Allocating Superfund Costs: Cleaning Up the Controversy

23 ELR 10133 | Environmental Law Reporter | copyright © 1993 | All rights reserved

23 ELR 10133 | Environmental Law Reporter | copyright © 1993 | All rights reserved

Allocating Superfund Costs: Cleaning Up the Controversy
John C. Butler III, Mark W. Schneider, George R. Hall, and Michael E. Burton
Editors' Summary: To promote prompt cost recovery and equitable allocation of liability for remediation costs assessed under CERCLA, Congress granted liable parties the right to sue other potentially responsible parties (PRPs) for contribution. The exercise of this right, however, has resulted in inequity and undue complexity. In trying to determine what constitutes a fair allocation of remediation costs, courts have applied allocation techniques that ignore basic economic principles.

This Article surveys the current state of Superfund cost allocation and proposes an old method to address this new problem. The authors review the relevant provisions of CERCLA, the legislative history, and the attempts that courts have made to determine PRPs' fair share of remediation costs. The authors discuss the problems created by the existence of "common costs" at Superfund sites and propose three principles for evaluating cost allocation techniques. Applying these principles to each of the major allocation techniques currently being used, they find each technique wanting, and advocate the use of a stand-alone cost allocation method. This method, which has been used for decades in water resource projects, would allocate identifiable, direct cleanup costs to the responsible parties, and common costs according to the relative costs of cleaning up each PRP's waste as though that waste were the only waste at the site. The authors conclude that this method, although not necessarily appropriate in all situations, would provide a logical approach to cost allocation in a great variety of cases.

Mr. Butler, Dr. Hall, and Dr. Burton are partners in the Washington, D.C., office of the economic and management consulting firm Putnam, Hayes & Bartlett, Inc. Mr. Butler and Dr. Hall are managing directors of the firm, and Dr. Burton is a principal. Mr. Schneider is a partner in the Seattle law firm of Perkins Coie. Mr. Butler's practice focuses on environmental, product liability, and insurance coverage matters. Dr. Hall, a former commissioner of the Federal Energy Regulatory Commission (FERC), focuses his practice on the economic regulation of public utilities. Dr. Burton focuses his practice on economic and financial analysis of both liability and damage issues arising in businesslitigation. Mr. Schneider practices in the area of environmental litigation with an emphasis on hazardous waste, costrecovery, contribution, and environmental insurance coverage issues. The authors thank Douglas W. Greene and George Hansen for their assistance.

[23 ELR 10133]

Comprehensive Environmental Response, Compensation, and Liability Act¹ (CERCLA or Superfund) § 113(f)(1)² allows a potentially responsible party (PRP) that is liable to the government for cleanup and investigation costs to seek contribution from other PRPs. CERCLA § 107(a)(4)(B)³ also allows private parties to recover response costs from other PRPs. Liability under CERCLA § 107 is joint and several, which relieves the government of the burden of joining all PRPs in a single action and proving their individual contributions to the site.⁴ The right of contribution allows parties that have assumed more than their "fair share" of cleanup costs to seek recovery of these costs from other PRPs.

This liability scheme was adopted by Congress to promote prompt cost recovery and equitable allocation of liability.⁵ However, as many observers have shown, it does just the opposite: it causes delay, increases transaction costs, and results in inequitable allocations of liability.⁶ The current [23 ELR 10134] CERCLA liability scheme creates complex analytical and practical problems for those attempting to resolve issues of liability allocation. An important question that has generated prolonged and intense controversy among PRPs is: what is a "fair share" or an "appropriate allocation" of the remediation costs?

This controversy has been exacerbated by the absence of a consistent judicial approach, coupled with the U.S. Environmental Protection Agency's (EPA's) reluctance to provide clear guidance. Consequently, a variety of different allocation techniques has been used to divide cleanup costs among PRPs, usually after intense debate in settlement negotiations or court proceedings about how equitable shares should be determined. One result is disparate and inconsistent allocations at Superfund sites.

Most of the cost allocation techniques in common use lack satisfactory intellectual foundations. They are, at best, pragmatic solutions rather than principled approaches to the cost allocation problem. The absence of techniques grounded in accepted principles leads to unnecessary wrangling among PRPs (and their representatives) with attendant delays and unnecessary costs. Also, the current state of the art in Superfund cost allocation complicates the task of judges if cleanup cost allocation cannot be settled by agreement among the parties and the cost apportionment issue must be litigated. It is time to clean up the cost allocation methodology mess. This Article describes and evaluates four fundamental alternative approaches to allocation of cleanup costs. The techniques analyzed in this Article have been used or advocated in litigation and settlement at various sites.⁷ After discussing strengths and weaknesses of these alternative techniques, the Article presents suggestions for improving Superfund remedial cost allocation.⁸

The Issue and Its Legislative Background

CERCLA provides the federal government with the statutory power to impose liability on PRPs for the costs of cleaning up certain hazardous waste sites. Once PRPs have been notified of their potential liability at a particular site, EPA, the Department of Justice, and, sometimes, the relevant state agencies begin negotiations with the parties to initiate site cleanup actions, or to receive reimbursement for past cleanup actions. If negotiations are unsuccessful, a lawsuit or an administrative order will be filed against the PRPs. Settlement negotiations are often conducted against a backdrop of pretrial litigation activities. Even when negotiations with the regulatory agencies are successful, a subset of PRPs sometimes agrees to pay for or conduct the site cleanup, thus requiring contribution litigation to obtain reimbursement from the nonsettling PRPs.

Additionally, in the absence of governmental involvement, PRPs have undertaken voluntary cleanups at a number of sites. In many instances, the party undertaking the cleanup has less than full responsibility for the contamination and thus seeks reimbursement from other PRPs. In these cases, many disputes arise over the allocation of costs.

In both the cost recovery and contribution contexts, a key issue in settlement or litigation is: who will pay which portion of the remediation costs? EPA has not facilitateda generic resolution of this issue.⁹ EPA's policy appears to address the question of cleanup cost allocation on a case-by-case basis, maintaining the maximum flexibility to propose solutions that "feel right." One major problem with EPA's approach is that it provides very little guidance for parties looking for a way to structure an allocation proposal for settlement negotiations or judicial proceedings.

If negotiations fail, cost allocation under CERCLA § 113(f)(1) is available. However, the statute and its legislative history offer the courts little guidance on how to allocate costs. The legislative history indicates that Congress intended courts to resolve claims for contribution case-by-case, taking into account equitable considerations.¹⁰ CERCLA § 113(f)(1)¹¹ states that "[i]n resolving contribution claims, the court may allocate response costs among the liable parties using such equitable factors as the court determines are appropriate." The equitable factors' standard has the merit of giving the judiciary a wide scope of decisional latitude but, as Laurie Burt and Robert Sanoff have pointed out, the "equitable factors' standard offers little concrete guidance as to the proper distribution of liability among the responsible parties."¹²

After CERCLA was enacted, the House Committee on the Judiciary offered¹³ as "relevant criteria" for apportioning liability, six factors (the Gore factors) contained in an amendment proposed¹⁴ by Rep. Al Gore (D-Tenn.) that was passed by the House¹⁵ but not incorporated into the final [23 ELR 10135] version of CERCLA. Some courts apportioning CERCLA liability have referred to them.¹⁶ The Gore factors are:

1. the ability of the parties to demonstrate that their contribution to discharge, release, or disposal of a hazardous waste can be distinguished;

2. the amount of hazardous waste involved;

3. the degree of toxicity of the hazardous waste involved;

4. the degree of involvement by the parties in the generation, transportation, treatment, storage, or disposal of the hazardous waste;

5. the degree of care exercised by the parties with respect to the hazardous waste concerned, taking into account the characteristics of such hazardous waste; and

6. the degree of cooperation by the parties with federal, state, or local officials to prevent any harm to the public health of the environment.¹⁷

The Gore factors provide little help in developing an equitable standard for cleanup cost allocation in many situations.¹⁸ In addition, recent judicial decisions seem to be moving away from using the Gore factors as a foundation for cleanup cost allocations. Burt and Sanoff make the following important point:

It is important to the Superfund process that potentially responsible parties can predict with some precision the equitable factors which courts will apply in resolving contribution claims. As parties in Superfund cases come to [understand] the standards that will govern contribution claims, they will be able to evaluate better and more promptly the risks and benefits of settlement in comparison to likely litigation outcomes.¹⁹

Unfortunately, we are a long way from this desirable situation; court decisions to date are not very useful in determining an appropriate cost allocation mechanism.²⁰

Review of Judicial Precedents

Many courts have recognized that, despite the Gore factors, they may consider any factor they deem to be in the interests of justice.²¹ Thus, courts have articulated case-specific factors, often concentrating on specific Gore factors or examining the importance of other factors that are relevant to equitable apportionment.

