- Email: [email protected]
Allocation by Contribution to Cost and Risk at Superfund Sites
Journal of Environmental Forensics (2000) 1, 117±120 doi:10.1006/enfo.2000.0013, available online at http://www.idealibrary.com on
Allocation by Contribution to Cost and Risk at Superfund Sites Brian L. Murphy* Sciences International, Inc., ComCenter 70, 6150 State Road, Bradenton, FL 34203, U.S.A. (Received 15 May 2000, revised manuscript accepted 13 June 2000) Because there is no speci®c legislative or regulatory guidance, there is no ``right'' way to allocate liability at Superfund sites. Allocation based on the cost of a remedy, and allocation based on the need for a remedy, i.e. risk-based allocation, represent two possibilities. Other allocation schemes can be located between these two philosophical poles. When waste streams and environmental impacts are qualitatively similar, allocation based solely on costs may make the most sense. When one or more potentially responsible parties (PRPs) have qualitatively dierent wastes or impacts, an allocation scheme based on both contribution to cost and to risk may be able to incorporate all PRPs. In any case, dissident PRPs, whose contribution to remedy costs is large but whose contribution to risk is small, may ®nd satisfaction in the courts # 2000 AEHS where there is precedent for risk-based allocation. Keywords: allocation; Superfund; risk.
Introduction
selection of equitable factors was found to be subject to an ``abuse of discretion'' standard of review. The application of those factors was found to be subject to a ``clear error'' standard of review. The equitable factors most often cited are the Gore factors, proposed by the then Representative Albert Gore in 1980, but not enacted. The Gore factors are: (1) the ability to distinguish the party's contribution to the nature and extent of the problem; (2) the degree of the party's involvement in the activities that caused the problem; (3) the degree of care exercised by the party; (4) the degree of cooperation of the party with governmental agencies; (5) the quantity of the hazardous waste involved; and (6) the toxicity of the waste. These factors have been found to be far from sucient, and in some cases may not be applicable at all. Furthermore, they are simply a list; they provide no conceptual framework for allocation. Thus CERCLA and its legislative history are not particularly helpful in specifying how liability is to be allocated. Courts have generally determined that there is a presumption of ``joint and several liability'' unless harm is distinct, or there is a reasonable basis for its division. What is ``harm'' at a Superfund site? It seems logical that it be closely identi®ed with the concept of ``risk''. A baseline risk assessment is conducted at all Superfund sites. Removal actions may precede completion of the risk assessment for urgent matters, but the continued remediation of a site is based on a ®nding that the computed baseline risks, either human or ecological, are unacceptable. Therefore, it is logical to identify the risks at a site, including those requiring removal action, with ``harm''. As discussed below, several Federal circuit court rulings also support this notion. Thus, an argument can be made that risk assessment is the appropriate tool to use in apportioning liability. However, in fact, risk is often not a consideration in
Two sections of the Comprehensive Environmental Response, Compensation and Liability Act or CERCLA, commonly referred to as Superfund, touch on liability allocation. Section 107 provides for recovery of remediation costs. Plaintis in a recovery action may be the US Environmental Protection Agency (EPA), or they may be a state. Courts are divided on whether a Potentially Responsible Party (PRP) may be a plainti. According to Aronovsky (2000), the recent trend has been to deny Section 107 to PRP plaintis. Although Section 107 speci®es joint and several liability, this is discretionary with the court. In particular, where a PRP can demonstrate distinct harm or divisibility of harm, that party may be responsible just for their contribution to harm. A distinct harm arises, for example, when there are separate groundwater plumes or areas of surface soil contamination. A divisible harm might be where there are successive site owners conducting the same operation. The basis for divisibility in that case might be years of operation. Section 113 allows a party who has incurred response costs to seek contribution from other PRPs. This section also provides contribution protection for parties that have settled with the government. Under Section 113, the liability of nonsettling PRPs is limited to their proportionate share. The nonsettling PRPs' liability may be determined in either of two ways: it may be determined by subtracting out the amount of prior settlements; or it may be determined by subtracting out the proportionate share of the harm for the settling PRPs. As Ferrey (1994) points out, the results of these two approaches may be quite dierent. There is no guidance in determining proportionate shares, beyond citing ``equitable factors'', in CERCLA. In a March 24, 2000 decision by the 9th Circuit Court, Boeing Company v. Cascade Corporation, the *E-mail: [email protected]
117 1527-5922/00/030117+04 $35.00/00
# 2000 AEHS
118 B.L. Murphy
apportioning liability. If the PRPs at a Superfund site agree on an allocation scheme, then that scheme is by de®nition satisfactory, assuming that there is no ``second guessing'' by other parties such as insurers. PRPs often decide to allocate liability based on the contribution of each to the cost of the remedy. Of course, surrogate measures for estimating contribution to costs may be used, such as counting barrels or estimating plume areas. In light of the above, there is no ``right'' way to allocate costs. However, there are approaches that have been tested in court, which are likely to reduce costs for individual PRPs. These approaches are grounded in the idea that ``harm'' is what creates the need for a remedy and may be identi®ed as risk. This paper presents a model that may be useful to PRPs in thinking about cost allocation and the possibility of cost reduction. The model represents a framework for balancing the points of view of PRPs whose contribution to a site may be qualitatively dissimilar.
