LaRue Tone Hossmer
Lessons from the Wreck of the Exxon Valdez: The Need for Imagination, Empathy, and Courage
Business Ethics Quarterly, Issue #1 (Ruffin Series) 1998
Investigations of large scale industrial accidents generally take one of two alternative approaches to identifying the cause or causes of those destructive events. The first is legal analysis, which focuses on the mechanical failure or human error that immediately precedes the accident. The second is socio-technical reasoning, which centers on the complexities of the interlocking technological and organizational systems that brought about the accident. Both are retrospective, and provide little insight into the means of avoiding industrial accidents in the future. This article looks specifically at the wreck of the Exxon Valdez. It looks at 6 levels of managerial responsibility within a firm, and suggests specific changes at all levels that should logically help in the prevention or mitigation of these high impact/low probability events. The most basic need, however, is for imagination, empathy, and courage at the most senior level of the firm.
|Copyright Loyola University of Chicago 1998|
Investigations of large scale industrial accidents today generally take one of two alternative approaches to identifying the cause or causes of those accidents. The first is legal analysis. This focuses upon one overriding courtroom concern: the need to prove that a reasonably simple mechanical failure or human error caused the death and destruction. Then, the organization that designed the mechanical system or employed the human being can be shown to be substantially at fault for failure to provide adequate supervision and reasonable care, and consequently should be held fully responsible for compensatory and punitive damages.
The intent of legal analysis is to affix blame and assess retribution for occurrences in the past; it is not -- except indirectly -- to avoid such accidents in the future. The finding is always that something broke or someone erred, and that even a minimally attentive organization should have detected that propensity for failure before it occurred. Statements are often made that the offending business firm has to learn that such lack of care is "totally unacceptable", and that the large damage assessments will now "get their attention", as if these horrendous industrial accidents were somehow similar to dinner table mishaps with a three-year old, and that attentive parents always move the milk back from the edge of the table before it is spilled.
The second major approach to large scale industrial accidents is much more understanding, even somewhat sympathetic, about the causes of those accidents. It focuses on the interconnected nature of technological processes and social institutions, and on the inability of human beings to fully comprehend all of the resultant interrelationships. This is "normal accident" theory.' The belief is that even a slight technical malfunction or human miscalculation in one sector can rapidly escalate through the interdependencies into a disastrous consequence for the total system.
The problem is that neither the legal nor the technical/social approaches to accident analysis really help in the prevention of these events. Both are retrospective in nature -- as, of course, is all empirical research -- but both look at the past through very specific lenses. Legal theory sees human beings as culpable; people who have been distracted by their economic self-interest from what is truly important for their society. Normal theory sees these same human beings as fallible: people who cannot understand what is truly important because of the complexity of the social and political matrix surrounding advanced technologies. Sheila Jasanoff recently summarized these two views with particular elegance and grace:
Accounts of disasters are written most often in a retrospective mode. The tragic event becomes the occasion for looking backward to a prior, more fortunate time, when foresight, prudence, good behavior, or divine grace might have unscrolled history towards a happier conclusion. Disasters involving technology, in particular, have tended through time to take the form of morality plays about human overreaching.2
The legal methodology is remarkably effective in affixing blame and assessing retribution for large scale industrial accidents after they have occurred. The technological/social approach is equally effective in distributing blame and ignoring retribution, also after the fact. Both judgmental methods add little, however, to our understanding of the most critical issue: how senior level managers might act to avoid these low probability/high impact accidents before they occur. Again, Prof. Jasanoff has a few words directly to the point: (We must) break out of the retrospective habits of thought that accidents and mishaps so often engender; to stop asking what caused the tragedy or who is to blame, and to consider instead how human beings and their faultprone institutions can learn to do better.3
The intent of this article is to examine the wreck of the Exxon Valdez in a nonjudgmental mode. The intent is to propose a third form of analysis: to look not at who or what was the direct cause of the accident (legal analysis), nor at the extenuating complexity that made the initial sudden accident and the subsequent slow response almost inevitable (technical/social analysis). Instead, the intent is to consider what might have been done at all layers of management to avoid the accident or speed the response and -- particularly important -- why those critical decisions and actions were not taken. The viewpoint of this article will be managerial avoidance, not legal retribution or "normal" explanation, as applied to the very damaging industrial accident that occurred in Prince William Sound in the spring of 1989.