Several courts attempting to apply the Gore factors have emphasized the importance of the degree of care exercised by a party with respect to the hazardous substances and the party's cooperation with the government. For example, in Amoco Oil Co. v. Dingwell,²² a group of generators who were signatories to an EPA consent decree sought approval of a settlement with the nonsignatory owner of the disposal site. The settlement allocated 65 percent of the cleanup costs to the owner and 35 percent to the generators. The court approved the settlement as fair, adequate, and reasonable, finding that it was equitable for the owner to bear most of the cost of cleanup.²³ After articulating the Gore factors, the court highlighted and applied three factors that it held to be relevant to equitable allocation of liability in this situation. The first was the parties' relatively equal involvement in creating the hazardous waste problem. The second was the greater care exercised by the generators and the owner's poor disposal practices, which caused the environmental damages. The third was the generators' cooperation with EPA, evidenced by their signing a consent decree to remedy the harm.²⁴

Similarly, in United States v. Tyson,²⁵ a group of generators that had signed a consent decree with EPA brought suit against the current owner-operator of the disposal site for contribution. The current owner had actual or constructive knowledge that the former owner had operated the site as a hazardous and industrial waste disposal facility, and allowed hazardous substances to leach into the groundwater and aquifer. The current owner did not notify regulators of the potential health hazard at the site, and was slow to cooperate with state and federal environmental officials once they identified the hazardous waste problem. When requested by state officials to clean up the site, the owner hired a contractor who had no experience handling industrial waste. The contractor did not adequately dispose of site waste. The owner refused to cooperate with EPA and did not sign the consent decree.²⁶ The court found these factors significant in allocating 50 percent of the response costs to the owner and 50 percent to the generators who, in contrast, had signed the consent decree and had already begun the cleanup.²⁷

Lack of cooperation with EPA also was significant in United States v. R.W. Meyer, Inc.,²⁸ in which the U.S. Court of Appeals for the Sixth Circuit affirmed a district court's allocation of cleanup costs. The district court found that the owner had constructed a defective sewer system that contributed to the contamination, and also bore significant responsibility simply by virtue of being the landowner.²⁹ [23 ELR 10136] Salient in the district court's equitable balance was the fact that the owner of the site neither assisted nor cooperated with EPA during its investigation and subsequent cleanup of the site. The district court allocated two-thirds of the liability to the generators and one-third to the owner.³⁰ The Sixth Circuit affirmed the district court's allocation, stating that "the trial court quite properly considered … not only the [owner's] contribution to the toxic slough … in a technical causative sense, but also its moral contribution as the owner of the site."³¹

By contrast, parties that have exercised care and cooperated with the government have benefitted in allocation decisions. For example, in United States v. Alcan Aluminum Corp.,³² the court apportioned response costs among generators based on the volume each contributed to the site, the degree of care they exercised in their waste disposal, and their degree of cooperation with local, state, and federal officials.³³ Although the court found the defendant contributed 12 percent of the volume to the site, it held the defendant liable for only 6 percent of the response costs because the defendant had demonstrated a greater degree of care and cooperation than the plaintiff.³⁴

In addition to the Gore factors, courts have considered other factors in determining an equitable allocation of costs. In United States v. R.W. Meyer, Inc.,³⁵ the Sixth Circuit noted that it was not limited to the Gore factors, stating that "the court may consider the state of mind of the parties, their economic status, any contracts between them bearing on the subject, any traditional equitable defenses as mitigating factors and any other factors deemed appropriate to balance the equities in the totality of the circumstances."³⁶

In cases involving site owners, the owner's state of mind, or knowledge of the contaminating activity, and whether the owner has profited from either the contaminating activity or subsequent cleanup have been important equitable factors. One such case is Weyerhauser v. Koppers Co.³⁷ In Weyerhauser, the court apportioned responsibility between the current owner and former operator of a wood treatment plant. The court discussed the Gore factors; however, in allocating cleanup costs, the court focused on the relative "benefits" received by the owner and former operator of the facility and knowledge of, and acquiescence in, the contaminating activities.³⁸ Specifically, the court held that the operator's activities were the sole cause of the environmental harm, and so it should bear the bulk of responsibility. However, the court held that the owner also should bear some liability, because it knew of the contaminating activities and received monetary benefits as a result of the contaminating activity. Accordingly, the court allocated 60 percent of the liability to the operator and 40 percent to the owner.³⁹

Similarly, in BCW Associates Ltd. v. Occidental Chemical Corp.,⁴⁰ the current owner and the current operator-lessee of a warehouse contaminated with lead dust brought suit against the former owner-operator after the current owner and the current operator-lessee had cleaned up the warehouse. The court found that the current owner benefitted financially from the cleanup of the site because it had paid an "as is" price and, after the cleanup, owned an environmentally safe property.⁴¹ The current operator-lessee, whose activities caused the threatened release of the lead dust into the environment, also benefitted from the cleanup, because the cleanup increased the value of its lease and satisfied its obligations under the Occupational Safety and Health Act.⁴² The court allocated one-third of the cleanup costs to the current owner, one-third to the prior owner-operator whose operations had generated the lead dust, and one-third to the current operator-lessee.⁴³

South Florida Water Management District v. Montalvo⁴⁴ involved allocation of cleanup costs between the current owners of land contaminated with pesticides and the prior operators whose activities caused the contamination. Significantly, the court found the current owner did not benefit from the land's contamination.⁴⁵ However, the owner was aware of the prior operator's activities, which contaminated the land. The court thus allocated the "lion's share" of liability, 75 percent, to the prior operator, which solely caused the harm, and the court then assessed 25 percent of the liability to the current owner because of its knowledge and acquiescence.⁴⁶

In Ellman v. Woo,⁴⁷ the court apportioned the future costs of cleanup between the owner of a site and a lessee of the site. The owner sued the lessee for contribution because of its role in site contamination, part of which resulted from petroleum hydrocarbons for which neither the owner nor the lessee was responsible, and part of which consisted of chemicals released during the operation of the lessee's dry cleaning business. The owner knew of the petroleum contamination when he purchased the property.⁴⁸ He asserted that he was entitled to 100 percent contribution, since the petroleum contamination had originated at and migrated from a neighboring site. The lessee argued that all the response costs were related to the petroleum contamination, since the same cleanup effort would remedy both contamination problems. Refusing to absolve either party of liability, the court allocated 50 percent of the future cleanup costs [23 ELR 10137] to the owner and 50 percent to the lessee.⁴⁹ The court noted that it could not equitably relieve the lessee of liability for the contamination it caused. On the other hand, the owner (or some other party that was responsible for the petroleum contamination) could not equitably be relieved of all its costs simply by the "fortuitous" occurrence of the lessee's activity.⁵⁰ Moreover, the owner's knowledge of the contamination at the time he purchased the property bolstered the court's determination that it was equitable for the owner to bear one-half of the cleanup costs.⁵¹