Basis for Risk-based Apportionment Risk-based methods have a foundation in common law concepts of ``contribution to harm''. The cases below, discussed in greater detail by Rockwood and Harrison (1993), deal with the requirements of establishing distinct or divisible harm. In the process, they illuminate what ``harm'' is in a Superfund context. In United States v. Alcan Aluminum Corporation, 964 F.2d 711 (3rd Circuit Court, 1992) (hereafter known as Alcan 1992), the court rejected the Government's argument that commingled wastes are necessarily indivisible, and held that if Alcan could prove that its waste did not contribute to the need for a remedy then Alcan would have no liability for remedy costs. Alcan argued that although its waste emulsion contained trace quantities of metals that were CERCLA hazardous materials, they were present below the naturally occurring background level. The 3rd Circuit Court took note of United States v. Chemdyne Corp., 572 F. Supp. 802 (S.D. Ohio 1983), in reaching its conclusions. In Chemdyne, the court turned to the Restatement (Second) of Torts, which states ``Damages for harm are to be apportioned among two or more causes where (a) there are distinct harms, or (b) there is a reasonable basis for determining the contribution of each cause to a single harm''. The court also noted that volume of waste alone is not an accurate predictor of harm because substances may dier in their toxicity and migratory potential, thereby calling attention to precisely the factors that enter into a risk assessment. The waste involved in United States v. Alcan Aluminum Corp., 990 F.2d 711 (2nd Circuit Court, 1993) (Alcan 1993) was the same as in Alcan 1992. The court rearmed that Alcan could escape any liability if it showed that its emulsion did not contribute to the cleanup costs when mixed with other hazardous wastes or, alternatively, that cost apportionment could be based on the relative contribution to harm. The 2nd Circuit Court said that Alcan had the burden of demonstrating the separability of harm and that this would require ``an assessment of the relative toxicity,
migratory potential, and synergistic capacity of the hazardous waste at issue''. Once again, this is a description of the elements of a risk assessment. The Alcan decisions upheld lower court decisions stating that there is no threshold concentration involved in being named a PRP. That is, disposal of wastes with chemical concentrations that are below ambient levels is not a defense against being named a PRP. However, once named a PRP, it is still the case that if the harm is separable, an ambient concentration defense may eliminate any liability. Finally, In re Bell Petroleum 3 F.3d 889 (5th Circuit Court, 1993), the court held that a preponderance of evidence was sucient in establishing divisibility. The court stated: ``The fact that apportionment may be dicult, because each defendant's exact contribution to the harm cannot be proved to a certainty, or the fact that it will require weighing the evidence and making credibility determinations, are inadequate grounds upon which to impose joint and several liability''. In essence, the 5th Circuit Court ruled against EPA's argument that because dierent allocation schemes provided dierent results, no allocation was possible. In remanding the case to the district court, the 5th Circuit Court ruled that a PRP need only provide a reasonable basis for apportioning liability.
