (Sax, 1993; McElfish, 1994). Requiring landowners to assume the risk that a rule change may alter permissible uses should encourage both adaptive behavior on the part of landowners (Sax, 1993, p. 1,449) and market price reflection of regulatory risks ( Mandelker, 1995). On the other hand, some regulatory changes are difficult to foresee. The more drastic the regulatory about-face, the more unfair it may appear to impose its costs on landowners. Since the need for restriction of private land uses to protect species has only recently been recognized, it may be unfair to expect landowners to have anticipated species conservation regulations in making their investment decisions.

Whether fairness requires compensation for the impacts of species conservation regulations depends on whether species protection is viewed as a duty or land development is considered a right. Landowners who consider species conservation a benefit rather than a right will object to the current regime, which imposes the bulk of the costs of conservation on landowners. Their objections could be answered by a compensation scheme like those discussed in section III. However, those who consider conservation a duty owed by every landowner to the public may object to compensation. But even they might endorse exceptions in cases of disproportionate cost or unanticipated rule changes. Current law addresses the issues of disproportionate costs and unanticipated rule changes through the ad hoc case-specific law of takings, which ameliorates some of the harsh impacts of the general no-compensation rule but inevitably does so by drawing lines that are difficult to justify. Reducing the ability of landowners to evade regulation would be most attractive to those who view species conservation as a duty of landowners. This approach also would leave the problems of disproportionate impact or unanticipated regulation to constitutional takings law.

In considering species conservation on private land, policymakers must take into account two valuable and potentially conflicting goals, economic development and species conservation. Because a landowner typically does not capture the full value of species conservation, her preferences on land use will not always coincide with social preferences. Further, even with land use regulations in place, a landowner will retain some control over information availability through access restrictions and over development and conservation values through investment choices. The landowner has both the incentive and the ability to take actions that yield inefficient outcomes. Under current law, the landowner has incentives to invest in lowering conservation value and to deny access to regulators in order to prevent collection of information.

Two approaches could make for more efficient conservation policy on private land. One is to align landowner and social preferences through payment of compensation equal to the portion of conservation value the landowner is otherwise unable to capture. If this is infeasible, compensating for lost market value will correct many of the current perverse incentives. A second approach is to limit the ability of landowners to affect the regulatory outcome by authorizing on-site inspection by regulators of private land or basing land use decisions on factors not easily manipulated by landowner investment, such as soil type, topography, and climate. Both points are likely to be very contentious, especially with the rise of the property rights movement. It is worth noting that basing land use decisions on factors less prone to manipulation simultaneously would reduce the need for on-site information collection since regulators are more likely to examine historical data or use remote sensing to determine soil type, topography, and climate data.

Is it better to entice landowners to make socially efficient decisions by paying compensation or to limit the ability of landowners to affect outcomes through changes in the regulatory regime? The answer depends on both practical implementation difficulties and distributive justice considerations. Compensating landowners at all will seem unjust to those who view species conservation as a duty. Some of this group, however, may drop their objections to compensation if they are convinced it will lead to a more effective (rather than simply more efficient) species conservation policy. Given the current ability of landowners to evade legal sanctions for activities harmful to species, carrots may indeed be more effective than sticks. A different group will see allowing searches of private property or imposing new regulations that deprive landowners of substantial value without compensation as unjust. They are not likely to change that view unless they can be persuaded of the importance of species conservation.

Choosing a species conservation policy for private land is difficult because both efficiency and fairness concerns are raised, and the two may not be readily reconciled. Fairness (or justice) appears the more intractable problem because it cannot be separated from value judgments. Perhaps the role of economists in this situation is to point out the various alternative means of achieving a more efficient result, to acknowledge the consequences of those alternatives for fairness, and to await a societal judgment on the appropriate choice.

Arrow, K., and A. Fisher, "Environmental Preservation, Uncertainty, and Irreversibility", Quarterly Journal of Economics, 88:2, 1974 , 312-319.

Ballard, C., J. Shoven, and J. Whalley, "General Equilibrium Computations of the Marginal Welfare Costs of Taxes in the United States", American Economic Review, 75:1, 1985 , 128-138.

Bean, M., "Looking Back over the First Fifteen Years," in Balancing on the Brink of Extinction: The Endangered Species Act and Lessons for the Future, K. A. Kohm , ed., Island Press, Covelo, Calif., 1991 .

Blume, L., D. Rubinfeld, and P. Shapiro, "The Taking of Land: When Should Compensation be Paid?" Quarterly Journal of Economics, 99:1, 1984 , 71-92.

Boyce, M. "Population Viability Analysis", Annual Review of Ecology and Systematics, 23, 1992 , 481-506.

Coase, R., "The Problem of Social Cost", Journal of Law and Economics, 3:1, 1960 , 1-44.

Defenders of Wildlife, Building Economic Incentives into the Endangered Species Act, 2nd Edition, Defenders of Wildlife, Washington, D.C., 1993 .

Fischel, W., Regulatory Takings: Law, Economics, and Politics, Harvard University Press, Cambridge, 1995 .

Fischel, W., and P. Shapiro, "A Constitutional Choice Model of Compensation for Takings", International Review of Law and Economics, 9:2, 1989 , 115-128.

Freyfogle, E., "The Owning and Taking of Sensitive Lands", U.C.L.A. Law Review, 43:1, 1995 , 77-138.

Hausman, J., and J. Poterba, "Household Behavior and the Tax Reform Act of 1986", Journal of Economic Perspectives, 1:1, 1987 , 101-119.