An owner's lack of knowledge of pollution problems has been held to be an important equitable factor. In Gopher Oil Co. v. Union Oil Co.,⁵² the U.S. Court of Appeals for the Eighth Circuit affirmed a district court decision holding that the current owner-buyer of an industrial site was entitled to 100 percent contribution from the site's seller, which formerly owned and operated the site, despite an "as is" clause in the sale contract, because the seller had fraudulently informed the buyer that the site was free of pollution. The district court found that the seller knew of, and was responsible for, the contamination, and that the buyer neither contributed to nor had specific knowledge of the contamination.⁵³

A party's lack of knowledge was also an important equitable factor in Danella Southwest, Inc. v. Southwestern Bell Telephone Co.⁵⁴ In Danella Southwest, the U.S. District Court for the Eastern District of Missouri held that a contractor was not liable for contribution of response costs incurred by a telephone company. The contractor excavated dirt pursuant to a contract with the telephone company. The court held that the contractor did not know, but the telephone company should have known, that the dirt contained dioxin. As a transporter, the contractor was potentially liable for contribution, but the court held that a balancing of the equities relieved the contractor of any CERCLA liability. The court found significant the contractor's lack of knowledge that the dirt was contaminated, the telephone company's negligent failure to learn of the contamination, and the contractor's performance of the excavation in a "workmanlike manner."⁵⁵

Some courts have suggested that the relative fault of the parties is a relevant factor in allocating CERCLA liability.⁵⁶ In Environmental Transportation Systems, Inc. v. Ensco, Inc.,⁵⁷ a subcontractor that transported hazardous materials sued the contractor for contribution after the subcontractor's truck crashed en route to the disposal site, causing the hazardous materials to spill from the truck. The court stated that CERCLA § 113(f)(1) contemplates an assessment of the relative fault of each responsible party and held that since the subcontractor was wholly responsible for the spill, the contractor had no liability.⁵⁸

Courts have taken different approaches on the question of whether costs should be allocated based solely on the volumes of hazardous substances generated by each PRP. In United States v. Western Processing Co.,⁵⁹ the court allocated response costs based solely on the relative volumes of hazardous substances generated by parties that had settled the claims at issue and those that had not.⁶⁰ However, in Alcan Aluminum,⁶¹ the court held a generator liable for only 6 percent of the cost of cleanup, despite its 12 percent volumetric contribution, because of the generator's superior cooperation with the government and the degree of care it exercised. In addition, most courts have not considered toxicity in the equitable balance.⁶²

Courts have found significant a party's cooperation with the government and its care with respect to the hazardous substances.⁶³ In cases involving allocation of liability to owners, courts have found salient the owner's knowledge of, and acquiescence in, contaminating activity and whether the owner received financial advantage from the contaminating activity or from the subsequent cleanup.⁶⁴ While these factors are case-specific, they suggest that some courts are moving away from precise, causative, and scientific allocation and toward a broad and equitable understanding of responsibility for environmental harm.

Recent judicial decisions also suggest that many courts are moving away from mechanistic allocations based on volume, or other measures of usage, toward benefit-based allocation that reflects equitable factors. This is a very desirable trend. While most courts have not relied on, or developed, mathematical support for their allocation decisions, the attempt to find a single formula for allocating all remediation costs at most Superfund sites is doomed to failure because many of the costs are common costs. Recognizing this fact is essential to the development of an equitable and intellectually sound approach to CERCLA cost allocation. The remainder of the Article explores how equitable cost allocation principles can be developed.

Principles of Cost Allocation

One reason that it is difficult to allocate Superfund cleanup costs among PRPs is that a substantial part of the remediation costs at a typical site is what economists call "common costs." In this context, common costs are costs that cannot [23 ELR 10138] be uniquely attributed to one waste type or party. For example, remedial investigation/feasibility study costs are common to all the waste at a site. Alternatively, the costs of incinerating dioxin-contaminated soil or treating TCE-contaminated groundwater are usually not common costs, but can be attributed to dioxin wastes and trichloroethylene (TCE) wastes, respectively. There is no clear economic principle that says that one division of common costs is inherently and logically superior to another. This is an area where a value judgment must be made. However, the lack of scientific theories on dividing common costs fairly does not mean that principled approaches to solving the Superfund cost problem cannot be developed. After all, in our economy it is more common than not for people to engage in activities as a group, with the costs of the activities reimbursed by the individuals involved. It is necessary to refer to first principles and practice in other areas to develop some general rules, while recognizing that acceptable rules or principles will be different from mathematical certainties. On these bases, the authors propose the following principles.

The first fundamental principle of cost allocation is that a party or class of parties should bear those costs that can be directly attributed to them. This principle is generally accepted in all situations requiring costs to be allocated.⁶⁵ In different cost allocation contexts, such clearly traceable costs are variously referred to as "attributable," "direct," or "variable" costs.⁶⁶ To reemphasize this fundamental point, it is fair and economically efficient that costs that can be clearly traced to the actions of a specific entity should be paid by that entity.⁶⁷ In public utility regulation, this cost causation principle is widely accepted as being consistent with, and required by, the public interest. It should be the cornerstone of Superfund cost allocation.

The second fundamental principle of cost allocation is the obverse of the first: costs that cannot be logically and directly attributed to a party or class of parties should not be borne by that party or class based on cost causation. It is a fundamental economic axiom that there is no logical or nonarbitrary way to allocate common costs to any specific party or class on the basis of cost causation.⁶⁸

Common costs are incurred when a specific operation yields two or more products, services, or benefits, resulting in economies or savings that would not be realized if the products, services, or benefits were produced separately. The classic example is the production of wool and mutton. Because the savings result from a common activity rather than separate activities, it is inherently impossible to divide the common costs among the individual products, services, or benefits that are produced. Put in other terms, producing several goods and services together yields "economies of scope." Economies of scope result in savings that belong in common to the different products, services, or benefits produced together; these economies cannot be logically attributed to any single product, service, or benefit on the basis that the particular product, service, or benefit produced or caused those economies.

A third fundamental principle is that an equitable allocation of costs that cannot be allocated on the basis of causation should be related to benefits. Admittedly, such an allocation must be based on a value judgment. Nevertheless, it is a value judgment that is commonly made and widely accepted in our society. For example, it is typical that the cost of an office lunch is divided among those attending in such a way that the people who ordered drinks often are expected to contribute more to the common bill than people who did not order drinks. This principle has also been applied when the stakes are much higher. For example, in federal water resource development projects, costs in the hundreds of millions of dollars are divided on the basis of the benefits to various customers.⁶⁹ Moreover, judicial decisions on Superfund cost allocations are increasingly relating relative benefits and equitable cost allocations.⁷⁰

There is no precise, on-point analogy from cost allocation in other areas to cost allocation in Superfund situations. Nevertheless, examples of cost allocation among several parties or classes of parties is very common. For instance, the U.S. Postal Service allocates costs among different classes of postal service; utilities, railroads, pipelines, and other common carriers allocate costs among different classes of customers. Governmental regulators wrestle with whether proposed cost allocations are "just and reasonable" or "fair and equitable." These decisions are similar to the decisions of judges or settlement negotiators trying to equitably divide the costs of cleaning up Superfund sites.

On the other hand, allocating remediation costs at Superfund sites is unique, because in that situation it is not necessary to produce an efficient pricing mechanism that will create incentives for using the facility. Typically, Superfund sites are inactive waste facilities. The hazardous waste has been dumped at the site long ago and the facility [23 ELR 10139] is no longer accepting additional waste. Thus, determining who will use the facility and how to control this use by charging efficient prices is not at issue. In a sense, the cost of cleanup is a "sunk" cost since the decision to dump waste at the site has already been made, Congress mandated that the environmental hazards created by that dumping must be corrected, and no new products or services will be generated at the site. Therefore, the goal in allocating Superfund cleanup costs is to expedite settlement among PRPs by producing an outcome that is equitable. Such an outcome would probably minimize negotiation and litigation costs.

