The Case of the California LNG Texminal

Working Paper A D E S C R I MODEL ~ OF CHOICE FOR SITING FACILITIES: THE CASE OF THE CALIFORNIA LNG TEXMINAL Howard Kunreuther John Lathrop Joanne Linnerooth J u l y 1981 WP-81-106 International Institute for Applied Systems Analysis A-2361 Laxenburg, Austria NOT FOR QUOTATION WITHOUT PERMISSION OF THE AUTHOR A DESCRIPTIVE MODEL OF CHOICE M)R SITING FACILITIES: THE CASE OF THE CALIFORNIA LNG TERBdINAL Howard Kunreuther John Lathrop Joanne Linnerooth J u l y 1981 WP-81-106 Working Papers are interim reports on work of the International Institute for Applied Systems Analysis and have received only limited review. Views or opinions expressed herein do not necessarily represent those of the Institute or of its National Member Organizations. INTERNATIONAL IYSTITUTE FOR APPLIED SYSTEMS ANALYSIS 2361 Laxenburg, Austria ABSTRACT T b s paper develops a descriptive framework to provide a basis for prescriptive considerations for improving societal decision processes, such as that concerned with the siting of energy facilities that offer longrun benefits but a t the risk of catastrophic consequences. The descriptive model incorporates multiple parties and multiple concerns in a dynamic setting. It emphasizes the potential for conflict emerging among t h e interested parties as a result of their differing objectives, mandates, and information sources. Of particular importance is the decentralized and sequential nature of the process. On the prescriptive side, the paper explores the role that analysis, including risk analysis assumption analysis, and multi-attribute utility analysis, can play in improving the decision process. The descriptive model and prescriptive considerations are d u s t r a t e d by a concrete example-the siting of a liquefied natural gas terminal in California. CONTENTS I. INTRODUCTION 11. A D E S C R I P T I V E MODEL O F C H O I C E A. B. Basic Concepts 1. Many Interested P a r t i e s 2. Sequential D e c i s i o n P r o c e s s 3. Differences in Information Utilized A Formal Structure 1. Rounds 2. Problem Bounds 3. Initiating Events 4. A l t e r n a t i v e s , A t t r i b u t e s and D e c i s i o n s 5. Interaction P h a s e 6. Concluding a Round 111. A P P L Y I N G FIAMP: TERMINAL A. B. C. D. IV. N a t u r e of t h e P r o b l e m I n t e r e s t e d P a r t i e s and R e l e v a n t A t t r i b u t e s The Decision Process R e s u l t s and I n t e r p r e t a t i o n of t h e MAMP D e c i s i o n SUGGESTIONS FOR P R E S C R I P T I V E MEASURES A. B. V. THE S I T I N G O F THE C A L I F O R N I A LNG A n a l y s i s t o Overcome Sequential D e c i s i o n Shortcomings A n a l y s i s t o C l a r i f y P r o b l e m B o u n d s and P a r t y Positions SUMMARY AND FUTURE RESEARCH REFERENCES A DESCRIPTIYE MODEL OF CHOICE FOR SITING FACILITIES: THE CASE OF THE CALIF'ORNLA LNG TERMINAL Howard Kunreuther, John Lathrop and Joanne Linnerooth* I. INTRODUCTION** Large-scale, novel technologies such as nuclear power or liquefied energy gas promise to yield benefits to society, but only at the cost of potential catastrophic losses. Thus the siting of the facilities for these technologies presents a formidable challenge to political risk management processes. There are two features of these problems which make them particularly difficult to structure analytically. First, the decision on whether to site a facility in a particular location affects many different individuals and groups in society rather than being confined to the nor- 'Authors' names are in alphabetical order. **The resevch reported in this paper is supported by the Bundesministerium fuer Forschung and Technologic, F.R.G.,contract no. 321/75QI/RGB8001. While support for this work is g r a t e m y acknowledged, the views expressed are the author's own and are not n e c e s sary shared by the Tonsor. We greatly benefited from earlier discussions with Nino Majone and wish t o thank Eric Ferguson, Paul Kleindorfer. Chris Mandl and Michael Stoto for their helpful comments on a preliminary draft of this paper. ma1 relationship of a private market transaction, when one purchases items like food, appliances or a car. If a gas terminal is sited, there is a class of firms and consumers who stand to gain from the new energy source. Among other possible social costs, the individuals living near a proposed site will be exposed to a new hazard. Their views as to whether the benefits are worth the extra risks from the new technology a r e likely to differ from each other. Those who a r e not directly affected by the proposed site will see the problem somewhat differently. Hence there is the potential f o r conflict between interested parties. A second feature of the siting problem is the absence of a data base which provides conclusive statistical evidence on the likely performance of the new technology and the probability distribution associated with potential accidents. Each of t h e interested parties may thus provide different estimates of the chances and consequences of certain events. There a r e n o objective measures to settle these differences. This paper proposes a descriptive model of societal choice for characterizing problems such as the siting of energy facilities where the above two features are present. We illustrate the model in the context of a particular case study-the siting of liquefied natural gas terminals in California. The final portion of the paper suggests how this framework for descriptive analysis might provide insights to aid political decision processes. 11. A DESCRIPTlTE MODEL OF CHOICE A. BASIC CONCEPTS 1. Many Interested Parties An important feature of the societal decision-making problem represented by the siting of a facility is the presence of a number of different interested parties, each with its own objectives, data base and constraints (Keeney 1980). For instance, locating a n energy facility in a particular region will be perceived differently by each of the following relevant parties: G o v e r n m e n t Agencies Government organizations a r e constrained in their actions by legislative mandates. For example, in the siting of LNG facilities, the Department of Energy must determine whether t h s new source of energy is in t h e national interest. Regulatory agencies a r e an integral part of the political process. Their actions influence the nature and distribution of t h e public's preferences and provide advantages to some interests relative to others in their efforts to affect the outcome of the societal decision process (Jackson and Kunreuther 1981). These organizations a r e qualitatively different from those in the following three categories, in t h a t government agencies, although they may in some cases have a n interest in a particular outcome, are basically the rule makers and "referees" of t h e process, while t h e other parties are the "players" with costs and benefits directly linked to the outcome of the process. Industry Firms whlch a r e involved in the construction and operation of a n LNG plant will focus on the potential economic benefits to them in relation to t h e cost. In addition to the cost of building and operating t h e facility, safety features of the plant are also a concern since t h e firm may be partially or fully responsible for the consequences of any accidents to the public. Local Community Residents in the community where the plant is to be sited may have different views of the situation. Some might see the economic benefits, e.g., reduced property taxes or employment, as justifying the increased risk associated with t h e plant. Others may feel differently. For instance, they may view the additional risk as unacceptable a n d / o r they are opposed to new technologies on ideological grounds (Otway and von Winterfeldt). P u b l i c Interest Ciroups Recently we have seen t h e rise of very intense public interest groups. These organizations generally represent the interests and preferences of one component of the public. For example, the membership of the Sierra Club is concerned with the effects that the siting of any new facility will have on the environment. Wilson (i975) and Mitchell (1979) have pointed out t h a t those a t t r a c t e d to such organizations have strong, particular interests w h c h dictate t h e agenda of the organization and influence the type of information t h a t is collected and processed. 