The Bipolar Allocation Model Even if risk assessment is the appropriate tool for dividing harm, it is overkill in many cases, particularly those where there is a commonality of waste. There are several common situations where the harm due to multiple PRPs is not distinct but requires division. For example, consider the situations of: (1) commingled groundwater plumes; (2) hazardous waste disposal sites accepting a variety of wastes; and (3) successive ownership of a site. If these situations result in contamination by similar chemicals, a straightforward allocation based on contribution to the cost of a remedy may make sense. However, when one or more PRPs' wastes dier signi®cantly from the others in the risk they pose, those PRPs may wish to consider a riskbased approach. The basic model is shown in Figure 1. Allocation based on the need for a remedy and allocation based on the cost of a remedy represent two poles, with a speci®c allocation aligned anywhere in between. Economic principles of cost allocation, including the stand-alone cost method, have been discussed by Butler et al. (1993) and by Wise, Maniatis and Koch (1997). The cost allocation matrix approach of Hall, Harris and Reinsdorf (1994) is also based on determining contribution to the cost of a remedy. Marryott
Need for a remedy
Cost of a remedy One solution to liability allocation
Figure 1. The Bipolar Allocation Model.
Allocation by Contribution to Cost and Risk at Superfund Sites 119
et al. (2000) present a stand-alone cost type model in which a weighted sum of contaminant mass in the plume, and plume volume, serve as a surrogate for remediation costs. The basic equation for calculating stand-alone costs is: SACi fi X SACi
1
where SACi is the stand-alone cost for the waste stream due to ith PRP generator/transporter. This equation does not address how liability is to be allocated between the generator, transporter and site owner, nor does it address orphan shares, such as from unidenti®ed or defunct parties. It is solely an allocation by waste stream. Equation (1) states that each PRP pays in proportion to the cost that would have been incurred if there were no other PRPs at the site. Because of redundancy of cost items and economies of scale, the total cost of a remedy will generally be less than the denominator of equation (1), and hence each PRP will actually pay less than their computed standalone cost. Risk-based allocation methods have been discussed by Murphy (1996) and by Mink, Nash and Coleman (1997). The risk contribution analogue to equation (1) is: SARi gi X SARi
2
where gi is the allocation fraction for the ith generator/ transporter based on the stand-alone contribution to risk SARi : The analogy with stand-alone costs is incomplete, however; the total risk is equal to the sum of the individual PRP-caused risks rather than being less. Of course, cost allocation may be a mixture of costbased and risk-based: hi afi 1 ÿ agi
3
where a is a constant. As a decreases from one, a ``contribution to the need for a remedy'' component is mixed in with the ``contribution to the cost of a remedy''. The kind of information needed to calculate fi or gi diers. For example, in computing stand-alone costs, well installation costs may vary as plume area, and groundwater treatment costs may vary as contaminant mass in the plume. How long a pump and treat remedy needs to be maintained will depend upon the ratio of individual chemical concentrations to acceptable levels in groundwater, and on chemical properties that determine partitioning to soil. In computing standalone risks, concentrations and toxicities of speci®c chemicals will be required. Of course, as indicated above, it may be to the advantage of all PRPs to lower transaction costs by using surrogate quantities, rather than by attempting to collect the additional information that is necessary for re®ned or precise calculations. The key question for PRPs is where on the axis between the two poles in Figure 1 do they want to be? Where one PRP has contributed large volumes of
relatively innocuous waste, that PRP may favor a riskbased approach, based on the contribution to the need for a remedy. Thus, at least one answer to the question of ``How far toward the risk pole should the allocation formula be?'' is ``Far enough to bring in the PRPs whose waste was relatively innocuous''. Accommodating that PRP makes sense because it is the risk-based, rather than the cost-based methods, which have a common law foundation.