It is hard to think of benefits that PRPs receive from participating in Superfund remediation projects. Few PRPs feel themselves benefitted by receiving a bill from EPA. Nonetheless, the benefit/burden principle applies here, since Congress has mandated that PRPs clean up contaminated sites and the PRPs save money by engaging in joint cleanup action, rather than discharging their cleanup responsibilities separately. This benefit provides a basis for developing a principled approach to cost allocation — one that has long-standing parallels to other areas.⁷¹

Although not fundamental principles, there are other desirable properties for a Superfund cost allocation methodology. One of these is that the allocation procedure should be implemented at a reasonable cost. All things being equal, the lower cost allocation procedure is preferable. However, the greater the cleanup costs are at a given site, the more likely it is that additional investments are warranted in implementing more costly and better allocation procedures. If the cleanup costs at a site are relatively low, or the waste streams are relatively homogeneous and require similar remedial actions, a less costly method may be preferable. The additional resources spent in implementing a more costly allocation procedure must provide a better allocation of costs. Also, the analytic procedure should be clear so that parties can understand the process used to generate the allocations. If a procedure involves a complex "black box" resolution of the allocation issue, it is less likely to be acceptable since PRPs will have difficulty in evaluating whether the allocation is equitable or not. If costs can be allocated on the basis of cost causation and relative benefits received, in a manner that is understandable and at a cost that is reasonable, then the goal of expediting settlement by reducing negotiation and litigation costs is more likely to be achieved.

Alternative Allocation Methodologies

There are four basic cost allocation approaches used in Superfund situations. These approaches allocate costs on the basis of volume, relative toxicity, a combination of toxicity and other physical factors, or stand-alone costs. The following discussion evaluates each approach based on the principles enumerated above.⁷²

Volumetric Apportionment Method

Allocation on the basis of relative volumes shipped by each PRP, or on the proportionate volume deposited at the site by each PRP, has been the methodology used at many sites. Volumetric apportionment methods rely on an analysis of what is commonly referred to as "waste-in data." These data are a compilation of information on the quantity and type of waste brought to the waste site. The information is usually obtained from records maintained by site operators, transporters, generators, state records, and on-site inventories. Gathering these data is usually a lengthy and complex task. For instance, at some sites data on volumes of wastes contributed by various parties are missing, incomplete, falsified, or fraudulent. Data from one source may be inconsistent with data from other sources. Not only are historical records from operators, transporters, and generators often inconsistent, but historical records may well be inconsistent with the findings of current on-site inventories. Depositions of employees of the waste facility and transporters are sometimes taken to assemble the record on the site's wastes. In any event, the PRPs will usually have to agree on which sources of information will be relied on in developing the waste-in list before cost allocation based on volume can begin.

The major benefit of using a volumetric approach in allocating costs is its simplicity and understandability. Although establishing waste-in volumes is not an easy matter, once waste-in volumes have been determined, the allocation technique is simple and straightforward. Of the four basic cost allocation approaches used in Superfund situations, the volumetric approach is probably the least costly to implement. Thus, if the Superfund site contains a set of relatively homogeneous waste types that require similar remedial actions, a volumetric approach may be attractive and reasonable. However, it is the exceptional Superfund site that has such characteristics.

The major disadvantage of a volumetric approach is that it violates the cost causation principle.⁷³ By allocating volumes without regard to directly assignable costs, cross-subsidization among PRPs will result. PRPs are not likely to view this as equitable and, thus, are less likely to accept such an arrangement. Even for those costs that are not attributable to any responsible party, allocating these common costs by volume is likely to raise issues of fairness among the PRPs since volumetric allocations do not distinguish between hazardous and nonhazardous wastes. Clearly, remedial action is more costly with hazardous waste than nonhazardous waste, which is not taken into consideration by volumetric allocation. Finally, choice of the unit of measurement for estimating volumes can result in widely different allocations. Depending on whether volume is measured in gallons, cubic yards, number of drums, or some other unit, the allocation results will differ. This arbitrary result also raises issues of fairness to PRPs.

The basic problem with this allocation technique is that a substantial portion of the costs of a typical cleanup project is not related to volume, at least not on a proportionate basis. An example is planning and set up costs: a small project is likely to require as much, or almost as much, of such "overhead" activities as a large project. In other words, [23 ELR 10140] there are typical economies of scale in Superfund cleanup efforts. By implicitly assuming that cleanup costs are related to volumes deposited on a direct one-to-one basis, the volume allocation technique is at odds with the basic engineering cost principles of remediation.

Often, the volumetric technique is used as a benchmark or starting point for settlement negotiations or judicially imposed allocations at many sites. However, there may be a trend in judicial decisions away from such a mechanistic approach toward allocation techniques that permit a greater weighing of relative equities.⁷⁴ This is a salutary trend, considering the deficiencies in the logical foundation of this technique.

Relative Toxicity Apportionment Method

Toxicity-based approaches to allocating remediation costs have been considered by PRPs at several sites. The relative toxicity approach classifies the wastes to be treated into groups of wastes with similar toxicity. A hazard or toxicity score for each group is then generated. The volume of each waste group is multiplied by its toxicity score to get a toxic equivalent volume. The ratio of each PRP's toxic equivalent volume to the total toxic equivalent volume of all the wastes at the site represents its cost apportionment share.

To perform an allocation using this technique, waste volume and waste type data are needed. The descriptions of the types of wastes that are obtained from site operators and transporters are usually not specific enough to establish categories of relatively homogenous toxicity. Often waste site operators will describe their wastes as solvents, solids, sludges, waste oils, municipal waste, and so forth. The toxicity of various wastes within each of these categories can vary by many orders of magnitude. Therefore, the generator PRPs need to provide more detailed information on the composition and quantity of wastes sent to the site. An expert toxicologist is also needed to analyze the waste information and develop the relative toxicity scores for each waste type.

The toxicity apportionment method addresses one drawback of the volumetric approach — the assumed homogeneity of waste types — since more highly toxic wastes receive a higher cost share than less-toxic wastes, all other things being equal. However, as with an allocationbased on volume, the relative toxicity method is not likely to reflect the causation of costs. The relationship between the degree of toxicity of wastes and Superfund remediation costs has never been established.⁷⁵ Also, this allocation method is inherently more costly than the volumetric approach and may be as costly as the other approaches.

In addition, the appropriateness of a particular toxicity score may be questionable, because considerable judgment must be used in assigning the scores, and because there is no direct correlation between toxicity and cleanup costs. Thus, the PRP's lack of understanding of the resulting allocation is an obstacle to resolving the allocation issue.

As with the volumetric technique, the relative toxicity approach also has the problem of arbitrariness. Because toxicity must be calculated by some measure of waste volume, this approach also results in a different set of allocations depending on the measure of volume used. These and other problems with this approach are recognized by most PRPs so that the relative toxicity allocation method is rarely offered as a method of allocating cleanup costs, even though the relative toxicity of wastes are often cited as one reason for inequities with the volumetric allocation approach.

Combination of Toxicity and Other Physical Factors Apportionment Methods

This approach scores each waste type for toxicity and several physical factors, in addition to estimating volume. One such approach, developed by G&E Engineering, Inc., was considered at the Petro Processors site in Louisiana.⁷⁶ G&E Engineering applied its Cleanup Cost Allocation (CCA) model to the site to calculate proportionate liability shares for each of the generators. The CCA model considers the volume, toxicity, and physical characteristics of each of the generator's individual wastes. Each waste is individually scored for each generator and then a total weighted average score is developed for each generator to determine its liability share. The toxicity of each weaste is measured as the highest applicable score in a number of toxicity/hazard tests. The physical threat of each waste is estimated through the combination of a number of physical factors, including its solid/sludge/liquid state and mobility, among others. The CCA model considers only waste characteristics, and not actual site conditions, occurrences, or more importantly, actual cleanup costs.