The events surrounding the wreck of the Exxon Valdez are well known and quickly summarized. At 9:30 P.M. on Thursday, the 22nd of March 1989, the largest tanker owned by the Exxon Shipping Company left the oil terminal at Valdez, Alaska fully loaded with 1.26 million barrels of oil. Initially the ship was under the command of Captain William Murphy, the harbor pilot. Harbor pilots are responsible for steering both incoming and outgoing tankers through the Narrows, a V2 mile wide approach to the port of Valdez. After exiting the Narrows and achieving the sea lanes in Prince William Sound, Captain Murphy turned over command to Captain Joseph Hazelwood and left the ship. Captain Murphy testified later that he smelled alcohol on the breath of Captain Hazelwood, but that he made no comment and took no action. He knew that it was common practice for both the officers and crew of oil tankers to drink while in port. There was, it was said, little else for them to do during a two day layover 2,800 miles from home.
Figure 1 (pg.122) maps the events described in the paragraphs below)
Captain Hazelwood, immediately after assuming command, radioed the Coast Guard and requested permission to alter the course of the ship to avoid large chunks of ice that had broken loose from the Columbia Glacier and were floating in the outbound shipping lane. The permission was granted. Captain Hazelwood altered the course southward, turned over command of the vessel to the Third Mate, Mr. Gregory Cousins, and went below to his cabin. Mr. Cousins was not licensed to pilot a ship in the sea channels approaching Valdez. Mr. Cousins and others later testified that it was common practice to turn over command of oil tankers to nonlicensed officers, even in restricted waters. There was, it was said, little choice for the staffing -- both officers and crew -- had been dramatically reduced on the new, largely automated tankers as part of a cost savings move.
Captain Hazelwood had set the automatic pilot to steer the ship southward, across the inbound shipping land. He had instructed Mr. Cousins to maintain that course until after the ice chunks from the glaciers were past, and then to turn back to the northwest when abeam of Busby Light, a well known marker in the Sound. No inbound traffic was expected, and permission for this course change had been granted by the Coast Guard, so no danger was anticipated even though the ship was on a course headed directly towards Bligh Reef, the best known hazard in the Sound. At 11:55 Mr. Cousins ordered a course change of 10 degrees right rudder to bring the tanker back towards the proper lane and away from the reef. Either the automatic pilot was still connected, which overrode the manual steering, or the rudder did not respond promptly to the wheel, or Mr. Cousins had waited until too late to initiate the change. The tanker, nearly 1,000' long and weighing over 280,000 tons, continued inexorably on its heading and at 12:05 am hit the reef with awesome impact. 11 million gallons of crude oil (fortunately only 27% of the cargo) spilled from badly ruptured tanks on the port side of the vessel.
The emergency response team at Valdez had been disbanded, as part of a cost cutting move. 18 years of accident free operations had found no occasion for their services. The emergency response equipment had been used for other purposes, or shipped to other locations, or allowed to slowly deteriorate, for exactly the same reason. The lighting system needed for emergency work at night was not available. The containment booms needed to surround an oil spill of this size were not accessible. The ocean-going barge needed to transport those booms was not serviceable. The employees needed to load and then lay the booms were not trained. Despite the existence of a 250 page emergency response plan that promised to first contain and then start the recovery of any spill within 5 hours of its occurrence, this spill remained uncontained for 59 hours, and the recovery, once begun, was totally ineffectual due to a combination of deteriorated equipment, limited supplies, and inadequate training.
Wave action quickly converted the uncontained crude oil to a "mousse" mixture of oil and water that quadrupled the volume. This emulsified mixture lay 4 to 5 inches thick upon the surface of the sea and, driven by the strong winds common to the area during the early spring, began to coat the first of nearly 350 miles of shoreline. The worst accident in the history of the North American petroleum industry had just occurred.