Case Studies Two case studies are included to illustrate how a riskbased approach can be useful when disparate wastes are involved, or when there are dissenting PRPs. US v. Atlas Minerals and Chemicals, Inc. This case is an example of considering both contribution to the need for a remedy, and contribution to the cost of a remedy, when PRPs dispose of disparate wastes. Both municipal solid waste (MSW) and industrial waste had been disposed at the Dorney Road Land®ll near Allentown, PA. The court found (41 ERC 1417, DC EPA, 1995) that the low-toxicity, high-volume MSW streams had created most of the remedy costs and that the high-toxicity, low-volume waste streams had created most of the need for a remedy. On this basis, costs were divided between the MSW and industrial parties. (The court also divided costs associated with the City of Allentown's MSW waste stream between the City and the transporter on the basis that the disposal site had been mutually selected.) Within each of the two categories costs were allocated by ``waste strength score'' weighted volume. Waste strength score appears to be intended as a surrogate for risk. The waste strength score of an individual waste stream was de®ned as the sum of the waste strength scores for the organics and inorganics in the waste stream, plus a physical state score. Liquid wastes had a higher physical state score than solid waste. The waste strength score of organic and inorganic chemicals was de®ned as a score for their concentration in individual waste streams multiplied by a toxicity/mobility score. Lowry Land®ll The author assisted two PRPs at the Lowry Land®ll outside Denver who had disposed of large volumes of innocuous waste. Several risk-based arguments led to a successful settlement with the other PRPs. For both PRPs the hazardous materials in the waste; arsenic, lead and cadmium, were present at levels below the naturally occurring levels in soils at the site. Although, as noted above in our discussion of the Alcan decisions, this is not a basis to avoid being named a PRP, it is relevant to issues of distinct or divisible harm. In the case of one waste, which had been disposed of in dry form, it was possible to show that net in®ltration of precipitation was insucient to carry hazardous materials to groundwater. The other waste had been disposed of as a slurry and so groundwater
120 B.L. Murphy
contamination had to be considered. EPA had de®ned acceptable concentrations at a compliance point boundary. Thus, for each chemical we were able to de®ne a ``noncompliance'' quotient as the ratio of an actual concentration in groundwater to what EPA had deemed an acceptable concentration. Adding noncompliance quotients, we were then able to de®ne a noncompliance index both for our client and for the site as a whole. The noncompliance quotient and index are analogous to the hazard quotient and index that EPA uses to de®ne risk. The noncompliance index is a measure of the need for a remedy. It turned out that our client's waste had contributed about two parts in a million to the site noncompliance index. Risk assessments do not ordinarily deal with the element of time. However, in this case we noted that our client's contaminants would not arrive at the compliance boundary for 1100 years because of the slow groundwater velocity and because of retardation. Thus, the question was raised as to whether the remedy was relevant to this PRP's waste at all. If one believed the remedy would still be relevant at that distant point in time, could payment be delayed until then? In that regard we noted that just $1,130 placed in trust at 1% would equal the total remedy cost of $65,000,000 in 1100 years.
References Aronovsky, R.G. 2000. New Trends in Groundwater Contamination Litigation. In: Proceedings: Environmental Litigation: Advanced Forensics and Legal Strategies. University of Wisconsin, College of Engineering and Professional Development, San Francisco, CA. April 13±14, 2000. Butler, J.C. III, Schneider, M.W., Hall, G.R. and Burton, M.E. 1993. Allocating superfund costs: cleaning up the controversy. Env. Law Reporter 23, 10133±10144. Ferrey, S. 1994. The new wave: Superfund allocation strategies and outcomes. BNA Environmental Reporter 25, 790±803. Hall, R.M., Harris, R.H. and Reinsdorf, J.A. 1994. Superfund response cost allocations: The law, the science and the practice. The Business Lawyer 49, 1489±1540. Marryott, R.A., Sabadell, G.P., Ahlfeld, D.P., Harris, R.H. and Pinder, G.F. 2000. Allocating remedial costs at superfund sites with commingled groundwater contaminant plumes. Environmental Forensics 1(1), 47±54. Mink, F.L., Nash, D.E. and Coleman, J.C. II 1997. Superfund site contamination: Apportionment of liability. Natural Resources and Environment 12, 68±79. Murphy, B.L. 1996. Risk assessment as a liability allocation tool. Env. Claims Jour 8, 129±144. Rockwood, L.L. and Harrison, J.L. 1993. The Alcan decisions: Causation through the back door. Environmental Law Reporter 25, ELR 10542. Wise, K.T., Maniatis, M.A. and Koch, G.S. 1997. Allocating CERCLA liabilities: The applications and limitations of economics. BNA Toxics Law Reporter 11, 830±833.