The CCA model appears to satisfy none of the principles of cost allocation discussed above. The technique is more costly than a simple volumetric approach, but does not result in a clearly superior allocation of costs. Equity concerns are raised because the model requires many arbitrary assumptions. The technique is complicated, making it difficult to evaluate the allocation results. Moreover, the technique does not attempt to assign direct costs to those responsible for them. In fact, it develops allocations that depend only on toxicity and physical factors of wastes and fails to account for site conditions.

PRPs in the Petro Processors settlement process disputed the responsibility share produced by using the model. They claimed that it did not accurately reflect the environmental threats posed by the wastes at the site or the real costs of removing those wastes.⁷⁷ For example, all highly mobile [23 ELR 10141] wastes were given high scores and, thus, high liability shares, although not all such wastes actually leached at the site causing extraordinary cleanup costs. Similarly, highly toxic wastes were given high scores and shares although a properly maintained barrel containing highly toxic waste may present a cleanup cost similar to that of a barrel containing substantially less toxic materials.

In sum, while the CCA model attempted to consider a wide range of factors in apportioning liability, it considered only waste characteristics, and not actual site conditions, occurrences, or more importantly, actual cleanup costs. Thus, it is not likely to promote acceptable allocations at most waste sites.

Another engineering cost allocation model, the Cause/Effect Allocation (C/EA) model developed by Kellogg Corporation, does consider the actual cost of cleanup in allocating costs. The Kellogg model purports to allocate costs to site generators based on the cause/effect relationships between wastes disposed of at the site and the incurrence of remedial costs.

Under this model, cleanup or response costs are first allocated to remedial action categories. Those costs not allocated to specific remedial action categories (i.e., common costs, which the model does not recognize as such) are either allocated volumetrically, as in one observed application of the model, or in proportion to costs previously assigned.⁷⁸ Then, waste types are assigned to each remedial action category. If two or more waste types are assigned to the same remedial action category, the model allocates costs to these waste types in proportion to relative volume or to the alleged relative cost impact that each waste type has on the remedial action. However, when a specific operation yields two or more products, the resulting costs are common costs.⁷⁹ By definition, common costs cannot be allocated on the basis of cost causality. Therefore, the Kellogg approach, which claims to be a cause/effect model, violates one of the allocation principles described above,⁸⁰ since it attempts to allocate common costs based on causation that, by definition, does not exist.

The final step in the model is assignment of each waste type's volume to generators so that costs of each remedial action category linked to waste types can be allocated to generators. In short, the C/EA model assigns the volume of each waste type to a generator or set of generators and this, in turn, is used to allocate the remedial action costs. Those cleanup costs not assigned to specific waste types are allocated by volume, or in proportion to the cost allocation of those costs assigned to waste types. Obviously the selection of waste types, the classification of generator waste by waste type, and the selection of remedial action cost categories involve a number of arbitrary assumptions and can have a significant effect on the resulting allocation.

Unlike the CCA model, the C/EA model has the admirable trait of directly considering cleanup costs in developing liability shares. Unfortunately, the model goes too far in attempting to establish causality where none exists. As mentioned above,⁸¹ if two or more wastes require the same remedial action, the model attempts to allocate these wastes on the basis of causality. The problem is that causality cannot be established, just as in the wool and mutton example, where the cost of raising the sheep cannot be causally allocated between the output of wool and mutton.⁸² Notwithstanding this, the C/EA model attempts to allocate these common costs on a volumetric, or other, basis under the guise of causation.

Generally speaking, the problems associated with allocating common costs on the basis of volume or toxicity also apply to the CAA model and the C/EA model. However, these models are much more costly to implement than a straight volumetric approach, giving them a disadvantage over the volumetric approach. Moreover, given the complexity of each model and the many arbitrary judgments underlying their use, there are also issues of understandability and equity.

Stand-Alone Cost Allocation Method

The Stand-Alone Cost (SAC) approach would first allocate any identifiable direct cleanup costs to the responsible parties. Then, it would allocate common costs according to the relative costs of cleaning up each PRP's waste as if that waste were the only waste at the site.

Mathematically, this approach is defined as follows:

Share of Common Costs[i] - SAC[i] N SIGMA i-1 SAC[i]

v.here:

SAC[i] = the stand-alone cost for PRP i (or PRP group i)

N = the total number of PRPs (or PRP groups)

SIGMA = the sum of all SAC[i] terms

This calculated share of the common costs is multiplied by the total common costs associated with the site cleanup to arrive at the amount of common costs allocated to each PRP group. For example, assume only two PRPs are involved at a site. The total cleanup cost is assumed to be $300,000, of which $50,000is directly attributable to one PRP, $50,000 is directly attributable to the other, and $200,000 is not attributable to any one party (it is a common cost). Further assume that the stand-alone cost of removing the wastes of the first PRP is $150,000, and the stand-alone cost for the second PRP is $100,000. Therefore, the first PRP's share of common costs would be 60 percent [$150,000 ($150,000 + $100,000)] and the second PRP's share would be 40 percent [$100,000 ($150,000 + $100,000)]. According to the stand-alone cost allocation method, the cost of cleanup allocated to the first PRP would be 60 percent of the $200,000 common cost plus $50,000 in direct costs, or $170,000. Likewise, the second PRP would be responsible for 40 percent of the common cost plus $50,000 in direct costs, or $130,000.

The idea of factoring stand-alone costs into cost allocation is an old one, and merits consideration as a technique long used to achieve equitable cost allocation in situations in [23 ELR 10142] many ways similar to Superfund cleanups. The stand-alone cost approach came into prominence in the 1930s as a cost distribution tool for water resource projects.⁸³ This approach was proposed by some of the authors as a method for allocating common cleanup costs at the Western Processing Superfund site in Washington.⁸⁴

Allocating Superfund remediation costs based on standalone costs follows naturally from the concept of economies of scope. The stand-alone cost of a cleanup effort is analogous to the cost of producing a single product or service separately from, rather than in combination with, other items. Economies of scale, to the extent they exist at the site, are also reflected in the stand-alone cost calculations, since scale economies are present when the cost per cubic yard to treat a large volume of waste is less than the cost per cubic yard to treat a small volume. In addition, the stand-alone cost approach is a simple and straightforward intellectual concept that has proven useful in distributing or assigning shares of common costs among different groups, a task that is often required in economic regulation and public enterprise pricing.

From the early days of the Tennessee Valley Authority to the present, the stand-alone cost concept has been viewed as an equitable way of allocating common costs. The decision to use this method in water resource projects, for example, was based on the idea that fairness requires project beneficiaries to pay in proportion to the benefits they receive.⁸⁵ Intuitively, it seems unfair to require beneficiaries to pay more for a service or product obtained from a multipurpose project than for equivalent benefits they could obtain from a stand-alone project. Thus, the traditional argument for using stand-alone cost allocation in common cost recovery situations has been equity or fairness.⁸⁶ Since the equity/benefits approach to Superfund cost allocation has become increasingly attractive to courts, a close look at this technique is merited.

Starting in the 1970s, some economists argued that applying the stand-alone cost method to common cost recovery was not only a reasonable approach to equitable common cost sharing, but also advanced economic efficiency. This position was eloquently developed in a seminal article by Gerald R. Faulhaber.⁸⁷ Using game theory, Faulhaber demonstrated that the efficiency benefits associated with economies of scope will not be realized unless entities participate in multi-purpose projects, and entities will not voluntarily join in a project unless their individual costs as members of the project are below their stand-alone costs.

The essence of Faulhaber's contribution to understanding the efficiency properties of the stand-alone cost concept can be illustrated by a hypothetical example.⁸⁸ Consider three families, each of which needs a water supply. Assume that the cost of a well for family A would be $10,000; for family B, $20,000; and for family C, $30,000. If each family developed its own water supply the total cost for wells would be $60,000. Assume that a well to serve all three families would cost $30,000 and that a joint project would require tanks and pipe costing an additional $10,000 for a total project cost of $40,000. To simplify the example, assume there are no variable costs, i.e., the only costs are the costs of the well, tank, and pipes which are fixed costs, common to the joint project. There is, in this illustration, a potential saving (economy of scope) of $20,000, if the project is developed with $40,000 of common costs rather than $60,000 of individual stand-alone or alternative costs.