The costs of that accident, in terms of wildlife deaths, ecological damages, economic hardships, and corporate losses were truly horrendous, though the estimates given in Figure 2.1 and Figure 2.2 are acknowledged to be very imprecise. The exact number of wildlife deaths, for example, is clearly open to debate. No one counted the oil-soaked birds and mammals as they washed up along 350 miles of beaches, and many of the corpses either sank or were driven by coastal currents out to sea. In one of the more depressing epilogues associated with the wreck, a study designed to sharpen estimates of seabird mortality killed 219 more birds, immersed them in oil, attached radio transmitters, placed them in the Sound, and then attempted to trace their drift patterns. The results were inconclusive due, it was claimed, to the small size of the sample. Fortunately, suggestions for studies on a larger scale were quickly overridden.4
The extent of the ecological harms are also subject to debate. Studies of sea water contamination traces, salmon and herring run fishery catches, coastal zone life forms, and inland wildlife health impacts have cost more than $100,000,000 and total almost 50 0,000 pages, but are inconclusive.5 Scientists and technicians hired by the Exxon Corporation interpreted the data to show that by 1993 there was no evidence of continuing exposure of wildlife to the toxic polyaromatics in the oil, and that the process of recovery was nearly complete. Researchers from the National Oceanic and Atmospheric Administration and representatives of the numerous environmental groups maintained in contrast that oil from the wreck was still seeping from the beaches into mussel mats and coastal waters, and that the reproductive health of harlequine ducks, sea otters, herring, salmon, and all other forms of maritime, coastal, and inland life continued to be adversely affected. Full recovery was said by these scientists to be at least ten years in the future.6
Regardless of the exact number of wildlife killed and regardless of the precise extent of the recovery experienced, it doubtless is safe to assume that everyone associated with the company, the region, and the environment devoutly wishes that the accident had not occurred. The purpose of this article is to examine how it could have been avoided, how our "fault-prone institutions" might learn to do better. Obviously, a sober tanker captain and an experienced third mate would have helped in this particular instance, but why were they not upon the bridge? And, why were the equipment, supplies, and personnel for a prompt response not upon the scene. Let us look at the responsibility for these conditions throughout the company. It is certainly legitimate to think of institutions as consisting of layers of responsibility arranged in a rank order approximating a traditional hierarchy in which "doing" is gradually replaced by "planning" (Figure 3). These layers range from "operational" where the daily, repetitive activities of the business are performed to the "conceptual" where external trends are recognized, competitive factors are analyzed, and long-term strategies are developed. At Exxon, all seemed to have contributed to the initial occurrence and then to the ineffectual response of the accident:
Operational level. This is the level at which products are made, customers are served, and bills are paid. It is also the level at which input materials are purchased, transported, and processed. As the crude oil was being shipped from Valdez to Long Beach it was under the direct supervision of an intoxicated tanker captain and an inexperienced third mate.7 In one sense, those two conditions were the sole cause of the accident. But, it also has to be recognized that 23 others people were aboard the tanker that night, and everyone on duty must have known that the tanker was off course, headed for a known hazard. No one took any action. Even further, there were perhaps 200 people at the shore base who must have known that the equipment, supplies, and personnel needed to respond to an oil spill were not available. No one had taken any action. No one on the tanker or on the shore seemed to believe that it was part of their job description to prevent or mitigate a potentially very severe industrial accident.
Functional level. This is the level at which the marketing, production, and financial activities are directed and short-term plans for those functions are prepared. It is also the level at which revenues are maximized, costs are minimized, and assets are controlled within a given region or division. At Exxon, as part of a cost reduction program for the shipping division, the crew size on tankers had been cut from 44 to 24 persons; numerous crew members after the accident testified that they were continually tired from working long shifts which perhaps accounts for their lack of attention. As part of a cost reduction program for the Alaska region, the emergency response team at Valdez had been disbanded, and training of the assigned replacements had been delayed. 18 years of accident free operations had made such teams and training seem redundant. Further, as part of an asset reduction program for the Alaska region the emergency equipment and supplies at Valdez had not been maintained at a ready state.
Technical level. This is the level at which the informational, individual, and technological capabilities for improved performance are developed. It is, more specifically, the level at which operational and financial data are processed and transmitted, workers and managers are selected and trained, and new products and processes are designed and disseminated. At Exxon, there was no information system on risk exposure; staff managers had looked at the danger points where volatile petroleum products were transported, processed or stored, and evaluated safety measures and response capabilities, but those evaluations were not widely distributed throughout the firm. There was also no human resource policy on risk containment; staff managers here had proposed that employees at the danger points pass certain standards and obey special rules (i.e., no alcohol consumption within 8 hours of a working shift), but those standards and rules were never enforced. Lastly, there was no research and development effort on risk reduction; staff mangers and scientists in this area were not looking into better ways of avoiding or responding to petroleum fires and spills.