Allocation by Contribution to Cost and Risk at Superfund Sites Brian L. Murphy* Sciences International, Inc., ComCenter 70, 6150 State Road, Bradenton, FL 34203, U.S.A. (Received 15 May 2000, revised manuscript accepted 13 June 2000) Because there is no speci®c legislative or regulatory guidance, there is no ``right'' way to allocate liability at Superfund sites. Allocation based on the cost of a remedy, and allocation based on the need for a remedy, i.e. risk-based allocation, represent two possibilities. Other allocation schemes can be located between these two philosophical poles. When waste streams and environmental impacts are qualitatively similar, allocation based solely on costs may make the most sense. When one or more potentially responsible parties (PRPs) have qualitatively dierent wastes or impacts, an allocation scheme based on both contribution to cost and to risk may be able to incorporate all PRPs. In any case, dissident PRPs, whose contribution to remedy costs is large but whose contribution to risk is small, may ®nd satisfaction in the courts # 2000 AEHS where there is precedent for risk-based allocation. Keywords: allocation; Superfund; risk.
Introduction
selection of equitable factors was found to be subject to an ``abuse of discretion'' standard of review. The application of those factors was found to be subject to a ``clear error'' standard of review. The equitable factors most often cited are the Gore factors, proposed by the then Representative Albert Gore in 1980, but not enacted. The Gore factors are: (1) the ability to distinguish the party's contribution to the nature and extent of the problem; (2) the degree of the party's involvement in the activities that caused the problem; (3) the degree of care exercised by the party; (4) the degree of cooperation of the party with governmental agencies; (5) the quantity of the hazardous waste involved; and (6) the toxicity of the waste. These factors have been found to be far from sucient, and in some cases may not be applicable at all. Furthermore, they are simply a list; they provide no conceptual framework for allocation. Thus CERCLA and its legislative history are not particularly helpful in specifying how liability is to be allocated. Courts have generally determined that there is a presumption of ``joint and several liability'' unless harm is distinct, or there is a reasonable basis for its division. What is ``harm'' at a Superfund site? It seems logical that it be closely identi®ed with the concept of ``risk''. A baseline risk assessment is conducted at all Superfund sites. Removal actions may precede completion of the risk assessment for urgent matters, but the continued remediation of a site is based on a ®nding that the computed baseline risks, either human or ecological, are unacceptable. Therefore, it is logical to identify the risks at a site, including those requiring removal action, with ``harm''. As discussed below, several Federal circuit court rulings also support this notion. Thus, an argument can be made that risk assessment is the appropriate tool to use in apportioning liability. However, in fact, risk is often not a consideration in
Two sections of the Comprehensive Environmental Response, Compensation and Liability Act or CERCLA, commonly referred to as Superfund, touch on liability allocation. Section 107 provides for recovery of remediation costs. Plaintis in a recovery action may be the US Environmental Protection Agency (EPA), or they may be a state. Courts are divided on whether a Potentially Responsible Party (PRP) may be a plainti. According to Aronovsky (2000), the recent trend has been to deny Section 107 to PRP plaintis. Although Section 107 speci®es joint and several liability, this is discretionary with the court. In particular, where a PRP can demonstrate distinct harm or divisibility of harm, that party may be responsible just for their contribution to harm. A distinct harm arises, for example, when there are separate groundwater plumes or areas of surface soil contamination. A divisible harm might be where there are successive site owners conducting the same operation. The basis for divisibility in that case might be years of operation. Section 113 allows a party who has incurred response costs to seek contribution from other PRPs. This section also provides contribution protection for parties that have settled with the government. Under Section 113, the liability of nonsettling PRPs is limited to their proportionate share. The nonsettling PRPs' liability may be determined in either of two ways: it may be determined by subtracting out the amount of prior settlements; or it may be determined by subtracting out the proportionate share of the harm for the settling PRPs. As Ferrey (1994) points out, the results of these two approaches may be quite dierent. There is no guidance in determining proportionate shares, beyond citing ``equitable factors'', in CERCLA. In a March 24, 2000 decision by the 9th Circuit Court, Boeing Company v. Cascade Corporation, the *E-mail: [email protected]
117 1527-5922/00/030117+04 $35.00/00
# 2000 AEHS
118 B.L. Murphy
apportioning liability. If the PRPs at a Superfund site agree on an allocation scheme, then that scheme is by de®nition satisfactory, assuming that there is no ``second guessing'' by other parties such as insurers. PRPs often decide to allocate liability based on the contribution of each to the cost of the remedy. Of course, surrogate measures for estimating contribution to costs may be used, such as counting barrels or estimating plume areas. In light of the above, there is no ``right'' way to allocate costs. However, there are approaches that have been tested in court, which are likely to reduce costs for individual PRPs. These approaches are grounded in the idea that ``harm'' is what creates the need for a remedy and may be identi®ed as risk. This paper presents a model that may be useful to PRPs in thinking about cost allocation and the possibility of cost reduction. The model represents a framework for balancing the points of view of PRPs whose contribution to a site may be qualitatively dissimilar.