How should the common costs be allocated? If the common costs are divided evenly, family A would pay $13,333. This is greater than the cost to family A of drilling its own well. This result would not only be unfair, but unless family A is coerced, it is clear that family A will not participate in the project, and the potential efficiency gains from the joint project would not be realized.

Assume the costs were allocated on the basis of water consumption. This would be arbitrary and the result would also be unfair, because the costs of the joint well project are common and in no way affected by how much water each family uses. There may also be a practical problem. Assume family A consumes 20 gallons per day, family B also consumes 20 gallons per day, but family C consumes only 10 gallons per day. Allocating cost based on water consumption would require family A to pay more than it would if it built its own well system, since families A and B would each have to pay 40 percent of the project, or $16,000 each. This result would be both inequitable and inefficient.

This example reveals several general points of importance about the allocation of waste cleanup costs. The first is that it can be arbitrary to attempt to allocate cost recovery responsibility for a project based on cost causation because none of the three families can be said "to cause" any part of the $40,000 joint project. The costs are a consequence of their combined needs.

Second, there are several obvious benefits from joint participation in projects and attendant cost sharing. In the cited example, family C benefits greatly from the joint project; it avoids a $30,000 expenditure and the total three-family project cost is only $40,000. On the other hand, the project is not going to benefit family A greatly: it only avoids a $10,000 expenditure. Intuition calls for trying to factor benefits and the differences in benefits into a joint cost allocation scheme. Using the stand-alone cost methodology follows naturally from the concept of [23 ELR 10143] benefit sharing. Using this methodology, each family pays the same fraction of its stand-alone cost. The stand-alone costs are $10,000 for family A, $20,000 for family B, and $30,000 for family C. The stand-alone cost method would allocate $40,000 of common costs as follows: one-sixth to family A, one-third to family B, and one-half to family C, or $6,667 to family A, $13,333 to family B, and $20,000 to family C. As is evident, each family is allocated the same fraction of its stand-alone costs (i.e., two thirds). This methodology also has an efficiency element: if the cost allocation results in a charge that is greater than the avoided or stand-alone cost of any participant, the participant that is charged more than the avoided or standalone cost will go it alone and the efficiencies of joint action will be lost. Since the stand-alone cost allocation results in charges less than each family's stand-alone cost, the efficiencies of joint action are obtained.

The stand-alone cost method of allocating common costs when combined with the assignment of direct costs to PRPs addresses many of the equitable factors mentioned in the legislative history of CERCLA and in judicial decisions. As mentioned above, this method allocates common costs so that each PRP (or PRP group) pays the same fraction of its stand-alone cost. The stand-alone cost calculation also reflects the economies of scale that are connected to the benefits each party receives from its participation in joint cleanup activities. Several of the Gore factors, such as the amount of hazardous waste, the degree of toxicity, and the contribution of each party's waste to site contamination, are reflected in the allocation through the calculation of stand-alone cost and the assignment of direct costs. This approach weighs these factors based on their effect on cleanup cost, which seems more appropriate than weighing them based solely on waste volume or waste toxicity, since the purpose is to allocate cleanup cost, not waste or toxic equivalent volumes.

To understand the SAC methodology, it is important to understand the way in which it allocates not only common costs, but direct costs as well; in other words, its application of the cost causation principle. The first step in implementing the stand-alone cost technique is to assign any direct costs to the responsible parties. Then, the technique should be used to allocate the remaining common costs. This methodology, while more costly to develop, can be more cost effective than a volumetric approach if it encourages settlement of the allocation issue. One disadvantage, though, is that the cost of implementing this approach increases as the number of PRP groups increases due to the concomitant increase in the number of stand-alone cost options that must be estimated. Another disadvantage to this approach, and probably any numerical approach, is that it does not address every conceivable equitable factor. For example, the standalone cost method does not consider the degree of a PRP's cooperation with EPA and other regulatory authorities, which is one of the Gore factors.

Finally, it should be noted in passing that EPA has indicated it may use a variant of the stand-alone cost approach for determining appropriate Superfund cleanup cost settlements with generators and transporters of municipal solid waste.⁸⁹ Under the approach proposed by EPA, which is referred to as the "Unit Cost formula," costs are allocated on the basis of the relative cost of remediating an acre of municipal solid waste (MSW) and an acre of industrial hazardous waste. The unit cost formula is as follows:

[MSW Unit Cost/(MSW Unit Cost + Industrial Unit Cost)]

If EPA calculates a $94,000 average cost per acre to remediate an acre of MSW and a $2,279,000 average cost to remediate an acre of industrial hazardous waste, this would result in a 4 percent settlement contribution for generators and transporters of MSW [$94,000 ($94,000 + $2,279,000)]. This percentage would be multiplied by the total cost of cleanup at the site to determine the dollar contribution for MSW generators and transporters.⁹⁰

Conclusion

Because of the common cost nature of Superfund site remediation, cost allocation is inevitably based on value judgments and is not simply a matter of applying scientific formulas. Thus, in a settlement negotiation if the parties agree that a particular technique yields appropriate results, there is no reason to dispute their decision. However, what if the parties cannot agree? Or, what if agreement requires lengthy and expensive debate over the justification for various allocation techniques? To deal with such situations, the PRP community must understand the logical foundations of the techniques it uses and reach some common understandings about fairness and equity. To that end, PRPs should apply a set of principles, such as those set forth above, that allow them to determine which allocation technique or techniques will yield results superior to other techniques, depending on the features of the site involved.

Unfortunately, the allocation techniques most widely discussed in the Superfund literature and most often used in practice lack the logical underpinnings necessary for the development of principled positions. While there may be situations where cleanup is simple and a clear, logical nexus exists between cleanup cost, on the one hand, and volume and toxicity, on the other, most site remediation is so complex that applying volumetric or toxicity-based techniques would be arbitrary. These techniques ignore the economies of scale present in most cleanups, and they attempt to establish causation where no causation can be established due to the existence of common costs. Public utility regulators spent fruitless decades searching for a nonarbitrary way of allocating the common costs of providing telephone, electricity, gas, and other public utility services before finally recognizing that the task was impossible due to economies of scale and scope. For the environmental community to run down the same blind alley would be a tragedy. Although PRPs should bear the remediation costs properly attributable to their actions, when such attribution is impossible, [23 ELR 10144] it is vital that allocation not be based on an arbitrary choice of methodologies derived from false assumptions concerning causation. Accepting this principle is vital, because a significant part of cleanup costs at many sites is impossible to attribute to a particular PRP.

It is also essential to realize that cost allocation is intended to produce an equitable result and the principles used in producing that result should be generally accepted as fair. Although the Gore factors are often cited in discussions of equitable allocation of Superfund costs, they are so general that they do not provide clear guidance. The principle that common costs should be allocated in accordance with benefits derived seems to be widely accepted in this society. Some courts may be moving toward accepting this principle in Superfund cost allocation cases, and economists and other analysts can further this trend by developing more logical benefit-based allocation techniques. As part of that effort, they should consider the stand-alone cost technique, which has been used for decades to apportion costs among participants in water resource development projects.

There is no reason to believe that one technique will fit all cases. Equitable factors, technical complexities, and sitespecific issues require that the choice of an allocation technique be highly selective. Also, the circumstances of making this choice in the context of settlement negotiations prohibits the justified application of a single technique on a blanket basis. Nevertheless, in many situations serious attention should be given to the stand-alone cost technique. It is a principled method for allocating Superfund remediation costs and can be useful in various contexts.

1. 42 U.S.C. §§ 9601-9675, ELR STAT. CERCLA 007-075.

2. 42 U.S.C. § 9613(f)(1), ELR STAT. CERCLA 039.

3. 42 U.S.C. § 9607(a)(4)(B), ELR STAT. CERCLA 024-025.