2. Sequential Decision Process A second feature of the siting problem is t h a t the process is characterized by sequential decisions. March (1969) notes that individuals and groups simplify a large problem into smaller subproblems because of the difficulty they have in assimilating all alternatives and information. Often constraints due to legislation and legal considerations dictate the order in w h c h certain actions must be taken. In the case of siting LPJG terminals in California, for example, the California Coastal Commission was mandated to rank a s e t of alternative sites before t h e California Public Utilities Commission could select a site for approval. This type of quasiresolution of conflict through formal decentralized procedures is very common in organizations where departments with different goals and constraints attend to different decisions a t di'yerent times. Cyert and March (1963) provide empirical evidence suggesting t h a t the organization factors its decision problems into subproblems e a c h of w h c h a r e assigned to different units. If t h e process is sequential in nature then the setting of a n agenda is likely to play a role in determining the final outcome a s well as the length of time i t takes to reach it. By agenda setting we a r e referring t o the order in w b c h different subproblems a r e considered. There is strong empirical evidence from t h e field as well as from laboratory experiments (see Downs i973; Cobb and Elder i975; and Levine and Plott 1977) t h a t different agendas for the s a m e problem frequently lead t o different outcomes. There a r e two principle reasons for t h s . Once a particular decision has been made on a subproblem this serves as a constraint for the next subproblem. If the order of the subproblems is reversed then there would likely be a different s e t of choices to consider. Secondly, each subproblem involves a different s e t of interested parties who bring with them their own s e t of data to bolster their cause. The timing of the release of this information may have a n effect on later actions. For example, citizens groups normally e n t e r the scene with respect to siting problems only when their own community is being considered as a possible candid a t e . The d a t a on t h e risks associated with siting would be released a t a slower r a t e (but perhaps with g r e a t e r emphasis and more political impact) if only one site was considerd a t a time t h a n if all potential sites were evaluated simultaneously. 3. Differences in Information Utilized Finally, t h e r e is considerable evidence from field and laboratory experiments suggesting t h a t the small data base for judging outcomes and performance will cause systematic differences in the way information is utilized a t different stages of the decision process. Tversky and Kahneman (1974) describe one of these phenomenon under t h e heading of availability whereby one judges the frequency of a n event by the ease with which one can retrieve it from memory. Special reports on a recent disaster call attention to the dangers associated with a particular technology. These events may drastically change the estimates of specific outcomes and are also likely to stimulate efforts to induce new legislation to "prevent" the event from reoccurring (Walker 1977). Fischhoff, Lichtenstein and Slovic (1979) summarize their recent experimental studies by cataloguing the perceptions of individuals on the perceived probability of occurrence and consequences of different types of hazards. One of their principal conclusions is that these estimates tend to be labile and hence a r e likely to change over time because of events w h c h a r e highlighted by mass media coverage. T h s finding reinforces the importance of focusing on changing perceptions of interested parties over time and t h e need to study the dynamics of societal decision-making. B. A FORMAL STRUCTURE The above concepts are now incorporated into a model of a sequential decision process w h c h involves different interested parties a t each stage. The approach focuses on more than one attribute and many interested parties. Hence we have labeled it the Multi-Attribute, MultiParty (M.4VP) approach. In developing this structure we have been greatly influenced by the concepts discussed by Braybrooke (1974) where he looks upon t h e political s y s t e m "as a machine o r collection of machines for processing issues." (p.1). Since we view the siting problem a s a decision with multiple objectives, the notation we utilize below builds on t h a t of Keeney a n d Raiffa (1976) who a r e primarily concerned with a single decision-maker. Our i n t e r e s t h e r e is in extending t h e s e concepts t o t h e case where t h e r e a r e many interested parties who i n t e r a c t over a finite horizon. 1. Rounds The decision process c a n be s e p a r a t e d into different rounds w h c h we label by capital l e t t e r s , A,B,... A round is simply a convenient device t o illustrate a change i n t h e focus of discussions. This new focus o r direction c a n be triggered by ( i ) a k e y decision t a k e n ( o r a s t a l e m a t e reached d u e t o conflicts among p a r t i e s ) , o r (2) a change in t h e context of t h e discussions due to a n unanticipated event, t h e e n t r a n c e of a new p a r t y o r new evidence t o t h e debate. Rounds a r e simply a convenient way of segmenting t h e decision process; t h e r e is no assumption, however, t h a t they cannot be simultaneous o r overlapping. 2. Problem Bounds The decision process in e a c h round is bounded by a s e t of issues, decision constraints and procedures. Braybrooke (1974) refers t o a n "issue-circumscribing phase where t h e alternatives for discussion a r e bounded by generally a c c e p t e d though not necessarily irrefutabble, f a c t s and values, e.g., "it is technically feasible to i m p o r t LNG t o California" o r "California needs LNG". Other bounding constraints include legislative and legal mandates requiring specific parties to be p a r t of the debate; resource constraints whch have the effect of limiting certain parties from exerting an influence because they do not have adequate funds and means; and prespecified voting procedures indicating what parties have the power to influence the outcome of specific decisions and in what ways. As we have already noted the problem is also bounded by the decisions from earlier rounds. Clearly the agenda setting process will have an impact on final outcomes through t h s latter type of bounding constraints. 3. Initiating Events Generally, a round of more-or-less official discussions is initiated by a formal or informal request. Informal discussions may be initiated simply by such actions as a request for information on the part of one of the parties or a request for preliminary discussions. Because the particular form of these initiating requests may further define or limit the bounds of the discussion, the careful scrutiny of their wording is important. For example, it may make a difference in the decision process if the question is framed as "Is there a site w h c h is appropriate?", or "wbch of the sites x, y, and z is the appropriate site?" Braybrooke (1974)refers to the first question as a "whether question" and the second as a "which question. Whether questions demand more complicated considerations and detailed thinking whle whch-questions can be approached with simpler rules of thumb and heuristics. 