Basis for Risk-based Apportionment Risk-based methods have a foundation in common law concepts of ``contribution to harm''. The cases below, discussed in greater detail by Rockwood and Harrison (1993), deal with the requirements of establishing distinct or divisible harm. In the process, they illuminate what ``harm'' is in a Superfund context. In United States v. Alcan Aluminum Corporation, 964 F.2d 711 (3rd Circuit Court, 1992) (hereafter known as Alcan 1992), the court rejected the Government's argument that commingled wastes are necessarily indivisible, and held that if Alcan could prove that its waste did not contribute to the need for a remedy then Alcan would have no liability for remedy costs. Alcan argued that although its waste emulsion contained trace quantities of metals that were CERCLA hazardous materials, they were present below the naturally occurring background level. The 3rd Circuit Court took note of United States v. Chemdyne Corp., 572 F. Supp. 802 (S.D. Ohio 1983), in reaching its conclusions. In Chemdyne, the court turned to the Restatement (Second) of Torts, which states ``Damages for harm are to be apportioned among two or more causes where (a) there are distinct harms, or (b) there is a reasonable basis for determining the contribution of each cause to a single harm''. The court also noted that volume of waste alone is not an accurate predictor of harm because substances may dier in their toxicity and migratory potential, thereby calling attention to precisely the factors that enter into a risk assessment. The waste involved in United States v. Alcan Aluminum Corp., 990 F.2d 711 (2nd Circuit Court, 1993) (Alcan 1993) was the same as in Alcan 1992. The court rearmed that Alcan could escape any liability if it showed that its emulsion did not contribute to the cleanup costs when mixed with other hazardous wastes or, alternatively, that cost apportionment could be based on the relative contribution to harm. The 2nd Circuit Court said that Alcan had the burden of demonstrating the separability of harm and that this would require ``an assessment of the relative toxicity,
migratory potential, and synergistic capacity of the hazardous waste at issue''. Once again, this is a description of the elements of a risk assessment. The Alcan decisions upheld lower court decisions stating that there is no threshold concentration involved in being named a PRP. That is, disposal of wastes with chemical concentrations that are below ambient levels is not a defense against being named a PRP. However, once named a PRP, it is still the case that if the harm is separable, an ambient concentration defense may eliminate any liability. Finally, In re Bell Petroleum 3 F.3d 889 (5th Circuit Court, 1993), the court held that a preponderance of evidence was sucient in establishing divisibility. The court stated: ``The fact that apportionment may be dicult, because each defendant's exact contribution to the harm cannot be proved to a certainty, or the fact that it will require weighing the evidence and making credibility determinations, are inadequate grounds upon which to impose joint and several liability''. In essence, the 5th Circuit Court ruled against EPA's argument that because dierent allocation schemes provided dierent results, no allocation was possible. In remanding the case to the district court, the 5th Circuit Court ruled that a PRP need only provide a reasonable basis for apportioning liability.