Organizational level. This is the managerial level that is primarily concerned with issues of strategy implementation. It is the level at which tasks are assigned, structures are designed, controls are developed, and incentives are proposed. Exxon had inherent organizational complexities that made reporting relationships and control methodologies awkward if not ineffectual. It was a huge (100,000 employees, even after the downsizing), geographically separated, vertically integrated, global firm. There were multiple divisions and regions, separated by function, distance, culture, and even ownership. At the bottom of the organization chart were joint ventures such as the Alyeska Pipeline Company. Alyeska was coowned with six other oil companies, and charged with the operation of the shipping terminal at Valdez and the provision of emergency responses to local accidents. The performance evaluation and employee motivation systems in place at Exxon did not extend to the Alyeska joint venture, and did not include emergency response capabilities even if they had covered that remote outpost.
Conceptual level. This is the managerial level that is primarily concerned with issues of strategy formulation. It is the level at which industry trends are recognized, competitive factors are analyzed, and strategic alternatives are considered. The most critical external trend in the petroleum industry prior to the accident had been the decline of crude oil prices from $32.00 per barrel in 1981 at the height of the power of the O.P.E.C. alliance to $ 12.00 per barrel in 1986 at the nadir of that power. The price then rose slightly, and in 1989 had stabilized at approximately $16.00 per barrel. Vertically integrated oil companies for years had relied upon the ample margins available in the crude oil they produced from their own fields; now they were suddenly strapped for both profits and cash. Many ofthese vertically integrated oil companies had diversified during the 1970s and early 1980s into petrochemicals, and the rapidly growing/ high value-added markets for thermoplastics and agricultural fertilizers. These markets matured during the mid-1980s, which further cut margins, profits, and cash flow. Large companies such as Exxon were forced to reduce costs in order to survive, or at the least to avoid the wave of hostile take-overs that characterized the oil industry at this time, Those cost cuts spread throughout the organization, reducing indirectly the effort to avoid accidents and the directly the ability to respond to them. In the final analysis, the reduction in crude oil prices and the maturation of petrochemical markets appear to be as responsible for the wreck of the Exxon Valdez as the intoxicated tanker captain and the inexperienced deck officer.
The destruction resulting from the wreck of the Exxon Valdez on Bligh Reef in Prince William Sound was not an inevitable consequence of changes in the input factor price margins and the output market growth rates. Those changes made the damaging accident and the slow response more probable only because they were permitted to do so. No one at any of the levels of managerial responsibility felt that it was their duty to resolve the potential moral problems brought about by those changes as attentively as they worked upon the obvious financial problems. Moral problems are those that cause harm to others in ways outside their own control. Moral problems in management are particularly complex because they usually involve hurt or harm to some and help or benefit to others. That was certainly the situation at Exxon in the mid-1980s. Cost cuts and asset reductions produced immediate benefits for the stockholders, customers, and employees who remained with the firm; they also, however, created potential harms for the employees who were forced to leave and the communities who were forced to bear the risk of industrial accidents. The company employees who were either forced or chose to leave were vocal in their opposition, and were granted early retirement options. The local communities, however, had no voice in the strategic decision that inadvertently increased the risk to the environment upon which so many residents were dependent for their livelihood and their life style. There was no focus of managerial responsibility within Exxon to address this unstated moral problem with obvious benefits for some and potential harms for others. No senior executive at Exxon was willing to insist that cost savings for the company should not come about at the expense of local residents and their dependence upon the environment; consequently the probability of a disastrous industrial accident was greatly increased. Improvements in prevention means and response methods could have been made at every level had that insistence been voiced: Figure 4
Operational improvements. No one on the tanker or on the shore felt that it was part of their job description to report potentially hazardous conditions. It is easy to assume that the night of the accident was not the first time that the captain had been intoxicated nor the first time that the tanker had steered directly towards a known hazard with a 5 minute time frame for corrective action. It is also easy to assume that the night of the accident was not the first time that the emergency equipment was unavailable or unserviceable. Doubtless the first rule of accident prevention and amelioration is to have employees on the site personally involved in that prevention and amelioration, with a clear reporting channel for their concerns. Those concerns were not apparent and that channel was not available at Exxon.