4. Alternatives, Attributes, and Decisions No m a t t e r how a round is initiated i t is characterized by a unique problem formulation w h c h is presented in the form of a set of alternatives. We define the alternatives for Round A to be A 1 , A 2 , A 3 ,. . ; Round , B has alternatives B ' , B ~ , B . . ~, . There can be several decisions made in any round but by definition they a r e based on the same set of alternatives. Each alternative is characterized by a s e t of attributes X 1 . . . . . X n . The value of any attribute c a n change from round t o round on the basis of new information or perceptial changes. For certain attributes any party involved may have target or aspiration levels which determine whether he considers a particular alternative in Round A to be acceptable with respect t o attribute Xi.An aspiration level frequently used in siting decisions is whether the probability of a catastrophic accident a t a particular site is below a present acceptable level (1980). 5. Interaction Phase To understand a particular p a t t e r n of institutional choice it is necessary to analyze a s e t of policy actors !PI their interactions with one other a t different stages of the process, and the information available to them. We define Pk to be the kth interested party in t h e debate. His evaluation of alternative A' is given by and is based on his estimation of the levels of each attribute resulting from that option, the value to h m of each of those levels, and the relative importance he gives to each attribute. Another party might have different estimates of the effects of a n option, different costs and benefits resulting from those effects, or, assign different relative importance to e a c h of the attributes. Because of any of these dGferences one party may rank alternatives differently than another. As we shall see in our analysis of the California case t h s happened frequently. Thus in the case of two interested parties and two alternatives it is possible t h a t party 1 prefers A ] to A~ (i.e., A! > A ; ) , while party 2 has the reverse reaction (i.e.,A: > A ; ) . Conflict frequently emerges for this reason. Another important feature of the decision process is that t h e value of a n attribute to the same interested party can change over time because of new information. For example, if a report provides new insight into the seismic risk associated with a particular site this may cause a change in t h e perception of t b s attribute by one or more of the parties involved. That change may take the form of a different estimate of t h e level of the attribute for that site, or even a different relative importance for the attribute. The interaction among t h e parties is represented by the m a i n arguments each brings to the debate in support of or in rejection of each of t h e alternatives a t hand. Those arguments may relate to only one or two attributes. It is not suggested here that the arguments presented for or against a particular proposal necessarily reflect a concern of the party making the argument. For example, a party opposed to a site because of its concern for environmental quality may present a n argument using seismic risk as the main reason to reject the site. The argument attribute may be selected to maximize the effectiveness of the argument, not to reflect the actual concern of t h e party. The argument reflects a s t r a tegy on the p a r t of the actor in support of or opposition to the proposal. Arguments or positions are generally well thought out since it is import a n t that a policy maker take a position that cannot be shown to be inconsistent with past and ongoing policies o r decisions. The strategy of the actors can reveal a number of underlying motives and desires of those concerned and may be essential in understanding t h e interpretation and use of scientific evidence, including risk analyses. Observing the interaction phase over the various rounds c a n provide useful insights into the process. The stability of the system can, a t least partially, be judged by the degree to which the actors--people holding certain recognized positions (i.e., officials experts, group leaders) or collections of these people, whether formally organized institutions or loosely working alliances-remain the same over each successive round. The interaction phase c a n also demonstrate the extent to which t h e political debate approximates an ideal or logically full debate, where each argument is confronted--heeded or rebuffed--by e a c h a c t o r , , followed by counter arguments, and so on. A social debate on a n issue is merely symbolic if the actors respond to the arguments presented only as signals for asserting or reasserting their own unmodified views. In thls connection, the MAMP model can indicate to what extent the debate falls short of being a full debate, and for what reasons. 6. Concluding a Round The round is concluded by a decision, a stalemate, a char ge in information (changing the focus of the debate and hence initiating a new round), or an unanticipated event aborting the discussions and requiring a new round of inquiry. Each decision can, in turn, be described by t h e tradeoffs implicit in t h e choice made. These tradeoffs may not be explicitly recognized by the decision maker, or not explicitly analyzed in t h e process making the decision. The distinction between a "decision-maker" and a decision resulting from a process is an important one since t h e person responsible for the decision often cannot be identified (see Allison 1 9 7 1 and Majone 1979). Figure 1 provides a schematic diagram of the MAMP model. Each round is s e t in the context of certain constraints that circumscribe t h e issues and bound the problem. An initiating event determines a n initial s e t of alternatives l ~ ~ ~ ~w ~h c' h. , 'in] turn, induce a s e t of interested par- ties i P k j to e n t e r the scene. Each of these parties has its own preferences over the above alternatives (and perhaps some new ones); these preferences are defended by reference to the attributes. Conflict between parties is likely to emerge.The interaction process results either in a clear decision or a n outcome t h a t does not have t h e appearance of a decision but t h a t does conclude t h e round. The conclusion of round J can take one of two forms. If there is a feasible and agreed-upon solution or if no solution is possible, the process ends. However, if one or more parties is unsatisfied with the situation a t t h e end of the round, and has recourse to other channels, or if the round ends in a request for further action, problem bounds are created for Round J + 1 and t h e above sequence is repeated for another set of alternatives, interested parties (some or all may be the same as in J), etc. Standard Procedures Problem Bounds 1 1 Initiating Event Identification of A l t e r n a t i v e s - - -- - I Impacts on 1- + I I n p a c t s on I Preferences Preferences I tsrriatives Defe?.l ?reference I w + t Decisicn Taken c r "Non-Decision" Outcome I I 1 no Is Process Closed? Fiquze 1: H u l t i a t t r i S u t e , Dynamic, M u l t i p a r t y ( b4x4P , ) I !?ode1 of Choice A. NATURE OF THE PROBLEM hquefied natural gas (LNG) is a potential source of energy w h c h requires a fairly complicated technological process t h a t has the potential, albeit with very low probability, of creating severe losses. For purposes of transporting, natural gas c a n be converted to liquid form a t about 1/600 its gaseous volume. It is s h p p e d in specially constructed tankers and received a t a terminal where it undergoes regasification and is then distributed. The entire system (i.e., the liquefaction facility, t h e LNG tanke r s , t h e receiving terminal and regasification facility) can cost more than 31 billion to construct (Office of Technology Assessment 1977). In 1974, three LNG terminals were proposed for California. After seven years of negotiations, hearings and studies, on three levels of government, t h e r e is still no approved site for any of the proposed terminals in California. In this paper, we cannot begin to cover, much less assess, the many and varied aspects of this