The Bipolar Allocation Model Even if risk assessment is the appropriate tool for dividing harm, it is overkill in many cases, particularly those where there is a commonality of waste. There are several common situations where the harm due to multiple PRPs is not distinct but requires division. For example, consider the situations of: (1) commingled groundwater plumes; (2) hazardous waste disposal sites accepting a variety of wastes; and (3) successive ownership of a site. If these situations result in contamination by similar chemicals, a straightforward allocation based on contribution to the cost of a remedy may make sense. However, when one or more PRPs' wastes dier signi®cantly from the others in the risk they pose, those PRPs may wish to consider a riskbased approach. The basic model is shown in Figure 1. Allocation based on the need for a remedy and allocation based on the cost of a remedy represent two poles, with a speci®c allocation aligned anywhere in between. Economic principles of cost allocation, including the stand-alone cost method, have been discussed by Butler et al. (1993) and by Wise, Maniatis and Koch (1997). The cost allocation matrix approach of Hall, Harris and Reinsdorf (1994) is also based on determining contribution to the cost of a remedy. Marryott
Need for a remedy
Cost of a remedy One solution to liability allocation
Figure 1. The Bipolar Allocation Model.
Allocation by Contribution to Cost and Risk at Superfund Sites 119
et al. (2000) present a stand-alone cost type model in which a weighted sum of contaminant mass in the plume, and plume volume, serve as a surrogate for remediation costs. The basic equation for calculating stand-alone costs is: SACi fi X SACi
1
where SACi is the stand-alone cost for the waste stream due to ith PRP generator/transporter. This equation does not address how liability is to be allocated between the generator, transporter and site owner, nor does it address orphan shares, such as from unidenti®ed or defunct parties. It is solely an allocation by waste stream. Equation (1) states that each PRP pays in proportion to the cost that would have been incurred if there were no other PRPs at the site. Because of redundancy of cost items and economies of scale, the total cost of a remedy will generally be less than the denominator of equation (1), and hence each PRP will actually pay less than their computed standalone cost. Risk-based allocation methods have been discussed by Murphy (1996) and by Mink, Nash and Coleman (1997). The risk contribution analogue to equation (1) is: SARi gi X SARi
2
where gi is the allocation fraction for the ith generator/ transporter based on the stand-alone contribution to risk SARi : The analogy with stand-alone costs is incomplete, however; the total risk is equal to the sum of the individual PRP-caused risks rather than being less. Of course, cost allocation may be a mixture of costbased and risk-based: hi afi 1 ÿ agi
3
where a is a constant. As a decreases from one, a ``contribution to the need for a remedy'' component is mixed in with the ``contribution to the cost of a remedy''. The kind of information needed to calculate fi or gi diers. For example, in computing stand-alone costs, well installation costs may vary as plume area, and groundwater treatment costs may vary as contaminant mass in the plume. How long a pump and treat remedy needs to be maintained will depend upon the ratio of individual chemical concentrations to acceptable levels in groundwater, and on chemical properties that determine partitioning to soil. In computing standalone risks, concentrations and toxicities of speci®c chemicals will be required. Of course, as indicated above, it may be to the advantage of all PRPs to lower transaction costs by using surrogate quantities, rather than by attempting to collect the additional information that is necessary for re®ned or precise calculations. The key question for PRPs is where on the axis between the two poles in Figure 1 do they want to be? Where one PRP has contributed large volumes of
relatively innocuous waste, that PRP may favor a riskbased approach, based on the contribution to the need for a remedy. Thus, at least one answer to the question of ``How far toward the risk pole should the allocation formula be?'' is ``Far enough to bring in the PRPs whose waste was relatively innocuous''. Accommodating that PRP makes sense because it is the risk-based, rather than the cost-based methods, which have a common law foundation.