4. United States v. Chem-Dyne Corp., 572 F. Supp. 802, 13 ELR 20986 (S.D. Ohio 1983) (holding that CERCLA liability is joint and several). However, the recent decision in United States v. Alcan Aluminum Corp., 964 F.2d 252, 22 ELR 21124 (3d Cir. 1992), provides PRPs with a substantial basis for challenging the imposition of joint and several liability.

5. See H.R. REP. No. 253 (III), 99th Cong., 2d Sess. 15-16 (1986), reprinted in 1986 U.S.C.C.A.N. 3038, 3038-39.

6. See Insurance Issues and Superfund: Hearings Before the Senate Comm. on Environment and Public Works, 99th Cong., 1st Sess. 54, 58 (1985) (statement of John C. Butler III); WILLIAM K. REILLY, A MANAGEMENT REVIEW OF THE SUPERFUND PROGRAM 1-2 (1989); U.S. OFFICE OF TECHNOLOGY ASSESSMENT, COMING CLEAN — SUPERFUND'S PROBLEMS CAN BE SOLVED … 3 (1989).

7. There are many variants of the basic methodologies addressed in this Article. Examples of other methodologies are equal shares, previously assigned direct costs, and incremental costs methods. However, this Article is concerned with the fundamental properties of techniques in widespread use. Accordingly, it does not present an exhaustive catalogue of all methodologies.

8. Do the failures of the current allocation techniques matter? In some situations the answer is no. Once an agreement is reached among the PRPs on how to divide cleanup costs, it may not matter how the parties reach it. Nevertheless, logical foundations and intellectual principles do matter. First, in the absence of a logical approach to reaching a principled outcome, far too much time, energy, and money is devoted to determining what constitutes a fair division of cleanup costs. Second, failure to reach an agreement on cost allocation often leads to delays in site cleanup and an increase in contribution litigation. Third, when PRPs cannot agree, a court must impose a solution, and courts are asked to render decisions based on logic and principle. The present theory and practice of Superfund remediation cost allocation offer them scant assistance.

9. One commentator has concluded that

in practice, the broad CERCLA Settlement Policy provides little guidance on allocation and EPA has utilized [nonbinding preliminary allocations of responsibility], de minimis settlements, and mixed funding in only a handful of cases. EPA has been reluctant to involve itself in allocation issues or to facilitate settlement through use of these mechanisms as provided in CERCLA Section 122…. EPA has essentially left the PRPs to resolve the allocation issue, usually through a PRP allocation committee and sometimes with the assistance of an outside consultant to review and assemble available information.

Daniel H. Squire, PRP Allocation Issues, 18 CHEM. WASTE LITIG. REP. 1046, 1052-53 (1989).

10. H.R. REP. No. 253 (III), supra note 5, at 19, reprinted in 1986 U.S.C.C.A.N. 2835, 3038, 3041-42.

11. 42 U.S.C. § 9613(f)(1), ELR STAT. CERCLA 039.

12. Laurie Burt & Robert S. Sanoff, Allocating Contribution Shares in Superfund Cases, 20 CHEM. WASTE LITIG. REP. 203, 204 (1990).

13. H.R. REP. No. 253 (III), supra note 5, at 19, reprinted in 1986 U.S.C.C.A.N. 3038, 3042.

14. 126 CONG. REC. H9461 (daily ed. Sept. 23, 1980) (statement of Rep. Gore).

15. 126 CONG. REC. H9468 (daily ed. Sept. 23, 1980).

16. See Carroll E. Dubuc & William D. Evans Jr., Recent Developments Under CERCLA: Toward a More Equitable Distribution of Liability, 17 ELR 10197 (June 1987).

17. H.R. REP. No. 253, 99th Cong., 1st Sess. 19 (1985), reprinted in 1986 U.S.C.C.A.N. 2835, 3042. See United States v. A & F Materials Co., 578 F. Supp. 1249, 1256, 14 ELR 20105, 20108 (S.D. Ill. 1984); see also United States v. R.W. Meyer, Inc., 932 F.2d 568, 571, 21 ELR 21062, 21063 (6th Cir. 1991); Danella Southwest, Inc. v. Southwestern Bell Telephone Co., 775 F. Supp. 1227, 1234 (E.D. Mo. 1991); Purolator Prods. Corp. v. Allied-Signal, Inc., 772 F. Supp. 124, 134-35 (W.D.N.Y. 1991); Weyerhauser Co. v. Koppers Co., 771 F. Supp. 1420, 1426, 22 ELR 20168, 20170 (D. Md. 1991); United States v. Alcan Aluminum Corp., No. 87-CV-920, 1991 U.S. Dist. LEXIS 18721, at 1, 10 (N.D.N.Y. Dec. 27, 1991); United States v. Tyson, 19 CHEM. WASTE LIT. REP. (Computer L. Rep., Inc.) 1310 (E.D. Pa. Dec. 29, 1989); Amoco Oil Co. v. Dingwell, 690 F. Supp. 78, 86 (D. Me. 1988).

18. See infra notes 20-21 and accompanying text.

19. Burt & Sanoff, supra note 12, at 213.

20. Squire correctly states that "there continue to be few judicial decisions containing guidance on the appropriate allocation of liability among PRPs." Squire, supra note 9, at 1048.

21. See United States v. R.W. Meyer, Inc., 932 F.2d at 572, 21 ELR at 21064; Weyerhauser Co. v. Koppers Co., 771 F. Supp. at 1426, 22 ELR at 20170; United States v. Alcan Aluminum Corp., No. 87-CV-920, 1991 U.S. Dist. LEXIS 18721, at 10 (N.D.N.Y. Dec. 27, 1991).

22. 690 F. Supp. 78 (D. Me. 1988).

23. Id.* at 86.

24. Id.

25. 19 CHEM. WASTE LIT. REP. (Computer L. Rep., Inc.) 1310 (E.D. Pa. Dec. 29, 1989).

26. Id. at 1321-23.

27. Id. at 1323.

28. 932 F.2d 568, 21 ELR 21062 (6th Cir. 1991).

29. Id. at 571, 21 ELR at 21063.

30. Id.

31. Id. at 573, 21 ELR at 21064.

32. No. 87-CV-920, 1991 U.S. Dist. LEXIS 18721, at 1 (N.D.N.Y. Dec. 27, 1991).

33. Id.* at 2.

34. Id.* at 13-14.

35. 932 F.2d 568, 21 ELR 21062 (6th Cir. 1991).

36. Id.* at 572-73, 21 ELR at 21064.

37. 771 F. Supp. 1420, 22 ELR 20168 (D. Md. 1991).

38. Id. at 1426, 22 ELR at 20170.

39. Id. at 1427, 22 ELR at 20171.

40. 1988 HAZ. WASTE LITIG. REP. (Andrews Publications) 13,606 (E.D. Pa. Sept. 30, 1988).

41. Id. at 13,619.

42. Id. at 13,619-20.

43. Id. at 13,632-33.

44. No. 88-8038, 1989 U.S. Dist. LEXIS 17555, at 1 (S.D. Fla. Feb. 14, 1989).

45. Id.* at 11.

46. Id.* Notably, the court rejected the Gore factors, because "they provide little or no assistance in allocating costs between a landowner and a generator. They were clearly intended to distinguish among operators and generators." Id. at 6, n.2.

47. No. 90-0718, 1991 U.S. Dist. LEXIS 18750, at 1 (E.D. Pa. Dec. 16, 1991).

48. Id. at 2-4.

49. Id.* at 8.

50. Id.* at 10.

51. Id.* at 12.

52. 955 F.2d 519, 22 ELR 21005 (8th Cir. 1992).

53. Id.*

54. 775 F. Supp. 1227 (E.D. Mo. 1991).