procedure, (for a review, see Linnerooth 1980 and Lathrop 1981). Our purpose here is to give a "bare bones" account, to illustrate how the MAMP model can be used to structure and understand the decision process. B. INTERESTED PARTIES AND RELEVANT ATTRIBUTES To structure the siting process we need to have a good understanding of the different concerns of the interested parties. For the LNG problem there were three categories of concern which are relevant: risk aspects, economic aspects, and environmental aspects. Each of these concerns can be described by a s e t of attributes. Table 1 depicts an interested party/attribute matrix showing the main concerns of each of the relevant groups over this seven year period. The attributes listed have been selected to reflect the nature of debates in the process, that is, to reflect the attributes as perceived by the parties in the debate, rather than to characterize in some logical analytical manner the alternatives. For example, population risk (Xz) involves the risk to life and limb to neighbors of the LNG terminal due to accidents including earthquake-induced accidents. Earthquake risk (X3), which involves both population risk and su?ply interruption risk due to earthquakes, is included as a separate attribute since it was handled as such in the process. The filled cells in Table 1 indicate which parties pay particular attention to whlch attributes. Naturally, many of the parties care about all the attributes listed. However, either because of the incentives directly felt by the party or because of the role the party plays in society, each party makes its decisions as a function primarily of a particular subset of the X x X n X i X - X 3 X attributes. For example, w h l e the applicant is certainly concerned with environmental quality and risks to the population, its primary responsibilities and concerns a r e earning profits for shareholders and delivering gas reliably to consumers. Its actions a r e apt to be motivated by concerns for profits and gas supplies, and constrained by political and legal limits s e t by other parties' concerns for safety and the environment. Likewise, the Sierra Club cares about reliable gas supplies, but receives membership dues for being primarily concerned with environmental qual- ity. Consequently, in a situation where a proposed action increases the reliability of gas supply a t the expense of environmental quality it is reasonable for the applicant to favor the proposal and the Sierra Club to oppose it. These differences in primary concerns may determine a great deal of t h e behavior of t h e political decision process, and explain how that process is apt to differ from the single decision maker postulated by normative evaluation approaches. The important message of Table 1 does not lie i n the details of exactly which cells a r e filled, but lies in the generally great differences between columns of the table. That is, the different parties in the process c a r e about very different subsets of the attributes. The applicant, Western LNG Terminal Associates, was a special company s e t up to represent t h e LNG siting interests of three gas distribution utilities: Southern California Gas Company, Pacific Gas and Electric, and El Paso Natural Gas Company. As domestic gas supplies seemed to be diminishing in the late 19608s,the gas utilities perceived a n increased risk of supply interruption, which could be mitigated by additional supplies such as LNG. Quite naturally, the applicant was primarily concerned with profitability (xs) and secure supplies of gas (XI). A t the various government levels there are five principal parties. The Federal Energy Regulatory Commission (FERC) is the principal body in the Department of Energy w h c h determines whether a proposed LNG project is in the public interest and should be allowed. In making its judgment it considers primarily the following attributes: risk factors environmental guidelines as reflected in air quality (X4) (XI ,Xz , and X3), and use of land (X5), and the expected LNG price (X7). A t the s t a t e level, t h e California Coastal Commission (CCC) was created in 1976, and has the responsibility for the protection of the California Coastline. Its primary concerns with respect to LNG siting a r e with the use of land (X5) and the associated risks (X2 and X3)from building a terminal a t a specific site. The California Public utilities Commission (CPUC) is the principal s t a t e body involved in power plant issues and is primarily concerned with t h e rate-setting process. Hence it focused on t h e provision of energy to California residents and need for gas ( x ~ )and t h e proposed price of t h e product (X7). In addition, it has responsibility for evaluating the impact that a proposed facility would have on t h e environment and safety. The California s t a t e legislature is ultimately responsible for the outcome of any siting process. It determines w h c h s t a t e and local agencies have final authority to rule on t h e feasibility of a proposed site. In addition, it c a n s e t standards to constrain any siting process. Hence the concerns of t h e legislators range over economic, environmental and safety attributes as shown in Table 1. A t the local level, the city councils evaluate the benefits of a proposed terminal in their jurisdiction in t e r m s of the tax, business revenues, and jobs (Xs)it promises to provide. I t has to balance this positive feature with the impact t h a t the facility would have on land use (X5)and risk to t h e population (X2). Finally, t h e public interest groups, represented by t h e Sierra Club and local citizens groups, are primarily concerned with environmental and safety i,.sues. C . THE DECISION PROCESS The sitirg process in California ( w h c h is not yet terminated) can be characterized up to now by four rounds of discussions as shown in Table 2, w h c h provides a summary of the entire process. Each round, in t u r n , containa a summary of how t h e problem was defined, the initiating event, and how the discussions were concluded. The remainder of t h s subsection discusses in more detail t h e decision process w i t h n each of t h e rounds. The main elements of rounds A, B, C, and D are described in Tables 3, 4, 5 and 6, respectively. Round A began in September 1974, when the applicant filed for approval of three sites on the California Coast--Point Conception, Oxnard, and Los h g e l e s - - t o receive gas from Indonesia. The application raised two central questions w h c h defined the problem addressed in Round A: Does California need LNG, and if so, which, if any, of the proposed sites is appropriate? -21 - TaSle 2 : Summary o f Rounds i n C a l i f o r n i a LNG S i t i n g Case Date - RCUND A Problem D e f i n i t i o n : I n i t i a t i n g Event: Conclusion: Problem D e f i n i t i o n : I n i t i a t i n g Event: Conclusion: Should t h e proposed s i t e s b e approved? T h a t is: Does C a l i f o r n i a need LNG, and i f s o , which, i f any, c f tlle proposed s i t e s is appropriate? Applicant f i l e s f o r approval o f t h r e e s i t e s . A p p l i c a n t p e r c e i v e s t h a t no s i t e i s approvable without long delay S e p t e n b e r 1974 ( 3 4 months) Jluly 1977 How s h o u l d need f o r LNG be determined? I f need i s e s t a b l i s h e d , how s h o u l d an LVG f a c i l i t y be s i t e d ? A p p l i c a n t and o t h e r s p u t p r e s s u r e on s t a t e l e q i s l a t u r e t o f a c i l i t a t e LNG s i t i n g . J u l y 1977 Mew s i t i n g p r o c e s s s e t up t h a t e s s e n t i a l l y S e p t e s b e r 1977 (2 mnths) assumes a need f o r L V G , and is d e s i g n e d t o a c c e l e r a t e LNG t e r m i n a l s i t i n g . lrcblem Definition: I n i t i a t i n g Eve3t: Conclusion: Which s i t e s h o u l d b e approvee? Asplicant f i l e s for approval c f Point Conception s i t e . October 1977 S i t e approved c o n d i t i o n a l on c o n s i d e r a t i o n of additional seismic r i s k data. J u l y 1978 ROUND D Problem D e f i n i t i o n : I n i t i a t i n g Event: Conclusion: Is P o i n t Conception s e i s m i c a l l y s a f e ? Regulatory a g e n c i e s s e t up p r o c e d u r e s t o consider additional seismic r i s k data. (Xound s t i l l i n p r o g r e s s ) (10 sonths Zlements o f Round A Table 3 : Problem D e f i n i t i o n : I n i t i a t i n g Event: Alternatives: Should t h e proposed s i t e s b e approved? T h a t i s : i h e s C a l i f o r n i a need LNG, and i f s o , which, i f any, o f t h e proposed s i t e s is a p g r o p r i a t e ? Applicant f i l e s f o r approval o f t h r e e s i t e s . Site Site Site Site at at at at P o i n t Conception: Oxnard: Los Angeles: any combination o f : A; A1 A , A 2 , A 3 Interaction : Involved P a r t i e s Applicant A t t r i b u t e s Used a s Arqcments 1 X1 FERC CCC C i t y Councils S i e r r a CluS Local C i t i z e n s P6 X2 X5 P7 8 Key Decisions: 1 . 