Case Studies Two case studies are included to illustrate how a riskbased approach can be useful when disparate wastes are involved, or when there are dissenting PRPs. US v. Atlas Minerals and Chemicals, Inc. This case is an example of considering both contribution to the need for a remedy, and contribution to the cost of a remedy, when PRPs dispose of disparate wastes. Both municipal solid waste (MSW) and industrial waste had been disposed at the Dorney Road Land®ll near Allentown, PA. The court found (41 ERC 1417, DC EPA, 1995) that the low-toxicity, high-volume MSW streams had created most of the remedy costs and that the high-toxicity, low-volume waste streams had created most of the need for a remedy. On this basis, costs were divided between the MSW and industrial parties. (The court also divided costs associated with the City of Allentown's MSW waste stream between the City and the transporter on the basis that the disposal site had been mutually selected.) Within each of the two categories costs were allocated by ``waste strength score'' weighted volume. Waste strength score appears to be intended as a surrogate for risk. The waste strength score of an individual waste stream was de®ned as the sum of the waste strength scores for the organics and inorganics in the waste stream, plus a physical state score. Liquid wastes had a higher physical state score than solid waste. The waste strength score of organic and inorganic chemicals was de®ned as a score for their concentration in individual waste streams multiplied by a toxicity/mobility score. Lowry Land®ll The author assisted two PRPs at the Lowry Land®ll outside Denver who had disposed of large volumes of innocuous waste. Several risk-based arguments led to a successful settlement with the other PRPs. For both PRPs the hazardous materials in the waste; arsenic, lead and cadmium, were present at levels below the naturally occurring levels in soils at the site. Although, as noted above in our discussion of the Alcan decisions, this is not a basis to avoid being named a PRP, it is relevant to issues of distinct or divisible harm. In the case of one waste, which had been disposed of in dry form, it was possible to show that net in®ltration of precipitation was insucient to carry hazardous materials to groundwater. The other waste had been disposed of as a slurry and so groundwater
120 B.L. Murphy
contamination had to be considered. EPA had de®ned acceptable concentrations at a compliance point boundary. Thus, for each chemical we were able to de®ne a ``noncompliance'' quotient as the ratio of an actual concentration in groundwater to what EPA had deemed an acceptable concentration. Adding noncompliance quotients, we were then able to de®ne a noncompliance index both for our client and for the site as a whole. The noncompliance quotient and index are analogous to the hazard quotient and index that EPA uses to de®ne risk. The noncompliance index is a measure of the need for a remedy. It turned out that our client's waste had contributed about two parts in a million to the site noncompliance index. Risk assessments do not ordinarily deal with the element of time. However, in this case we noted that our client's contaminants would not arrive at the compliance boundary for 1100 years because of the slow groundwater velocity and because of retardation. Thus, the question was raised as to whether the remedy was relevant to this PRP's waste at all. If one believed the remedy would still be relevant at that distant point in time, could payment be delayed until then? In that regard we noted that just $1,130 placed in trust at 1% would equal the total remedy cost of $65,000,000 in 1100 years.
References Aronovsky, R.G. 2000. New Trends in Groundwater Contamination Litigation. In: Proceedings: Environmental Litigation: Advanced Forensics and Legal Strategies. University of Wisconsin, College of Engineering and Professional Development, San Francisco, CA. April 13±14, 2000. Butler, J.C. III, Schneider, M.W., Hall, G.R. and Burton, M.E. 1993. Allocating superfund costs: cleaning up the controversy. Env. Law Reporter 23, 10133±10144. Ferrey, S. 1994. The new wave: Superfund allocation strategies and outcomes. BNA Environmental Reporter 25, 790±803. Hall, R.M., Harris, R.H. and Reinsdorf, J.A. 1994. Superfund response cost allocations: The law, the science and the practice. The Business Lawyer 49, 1489±1540. Marryott, R.A., Sabadell, G.P., Ahlfeld, D.P., Harris, R.H. and Pinder, G.F. 2000. Allocating remedial costs at superfund sites with commingled groundwater contaminant plumes. Environmental Forensics 1(1), 47±54. Mink, F.L., Nash, D.E. and Coleman, J.C. II 1997. Superfund site contamination: Apportionment of liability. Natural Resources and Environment 12, 68±79. Murphy, B.L. 1996. Risk assessment as a liability allocation tool. Env. Claims Jour 8, 129±144. Rockwood, L.L. and Harrison, J.L. 1993. The Alcan decisions: Causation through the back door. Environmental Law Reporter 25, ELR 10542. Wise, K.T., Maniatis, M.A. and Koch, G.S. 1997. Allocating CERCLA liabilities: The applications and limitations of economics. BNA Toxics Law Reporter 11, 830±833.