Hermalin, B., "An Economic Analysis of Takings", Journal of Law, Economics, and Organizations, 11:1, 1995 , 64-86.

Houck, O., "The Edangered Species Act and Its Implementation by the U.S. Departments of Interior and Commerce", Colorado Law Review, 64:2, 1993 , 277370.

Innes, R., "An Essay on Takings", Choices, First Quarter 1995 , 4-7, 42-44.

Laffont, J-J., and J. Tirole, A Theory of Incentives in Procurement and Regulation, MIT Press, Cambridge, 1993 .

Lamberson, R., R. McKelvey, B. Noon, and C. Voss, "A Dynamic Analysis of the Northern Spotted Owl Viability in a Fragmented Forest Landscape", Conservation Biology, 6:4, 1992 , 505-512.

Lande, R., "Demographic Models of the Northern Spotted Owl (Strix Occidentalis Caurina)", Oecologia, 75:3, 1988 , 601-607.

Loomis, J., and D. White, "Economic Benefits of Rare and Endangered Species: Summary and Meta-Analysis", Ecological Economics, 18:3, 1996 , 197-206.

Mandelker, D., "Investment-Backed Expectations in Taking Law", Urban Lawyer, 27:2, 1995 , 215.

Mann, C., and M. Plummer, Noah's Choice: the Future of Endangered Species, Alfred A. Knopf, New York, 1995 .

McElfish, J., Jr., "Property Rights, Property Roots: Rediscovering the Basis for Legal Protection of the Environment", Environmental Law Reporter, 24:5, 1994 , 10,231.

McLaughlin, J., "Majoritarian Theft in the Regulatory State: What's a Takings Clause For?" William and Mary Environmental Law and Policy Review, 19:2, 1995 , 161, 195-198.

McNeely, J., Economics and Biological Diversity: Developing and Using Economic Incentives to Conserve Biological Diversity, IUCN, Gland, Switz., 1988 .

Mendelsohn, R., and M. Balick, "The Value of Undiscovered Pharmaceuticals in Tropical Forests", Economic Botany, 49:12, 1995 , 223-228.

Miceli, T., and K. Segerson, "Regulatory Takings: When Should Compensation Be Paid?" Journal of Legal Studies, 23:2, 1994 , 749-776.

The Nature Conservancy, Perspective on Species Imperilment: A Report from the Natural Heritage Data Center Network, The Nature Conservancy, Arlington, Va., 1993 .

Polasky, S., and H. Doremus, "When the Truth Hurts: Endangered Species Policy on Private Land with Incomplete Information", Department of Agricultural and Resource Economics, Oregon State University, 1995 .

Polasky, S., and A. Solow, "On the Value of a Collection of Species", Journal of Environmental Economics and Management, 29:3, 1995 , 298-303.

Sagoff, M., The Economy of the Earth, Cambridge University Press, Cambridge, 1988 .

Sax, J., "Takings and the Police Power", Yale Law Journal, 74:1, 1964 , 36-76.

_____, "Property Rights and the Economy of Nature: Understanding Lucas v. South Carolina Coastal Council", Stanford Law Review, 45:3, 1,433-1,455.

Simpson, D., R. Sedjo, and J. Reid, "Valuing Biodiversity for Use in Pharmaceutical Research", Journal of Political Economy, 104:1, 1996 , 163-185.

Soulé, M., ed. Viable Populations for Conservation, Cambridge, Cambridge University Press, 1987 .

Stevens, T., J. Echeverria, R. Glass, T. Hager, and T. More, "Measuring the Existence Value of Wildlife: What Do CVM Estimates Really Show"? Land Economics, 67:4, 1991 , 390-400.

Stigler, G., "The Theory of Economic Regulation", Bell Journal of Economics and Management Science, 2:1, 1971 , 137-146.

Stroup, R., The Endangered Species Act: Making Innocent Species the Enemy, PERC Policy Series Issue Number PS-3, 1995 .

Thomas, J., and J. Verner, "Accommodation with SocioEconomic Factors under the Endangered Species Act -- More than Meets the Eye", Transactions of the North American Wildlife and Natural Resource Conference, 57, 1992 , 627-641.

U.S. Department of Interior, Fish and Wildlife Service, "Endangered and Threatened Wildlife and Plants; Proposed Special Rule for the Conservation of the Northern Spotted Owl on Non-Federal Lands", Federal Register, 60:33, 1995 a, 9,484-9,516.

_____, "Endangered and Threatened Wildlife and Plants; Proposed Rule Exempting Certain Small Landowners and Low-Impact Activities From Endangered Species Act Requirements for Threatened Species", Federal Register, 60:139, 1995 b, 37,419-37,423.

Endangered Species Act, 16 U.S.C. §1531-1544.

Omnibus Consolidated Rescissions and Appropriations Act of 1996, Public Law 104-134, 110 Stat. 1321 ( 1996 ).

Babbitt v. Sweet Home Chapter of Communities for a Great Oregon, 515 U.S. 687, 1995 .

Lucas v. South Carolina Coastal Council, 505 U.S. 1003, 1992 .

Penn Central Transportation Company v. City of New York, 438 U.S. 104, 124, 1978 .

Pennsylvania Coal Company v. Mahon, 260 U.S. 393, 422, 1922 ( Brandeis, J., dissenting).

Pruneyard Shopping Center v. Robins, 447 U.S. 74, 83, 1980 .