2. 2 3 based on t k e fact that the could be preferrsd to %, and A3, d e c m s e in population r i s ~ outreighs t k increzse in ewirammtal degradation. would be qpruved &athe seismic risk is q=t b n an accatable t h r e ~ ~ b l d . 1 4 Conclusion: Applicant perceives a stalemate, i . e . , without long delay. t h a t no s i t e i s a p p r o v a l e Problem D e f i n i t i o n : I n i t i a t i n g Event: Alternatives: How s h o u l d need f o r LUG be determined? I f need i s e s t a b l i s h e d , now s h c u l d an LNG f a c i l i t y be s i t e d ? A p p l i c a n t and o t h e r s p u t p r e s s u r e on s t a t e l e g i s l a t u r e t o f a c i l i t a t e LNG s i t i n g . 1 Consider o f f s h o r e s i t e s : 2 C o n s i d e r remote o n s h o r e s i t e s : 3 Consider non-remote o n s h o r e s i t e s : 4 One-stop l i c e n s i n g : 5 6 L i c e n s i n g Agency: CPUC = B , CFC = B , CEE = a'* Any c o n s i s t e n t combination o f B through a ' . Interaction: A t t r i S u c e s Used f o r Arguments Invoived P a r t i e s Applicant CCC State Legislature X2 3 P 5 1 2 Key D e c i s i o n s : 3. I n i t F 4 1 l e q i s l a t i o n i n t r ~ d u c e dv h i c h Lnciu5ed a 4. F i n a l l e g i s l a t i c n p a s s e d which i n c c r p o r a t e d 1 , a 2 , a4 7 3- and B and B 5 5 . . Csnclus i o n : Passage cf LNG S i t i n g Act o f 1977 (S.a.1081) which d e f i n e s a c u s c o ~ t- a i l o r e d s i t i n g p r o c e d u r e f o r LNG. Some f e a t u r e s : CCC nominates and r a n k s s i t e s i n a d d i t i o n t o t h e one a p c l i e d f o r . CPUC s e l e c t s a s i t e from the CCC-ranked s e t , n o t n e c e s s a r i l y t h e top-ranked s i t e . --- Table 5 : E l e ~ e n t so f Round C Prcblem Definition: I n i t i a t i n g Event: dternatives: Which s i t e s h o u l d b e approved? A p p l i c a n t f i l e s f o r a p p r o v a l o f P o i n t Conception s i t e . (The o n l y s i t e o f i t s o r i g i n a l t..ree meeting t h e remote s i t i n g c o n s t r a i n t o f S.B.1081.) ( S i t e s nominated by CCC p l u s a p p l i e d - f o r s i t e . S i t e a t Camp P e n d l e t o n : S i t e a t R a t t l e s n a k e Canyon: 3 S i t e a t P o i n t Conception: S i t e a t Deer Canyon: C ) Interacticn: A t t r i b u t e s Used f o r Armments Involved P a r t i e s Applicant ERC CCC CP UC S i e r r a Club Local C i t i z e n s 7 'a Ksy D e c i s i o n s : 6. < approved conditional on further determination whether or not seismic risk a t Point CoM=eption is acceptable. 2 7. C3 considered acceptable. 8. Court requires F E Z to ansirier additicnal da'd t o Z e t m n b e whether o r not seismic risk a t Point Conception is acceptable. Conclusion : Fnc and BUC t o ccnsin'Py additional seismic data. Table 6 : Elements o f Round D Problem D e f i n i t i o n : I n i t i a t i n g Event: Ute-natives : Is P o i n t Concegtion s e i s m i c a l l y s a f e ? F'ERC and CTUC s e t up p r o c e d u r e s t o c o n s i d e r additional seismic r i s k data. D e c l a r e P o i n t Conception s a f e : Declare P o i n t Conception n o t s a f e : D Interaction: No i r i t e r a c t i o n y e t , a s s t u d y grocps f o r FERC and QUC seismic data i n preparation f o r hearings. C u r r e n t l y Accive P a r t i e s examine Attribute Considered Key D e c i s i o n s : None y e t . F u t u r e n e a r i n g s a r e t o d e t e r m i n e whether o r n o t s2ismic r i s k . i s a c ~ g - l e for Point Conception. Although the issue was a t t h s stage broadly circumscribed, the agenda f o r discussion was s e t more narrowly. The wheels of the process were set into motion, not by a broadbased energy-policy question initiated in Washngton, but by a proposal from industry for three preselected sites. The importance of t h s process--where the initiative is taken first by industry--in preselecting the agenda for debate cannot be overemphasized. The initiating proposal framed the problem as "Should the proposed LNG sites be approved?", and not "Should California have a n LNG terminal in view of the alternatives, costs, risks, etc.?" Setting the agenda in this manner did not preclude the "need" question from entermg the debate, but it did ensure t h a t the question was only considered in the contentious context of a siting application. Table 3 also specifies the relevant interested parties who were involved in the interaction phase of Round A. There wore four primary attributes which were utilized in the ensuing debate among the parties. The need for LNG or the risk of a n interruption in the supply of natural gas (XI) supported the locating of a terminal in a t least one of the three proposed sites. While environmental, land-use considerations (X3) suggested a non-remote site (Los Angeles and Oxnard), the risks to the population (Xz) argued for siting the terminal in a remote a r e a (Point Conception). Finally, concerns about earthquake risk brought about opposition to the Los h g e l e s site, which was found to be crossed by a significant fault. The interaction phase of round A (see Table 3) indicates the attributes used as arguments by e a c h of the major involved parties. It is important to distinguish t h s listing of attributes from the listing in Table 1. Whle Table 1 specifies w h c h attributes are of p r i m a r y concern to each party, Table 3 specifies which attributes were used as a r g u m e n t s by each party. Thus while the applicant is concerned with both profit considerations and supply interruption risk, naturally enough its arguments in support of each site stressed supply interruption risk. Two key decisions were made during Round A. First, the CCC, concerned about the catastrophic potential of LNG, indicated t h a t it might favor Point Conception over t h e non-remote sites. T h s decision represents a trading off of population risk against land-use environmental quality; in the words of the CCC, this was a "people over birds" decision. ( T h s decision did not, however, tie t h e CCC to approval of Point Conception. In fact such an approval could be difficult to obtain.) Second, the FERC indicated disapproval of the Port of Los Angeles as an acceptable site because a recently discovered earthquake fault increased the seismic risk above a n acceptable threshold. The round was concluded with a possible stalemate, a t least as perceived by industry (Ahern 1980). Los Angeles would not receive federal (FERC) approval, Oxnard was not likely to receive state (CCC) approval, and Point Conception would face very difficult approval challenges a t the county and s t a t e (CCC) levels because of its adverse land-use impacts. The stalemate of Round A bounded the problem for Round B. I t was clear to all t h e parties involved t h a t it was difficult, if not impossible, for t h e applicant to gain approval for a site under the existing siting procedure in California. In particular, t h e r e were possibilities of vetoing proposals a t either the federal, s t a t e , or local levels as evidenced by the respective reactions to the t h r e e proposed sites. Rather