55. Id. at 1234-35.

56. See United States v. Monsanto Co., 858 F.2d 160, 173 n.29, 19 ELR 20085, 20090 n.29 (4th Cir. 1988); Environmental Transp. Sys., Inc. v. Ensco, Inc., 763 F. Supp. 384 (C.D. Ill. 1991); Versatile Metals, Inc. v. Union Corp., 693 F. Supp. 1563, 1572, 19 ELR 20472, 20475 (E.D. Pa. 1988).

57. 763 F. Supp. 384 (C.D. Ill. 1991).

58. Id. at 388, 392.

59. No. C89-252M (W.D. Wash. Aug. 19, 1992).

60. The allocation in Western Processing concerned a minor portion of the total response costs and involved only four out of the original 250 parties. The claims for the overwhelming majority of costs had been settled by most of the other PRPs on the eve of trial. None of the remaining parties had developed any allocation approach for the remaining costs and all of them proposed that the court apply a simple volumetric allocation method, which the court did.

61. United States v. Alcan Aluminum Corp., No. 87-CV-920, 1991 U.S. Dist. LEXIS 18721, at 1 (N.D.N.Y. Dec. 27, 1991).

62. See, e.g., id.* at 10. In rejecting toxicity as a basis for allocation, the court stated that "[u]nder the circumstances, it would be inequitable for the court to split scientific hairs." Id.*

63. See, e.g., id.

64. See, e.g., BCW Assoc. v. Occidental Chem. Corp., 1988 HAZ. WASTE LITIG. REP. (Andrews Publications) 13,606 (E.D. Pa. Sept. 30, 1988).

65. This principle is firmly established in public utility regulation. FERC has expressed this key point as follows: "[a]llocating particular costs to those responsible for them — where that can be done feasibly — is fair, and altogether consistent with the public interest." 53 FERC P63,020, 65,222 (1990). FERC has also said that "[p]rinciples of fairness in ratemaking support the concept that those who are responsible for the incurrence of costs be the ones who bear those cost burdens." System Energy Resources, Inc., 41 FERC P61,238, 61,616 (1987). Classic economic treatises that explore the concept of reflecting cost causation in public utility rates are JAMES C. BONBRIGHT, PRINCIPLES OF PUBLIC UTILITY RATES (1961), and ALFRED E. KAHN, THE ECONOMICS OF REGULATION: PRINCIPLES AND INSTITUTIONS (1971).

66. The phrase "attributable cost" is used in postal rate regulation. "Direct" and "variable" costs are common in accounting analysis.

67. This principle is also expressed in the concept of "separable costs" in water resources. See ALVIN S. GOODMAN, PRINCIPLES OF WATER RESOURCE PLANNING 229-32 (1984).

68. KAHN, supra note 65, at 85; William J. Bawmol et al., How Arbitrary Is "Arbitrary"? - or Toward the Deserved Demise of Full Cost Allocation, PUB. UTIL. FORTNIGHTLY, Sept. 3, 1987, at 16-21.

69. This concept of equitable cost allocation is not new. James C. Loughlin has written that

[t]he "fairness" concept of equity in cost allocation has been emphasized by the [Federal Inter-Agency River Basin Committee's] Subcommittee on Benefits and Costs: "The objective of cost allocation is to distribute project costs equitably among the purposes served … by providing for proportional sharing of the savings resulting from multiple-purpose development" … [SUBCOMMITTEE ON BENEFTTS AND COSTS, FEDERAL INTER-AGENCY RIVER BASIN COMMITTEE, PROPOSED PRACTICES FOR ECONOMIC ANALYSIS OF RIVER BASIN PROJECTS 53 (1950)]. A similar notion of equity in cost allocation is provided in … [U.S. BUREAU OF THE BUDGET, CIRCULAR A-47, at 11 (1952)] and … [S. DOC. NO. 97, 87th Cong., 2d Sess. 12 (1962)].

The Efficiency and Equity of Cost Allocation Methods for Multipurpose Water Projects, 13 WATER RESOURCES RES. 11 (1977).

70. See supra notes 28, 32, 35, 37, 40, 44 and accompanying text.

71. See supra note 69 and accompanying text.

72. The following discussion compares each approach in the context of generator share liability allocation. The conclusions on the relative merits of each approach would be substantially the same for owneroperator or transporter liability allocation.

73. See supra note 65 and accompanying text.

74. See supra notes 28, 32, 35, 37, 40, 44 and accompanying text.

75. EPA, OSWER DIR. No. 9839.1, INTERIM GUIDELINES FOR PREPARING NONBINDING PRELIMINARY ALLOCATIONS OF RESPONSIBILITIES 3 (1987), ELR ADMIN. MATERIALS 35065.

EPA considered and rejected models based on toxicity because of the complexity of their application and the lack of agreement among the scientific community about degrees of toxicity of specific hazardous substances and synergistic affects. Also, toxicity is usually casually related to the cost of cleanup for only a few substances (e.g., PCBs, dioxin).

76. G&E Engineering, Inc., Abandoned Waste Site Cleanup Cost Allocation (CCA) Model, 7 CHEM. WASTE LIT. REP. 398 (1984). An approach similar to the G&E Engineering approach was offered at the Western Processing site. See deposition testimony of G. Graham Allen, United States v. Western Processing Co., Case No. C89-214M (W.D. Wash. Aug. 19, 1992).

77. See G&E Engineering, Inc., supra note 76, at 509-10. One of the authors disputed application of this model to allocation of costs at the Western Processing site. See deposition of John C. Butler III, United States v. Western Processing Co., Case No. C89-214M (W.D. Wash. Aug. 19, 1992) [hereinafter Butler deposition].

78. Both versions of the Kellogg model were applied in Western Processing. See Butler deposition, supra note 77.

79. See supra note 68 and accompanying text.

80. See supra note 68 and accompanying text.

81. See supra notes 78-80 and accompanying text.

82. See supra notes 68-69 and accompanying text.

83. In that period there was a national controversy over what constituted a fair distribution of the costs of the Tennessee Valley Authority (TVA) project. In particular, the investor-owned utilities asserted that TVA's electric power was subsidized. See JOSEPH S. RANSMEIER, THE TENNESSEE VALLEY AUTHORITY (1942). An authority on the application of the SAC concept to railroad regulation has explained that

[t]he concept of stand-alone costs was developed in the late 1930's to devise a fair system for allocating the common costs of TVA's multiple-purpose dams, which were designed to provide electric power, navigation, and flood control. The fundamental premise was that the users of a single function, such as navigation, should not be required to pay more for their service than if they were served by a dam constructed only for that specific purpose.

Merrill J. Roberts, Residual Railroad Rate Control: The Unmet Challenge of Deregulation, 23 LOGISTICS & TRANSP. REV. 91 (1987).

84. See depositions of George R. Hall & Henry Landau, United States v. Western Processing Co., Case No. C89-214M, slip op. (W.D. Wash. Aug. 19, 1992); Butler deposition, supra note 77.

85. RANSMEIER, supra note 83, at 277.

86. RANSMEIER, supra note 83, at 303; H.P. Young et al., Cost Allocation in Water Resources Development, 18 WATER RESOURCES RES. 463 (1982).

87. Gerald R. Faulhaber, Cross-Subsidization: Pricing in Public Enterprises, 54 AM. ECON. REV. 966 (1975).

88. This example is adapted from Faulhaber, supra note 87.

89. See EPA Policy to Ease Cities' Superfund Costs Appears To Be Election Year Hostage, INSIDE EPA WKLY. REP., June 5, 1992, at 9.

90. EPA's application of the stand-alone cost concept is flawed in at least three ways. First, EPA's closure cost estimates are understated for MSW sites, since they assume significantly less stringent closure requirements for MSW sites than for industrial waste sites. Second, EPA's calculation assumes equal volumes of MSW and industrial waste. This understates the MSW relative contribution to costs when MSW exceeds 50 percent of waste volume. (The converse is true when MSW is less than 50 percent of volume, holding all else equal.) Third, the selection of the discount rate is biased so that costs of remedial action for industrial wastes are increased compared to MSW cleanup costs.