than trying to operate withn the existing constraints of the process, t h e interested parties in the process frequently try to change the rules of the game (Majone 1979). T h s is precisely what triggered Round B, where the problem was redefined into two new questions: How should need for LNG be determined? If need is established, how should an LNG facility be sited? Round B was thus initiated when pressure to change the siting procedure was brought to the s t a t e legislature by the utility companies, t h e business community and the labor unions in California. Table 4 depicts t h e relevant alternatives which formed the basis for the debate on the elements of proposed legislation. The industry and business interests, saw the nearly inevitable problem of obtaining local approval for a project in the national interest, but with costs to t h e local community. So the utility companies went to battle for a bill (S.iO81)which would vest the CPUC with one-stop licensing authority, precluding any real interference from local communities. The environmental and local interests, on the other hand, objected to a onestop l i c e n s i q process and favored a bill which required remote-siting. The resulting legislation was a compromise between t h e environmentalists, who supported consideration of off-shore sites, and those who saw a n urgent need for an LNG facility t o assure energy and jobs. The CPUC was chosen over the more conservation-minded CCC or the California Energy Commission as the agency with s t a t e permit authority, preempting local governments. As a bow t o the conservationists, t h e CCC was given t h e mandate t o choose and to rank possible sites, and t o pass these rankings on to the CPUC. I t was agreed t h a t t h e site would not be offshore, as some environmentalists wished, nor could it be in a populated a r e a , as the gas utilities wished. Indeed, a nonpopulated area was strictly defined. There could be no more than an average of 10 people per square mile w i t h n one mile of t h e terminal, and no more than 60 people per square mile within four miles of the terminal. The passage of the Siting Act of 1977 (s.1081) opened up a new procedure for finding a n acceptable site and led to Round C with the following problem definition: Which site should be approved? The round was initiated by the CCC w h c h , after c o n s i d e r i q 82 sites meeting the remote-siting constraint, ranked the top four sites, Camp Pendleton, Rattlesnake Canyon, Point Conception, and Deer Canyon, in t h a t o r d e r , on the basis of seismic, soil, wind and wave conditions, rough cost, and coastal resource considerations. (Point Conception was included in the candidate set because S. ;OBI required that the applied-for site be included). These four alternatives form the background for the interaction amGng t h e interested parties in Round C as shown in Table 5. The CCC pased these rankings on to the CPUC which chose, by process of elimination, Point Conception, on the grounds t h a t the two higher-ranked sites would involve unacceptable delay and would cause unacceptable risk to transients (i.e., campers, swimmers, e t c . ) a t the nearby beaches and public parks. The CPIJC, however, could only conditionally approve Point Conception subject to the utility company's ability to show that earthquake faults discovered in the a r e a presented an acceptable risk to the terminal. At the federal level, the FERC staff determined t h a t the risks of both Oxnard and Point Conception were acceptably low, so that Oxnard should be p r e f e r r e d on land-use grounds; however, the FERC, choosing to avoid a federal-state confrontation ruled in favor of Point Conception. It would appear that with both the CPUC and the FERC in favor of Point Conception as a compromise site, the issue would be decided. But the checks-andbalances of t h e U.S. system assures that not total authority is vested in t h e legislative branches. After an appeal by t h e environmental and local interests, the Washington, D.C. Court of Appeals remanded the case back to the FERC on t h e grounds that not all available seismic risk data were considered by the FERC in its ruling. T h s decision concluded Round C. Round D is still in progress a t this time. As shown in Table 6 the initiating proposal is bounded by t h e activities in Round C which frame t h e alternatives as simply whether or not to declare the Point Conception site seismically safe. Only two parties, the FERC and the CPUC are currently active in the process, and they are considering only one attribute, --the seismic risk a t Point Conception. A final decision will depend upon whether the new studies show t h s risk to be above or below some acceptable level. D. RESULTS AND INTERPRETATION OF THE MAMP DESCRIPTlON The purpose of the MAMP representation is not simply to describe a political decision process, but to bring a structure to t h a t description t h a t might suggest institutional reforms. The California decision process illustrated in t h s paper can be interpreted from many varied perspectives. It is a good example of conflicting national and local interests; it is a study of the workings of the adversial nature of U.S. regulatory proceedings; it is a precedent-setting report of procedural practice for setting energy policy; as well as a n account of introducing a controversial large- scale technology with a small probability of a catastrophic accident. Clearly, a full exposition of these interpretations would go beyond the more narrow scope of this paper, i.e., to illustrate a framework of descriptive analysis. In this spirit we will briefly present two insights from this approach which a r e of particular concern to the IIASA Risk Group: the role t h a t risk and risk analysis have played in t h s process and the limitations that sequential decision-making places on t h e role of analysis in general. In t h e siting of large-scale, hazardous technologies, a great deal of attention has been paid recently to the topic of risk assessment, a s evidenced in a large and growing Literature (see Conrad 1980 and Schwing and Albers 1980). It is of interest, to examine the role risk assessments, which have been concerned primarily with risk to life and limb, have played in the California LNG case. D u r q the course of the LNG debate in California, six studies assessing t h e safety risks of the proposed terminals were conducted by t h e utility and local, s t a t e and federal government agencies (for a critical review of these studies, see Mandl and Lathrop 1981). The assessments of risk commissioned by the applicant and the FERC showed very low numbers on various probabilistic measures of risk (expected fatalities per year and individual probability of fatality per year), these numbers were interpreted to mean t h a t the risk was acceptable. ,Alternatively, a risk assessment commissioned by t h e Oxnard municipal government produced similarly low probabilistic measures of risk (though expected fatalities were 380 times higher than the applicant's assessment), but interpreted as unacceptably high! The explanation lies in t h e format for presenting t h e results. The Oxnard study described maximum credible accidents (MCAs) without accompanying probabilities. Opposition groups interpreted these results as evidence t h a t the terminal was not acceptably safe. The municipal government, itself, originally in favor of the site, began t o waver in its support, probably influenced by the apparent uncertainty of the risk and the strength of t h e opposition groups (Ahern 1980). In sum, risk assessments did not provide a single, coherent assessment of acceptability of the risk of a n LNG terminal; their results were subject to interpretation depending on party positions (Lathrop 1980). In fact, risk assessments were used both t o promote and to oppose terminal applications. Turning to the importance of sequential decision-making, the most fundamental message of the MAKP approach is t h a t political decision processes such as t h e one studied here tend to be disaggregated over agencies and over time. In t h e California case t h e disaggregation over agencies does not seem to have had a significant effect on the outcome. As indicated in t h e preceding tables, power was in t h e hands of the California Legislature, the CPUC, and FERC, each of w h c h considered most of t h e relevant concerns (see Table 1) and each of which was able to apply its own subjective weights. The important decision affecting the disaggregation over agencies in the process by the diminishment of local sovereignty (and applicant control) was embodied in the California LNG Siting Act (S. 1081). In contrast to t h e disaggregation over parties in the process, the sequential aspect of t h e decision process seems to have been crucial. In t h e seven year course of the process, the need for imported natural gas in California diminished greatly. Instead of examining t h s need, the process, "locked in" by previous decisions, is presently dictating a slowlypaced examination of seismic data. A second example of sequential constraints is the remote onshore condition of S. i 0 8 i , which represented a balancing of costs and benefits t h a t have probably changed over time. A final example of undesirable effects from sequential constraints concerns t h e risk of a n interruption in t h e supply of natural gas. Initially, the applicant stressed supply interruption risk due to shortage of natural gas as a major reason for importing LNG to three separate sites. During the course of the decision process, for reasons beyond the control of the applicant, t h e three sites were reduced to one site, and t h e number of storage tanks a t that site were reduced from four to two. The planned Point Conception throughput of 58,000 m 3 LNGIday, equivalent in energy flow to roughly 15 modern nuclear reactor units (Mandl and Lathrop 198i), is large for one geog r a p h c a l location. Because of t h s concentration in one small area, and the possibility of routine closures or nondelivery resulting from bad weather e t c . , the net result of the sequential decision process is t h a t a project originally meant to d e c r e a s e supply interruption risk has been shaped over time into a project t h a t may i n c r e a s e supply interruption risk. TY. SUGGESTIONS FOR PRESCRIPTIVE MEASURES The use of the MAMP model for structuring the decision process concerning LNG siting in California makes clear how t h e roles, agendas, values and power of the different parties, and the dynamics of their interaction, determine the way in w h c h the process steers its path through a complex multiattribute choice problem. Since a motivation of this descriptive analysis is to provide insights for proposing changes in institutional procedures, and especially for suggesting constructive uses of analyses to aid those procedures, i t is appropriate to suggest prescriptive measures. Of course,institutional change encompasses a range of possible measures from radical reforms in t h e power structures of society to more modest proposals for aiding existing decision structures through analysis in order to facilitate solutions. As we have shown in this paper, the model brings to light problems arismg from the sequential nature of a multi-party decision process that might be remedied by a m o r e imaginative use of analyses. In this spirit we select two problem areas to illustrate directions that analyses could take. A. ANALYSIS TO OVERCOME SEQUENTIALDECISION SHORTCOMINGS A general conclusion to be drawn from the MAMP description is t h a t t h e political decision process examined here solved a complex question one piece a t a time, m u c h as described in the work of Cyert and K .rch (1963) mentioned in Section 11. This disaggregation of problem solution by sequential constraints could be explained in t e r m s of information pro- cessing limitations of the process, t e r m s that correspond to some concepts of bounded rationality (Simon 1957). The examples presented a t t h e end of the previous section illustrate a fundamental problem with this disaggregation: The fact t h a t the political decision process is sequential, with each decision constraining t h e ones that follow, means that the outcome of t h a t process is apt to be inappropriate if important variables change over time. Whle the complexity of siting decisions may make problem disaggregation necessary, analyses can be designed to alleviate the undesirable effects of that disaggregation in a t least two ways. First, techmcal and economic analyses can make clear the implications for the overall process of the particular alternative and constraints being considered a t any one time. For example, the California process might have been helped considerably had the decisions to reduce the number of storage tanks a t Point Conception been accompanied by analyses making clear to all parties the resulting risk of a n interruption in supply. A second way in w h c h analyses might help alleviate disaggregation problems is to demonstrate the point a t which previously-set constraints may no longer be appropriate, and so help identify when previous parts of the process should be re-opened for consideration in the light of new data. While such a role for analyses could be seen as contributing to delays in the siting process, t h a t tendency could be corrected by incorporating delay and cost of deiay explicitly in the analyses. B. ILVALYSIS TO CLARIFY PROBLEM BOUNDS AND P M T I POSITIONS Before one begins t h e process t h e r e m u s t be a clear understanding of t h e problem bounds whlch a r e likely to emerge a t different stages of t h e process. To the extent possible t h e interested parties should b e made aware of t h e different decisions w h c h have to be m a d e , how long these decisions a r e likely t o t a k e a n d the mechanisms utilized to determine t h e final outcome. In other words, some type of normal process should b e established.. In the same spirit it appears appropriate for e a c h of t h e i n t e r e s t e d parties t o specify t h e constraints which guide its decisions and t h e assumptions upon which they are based. Each p a r t y might defend its position by stating the basis for its conclusion regarding site preference. Mitroff, Emshoff and Kilmann (1979) have proposed a technique called assumptional analysis whereby t h e parties undertake a dialectical approach t o t h e problem by being forced to defend t h e assumptions on which t h e y base their position with the opportunity to challenge t h e assumptions of others. Majone (1979) has suggested t h a t t h e knowledge base on w h c h to m a k e decisions for these type of problems is currently so inadequate t h a t s u c h a process would enable one to explore avenues of disagreement and improve his understanding of t h e problem. It is likely, however, t h a t conflicts will continue t o exist between t h e parties because they have different assumptions on which t h e y base their actions. As we have s e e n from the MAMP model, public i n t e r e s t groups like the Sierra Club focused entirely on environmental a n d safety considerations because t h a t is the basis for their existence; public utilities were primarily concerned with economic factors while government agencies were forced by legislative mandate to focus on specific attributes. There a r e two analytic approaches which may be helpful in structuring the multi-party problem so that policy alternatives can be evaluated before attempting to introduce them into the political process. Both approaches incorporate the differing values held and concerns felt by the different parties, and use those to represent and predict e a c h party's ranking of alternatives. The first is multiattribute utility (MAU)analysis, developed by Keeney and Raiffa (1976). In t h a t approach a utility function is developed for e a c h party w h c h is fitted to measures of the relative weights over attributes t h a t party seems to employ in making its choices. These measures can be developed in the course of direct value elicitations of members of e a c h group, or in some cases can be estimated from past decisions or policy stands. A second analytic approach, the analytic hierarchy process (AHP), has been recently developed by Saaty (1980). This approach is conceptually similar to the MAU approach, except t h a t it uses ratios of reIative importance over attributes, and ratios of relative impacts of each of the alternatives, as opposed to the more complete utility functions of the MAU approach. Both approaches r u n into difficulty in trying to represent equity considerations. In any problem of social conflict, such as the siting problem described here, the alternative selected depends crucially on the relative weights given to each party in the process. If full weight would be given to t h e applicant, i t would be free to select a denesly populated site; if full weight were given to neighbors of t h e proposed plant, no siting application would be successful without attractive compensation schemes. Neither analytic approach offers a convenient means to assign weights over parties. In both cases, some individual must assign t h e weights, which begs t h e question of who that individual should be. Keeney (1976) has suggested that there is some referee agency which implicitly assigns these weights, so the MAU (or AHP) process could simply elicit t h e weights from representatives of t h a t agency. In the California case, however, it is not clear whether that agency should be the FERC, t h e CPUC, or another agency. Perhaps t h e best way around t h e equity problem is to use t h e MAU. or AHP approaches with each of several plausible sets of weights over parties. A sensitivity analysis would then shed light on what conclusions could be drawn from a range of weights. Either t h e MAU or the AHP approaches can be combined with instruments of policy analysis, helping the analyst to determine t h e effects of, for instance, incentive schemes and regulations on t h e performance of alternative scenarios. For example, the relevant state o r federal agencies might consider the possibility of compensating those who a r e exposed to t h e risks of a terminal, similarly, additional safeguards might be considered. The MAU or AHP approaches would suggest how changes in t h e values of particular attributes might effect the relative desirability of different scenarios. Of course, other social and political considerations might enter these types of decisions. V. SUMMARY AND FUTURE RESEARCH In t h s paper we have developed a descriptive model of choice, t h e MAMP model, and drawn some presecriptive recommendations where analyses might improve societal decision making with respect to siting problems. The seven year experience in California in trying to site LNG terminals points out the need for a well-articulated set of procedures for dealing with energy problems in the United States. It also suggests t h e need for examining how other countries undertake their siting decisions. A t IIASA we are now completing a cross-cultural, comparative study of siting decisions in four countries: the Federal Republic of Germany (ATZ 1981), the United Kingdom (Macgill !981), the Netherlands (Schwarz 1981), and t h e United States (Lathrop 1981). We will apply t h e MAMP model to each of these cases in the hopes of learning about the similarities and differences among countries and developing a broader s e t of guidelines for improving the siting process based on these descriptive analyses. Problems of siting new technologies will become more import a n t as our energy resources become more scarce. We are hopeful t h a t through international comparative studies we will gain a b e t t e r appreciation as to what can be done in the future. REFERENCES Ahern, W. i980. "California Meets the LXG Terminal." Coastal Zone Management Journal, 7 pp. 185-22i. Allison, G.T. 1971. Essence of D e c i s k n . Boston: Little, Brown and Company. Atz, H. 1981. "Decision-Making in L?JG Terminal Siting: Wilhelrnshaven, F.R.G.", Draft Report, IIASA, Laxenburg, Austria. Braybrooke, D. 1974. T r a f f i c Congestion Goes Through t h e Issue- Machine. London: Routledge and Keg an Paul. Cobb, R . , and C.D. Elder. 1975. Participation in A m e r i c a n Politics: The D y n a m i c s of Agenda Building Baltimore, John Hopkins University Press. Conrad, J. ( e d . ) . i980. Society, Technology a n d Risk A s s e s s m e n t . London: Academic Press. Cyert, R., and J . March. 1963. A B e h a v i o r a l T h e o r y of t h e fim. Engle- wood Cliffs. N.J. Prentice Hall. Downs, J. 1973. "Up and Down the Attention Cycle." P u b l i c I n t e r e s t . Fischhoff, B., S. Lichtenstein, P. Slovic, R. Keeney, and S. Derby. 1979 "Approaches to Acceptable Risk: A Critical Guide." NUREG/CR-1614. Washngton: U.S. Nuclear Regulatory Commission. Jackson, J., and H. Kunreuther. 1981. "Low Probability Events and Determining Acceptable Risk: The Case of Xuclear Regulation." PP-81-7, IIASA, Laxenburg, Austria. May. Keeney, R. 1976. "A Utility Function for Examining Policy Affecting Sal- mon in the Skeena River," RM-76-5, IIASA, Laxenburg, Austria. Keeney, R. 1980. S i t i n g E n e r g y F a c i l i t i e s . New York: Academic Press. Keeney, R., and H. Raiffa. 1976. D e c i s i o n s with rUultiple Objectiues. New York: John Wiley. Kunreuther, H. 1980. "Societal Decision Making for Low Probability Events," WP-80-i64, IIASA, Laxenburg, Austria. Lathrop, J . 1980. "The Role of Risk Assessment in Facility Siting: An Example from California." WP-80-150. IIASA, Laxenburg, Austria. Lathrop, J. 1981. "Decision-Making on LNG Terminal Siting: California, U.S.A.", Draft Report, IXASA, Laxenburg, Austria. Levine, M.E., and C.R. Plott. 1977. "Agenda Influence and 'its Implications." V i r g i n i a Law R e v i e w , 63 (4). Linnerooth, .J. 1980. "A Short History of the California LNG Terminal." WP-80-155, IIASA, Laxenburg, Austria. Macgill, S.M. 1981. "Decision-Making on LNG Terminal Siting: Mossmorran-Braefoot Bay, United Kingdom." Draft Report. IIASA, Laxenburg. .Austria. Majone, G. 1979. "Process and Outcome in Regulatory Decisions." Ameri- c a n B e h a v k r a l Scientist 22 pp. 561-583. Mandl, C., and J. Lathrop. (Forthcoming). "Assessment and Comparison of Liquefied Energy Gas Terminal Risk." WP, IIASX, Laxenburg, Austria. March, J. 1969. "Bounded Rationality, Ambiguity and the Engineering of Choice," Bell Journal of Economics. 9:587-608. Spring. Mitchell, R.C. 1979. "National Environmental Lobbies and the Apparent Illogic of Collective Action." In Applying Public Choice Theory What a r e t h e Prospects. Edited by C. Russell. Washngton, D.C.: Resources for the Future. Mitroff, I., J . Emshoff and R. Kilmann. 1979. "Assumptional Analysis: A Methodology for Strategic Problem Solving ." M a n a g e m e n t S c i e n c e , 6 pp. 583-593. Office of Technology Assessment (OTA). 1977. Trunsportatim'L of Liqz~e- )led N a t u r a l Gus. Washington, D.C.: Office of Technology Assess- ment. Otway, H., and D. von Winterfeldt. (in press). "Judgments of Technologies and their Risks," proceedings of a workshop on risk research conducted by the ZIF, Bielefeld, FRG. November 1980. Saaty, T.L. 1980. The Analytic H i e ~ a r c h yProcess. New York: McGrawHill. Schwarz, M. 1981. "Decision-Xaking on LNG Terminal Siting: The Nether- lands." Draft Report. IIASA, Laxenburg, Austria. Schwing, R.C., and W.A. Albers (eds.). 1980. Societal Risk A s s e s s m e n t : s S a f e Enough? New York: Plenum Press. How S a f e i Simon, H.A. 1957. Models of Man: Social a n d R a t i o n a l . New York: Wiley. Tversky, A., and D. Kahnemann. 1974. "Judgment Under Uncertainty: Heuristics and Biases." S c i e n c e 185:1124-31. Walker, J. 1977. "Setting the Agenda in t h e U.S. Senate: A Theory of Problem Selection." B r i t i s h J o u r n a l of Political S c i e n c e . 7:423-445. Wilson, J.Q. 1975. Political O r g a n i z a t i o n New York: Basic Books.