Functional improvements. The emergency response teams were eliminated one at a time, at each of the ports, refineries, and oil fields. It would have been possible to mitigate that change by establishing two or three centralized response teams within the company, with contracts for air transport of personnel and equipment to the site of any future accident. Centralized response teams could be much better trained and much better equipped, at considerably lessened cost, as compared to the local groups. Doubtless the second rule of accident prevention and amelioration is to have employees ready to respond, properly trained and equipped, with a certain degree of esprit and elan. That centralized alternative had not been developed at Exxon.
Technical improvements. Oil companies, due to the large volume, toxic nature and inflammable characteristic of their raw materials and final products, are particularly susceptible to severe industrial accidents. It is certainly possible to identify the sites where those accidents are most damaging, and to be explicit about the risk exposure at each site. It is also possible to provide special policies (alcohol & drug prohibitions) and standards (workshift limitations) to contain those risks at each of the sites. And, it is even further possible to develop new technologies and design new methods to reduce those risks. The third rule of accident prevention and amelioration doubtless is for managers at all levels to be knowledgeable about the risks at each site, and to be encouraged to take the appropriate actions and make the necessary investments to contain or reduce those risks. That had not been done at Exxon.
Organizational improvements. Exxon is too large and complex a company to manage by "walking around". It is easy to assume that Lawrence Rawl, chairman of the company at the time of the accident, had never been to Valdez, and knew nothing of conditions there either on the tankers or upon the shore. Yet, within the corporate structure a senior officer could have been assigned responsibility for accident prevention and response, and to be the reporting focus for employee concerns. The control and incentive systems could have been changed to measure and reward functional and technical managers on their degree of attention to accident prevention and on their amount of preparation for accident response. The fourth rule for accident prevention and amelioration doubtless is to have someone clearly in charge, and to have control and incentive systems already in place to encourage positive efforts at containing and reducing the risks Neither had occurred at Exxon.
Conceptual improvements. The company was forced by the reduction in the margins of the crude oil they produced and by the increase in the competition of the petrochemical markets they served to change its strategy. The senior executives decided, based upon the company's obvious economies of scale and of scope, to make Exxon the low cost producer in the petroleum industry. There was extensive restructuring to reduce total employees from 145,000 in 1986 to slightly under 100,000 in 1988, which eliminated many of the existing positions and most of the current programs devoted to accident prevention and response. There was, however, no review of the strategic, program, and budgetary plans to determine who or what might be hurt or harmed, the extent of that hurt or harm, and the probability of its occurrence. The fifth rule for accident prevention and amelioration doubtless is to recognize that when the basic plans for an organization are changed there often are unintended consequences of those changes. That had not occurred at Exxon.
Moral improvements. None of these improvements which seem so obvious in hindsight, however, was going to occur unless someone at the most senior level in the company was willing to make them occur. The five improvements suggested above would not have been expensive for Exxon to implement. This was not an issue of economic efficiency versus environmental protection. This was simply an issue of a lack of managerial direction. None of the company's employees at any of the five levels of responsibility felt that it was part of their job to worry about the possibility of industrial accidents and to actively help in the prevention or mitigation of those accidents.
What could have been done to provide that direction? What could have been done by the most senior executives to ensure that employees at all levels of the firm worried about the possibility of industrial accidents and actively worked to prevent or mitigate them? The first step would have been to recognize the potential for harm to the legitimate interests of others. All firms have costs that are external to their production process. These are the "externalities" of economic theory. They are usually not formally recognized because they are borne by persons or groups outside the company. Someone within each firm has to take action to first formally recognize the costs and then to reduce or mitigate the harms This is the moral level of responsibility: A very simple but also very precise definition of this moral level of responsibility has recently been proposed by John Magee and Ranganath Nayak, chairman and senior vice president respectively of Arthur D. Little: When all interested parties cannot participate in making a decision -- which is often the case -- the decision-maker must take the legitimate interests of all constituents into account, whether they are there to argue their interests or not.8
No one at Exxon was willing to take the legitimate interests of all constituents into account, including the interests of those residents who happened to live in communities dose to Exxon's facilities for the potentially hazardous production, transportation, and processing of petroleum products, and including the interests of those persons who happened to value a clean environment for aesthetic rather than economic reasons. No one was willing to issue a statement of organizational values that listed all of the constituents and their legitimate interests in a dear order of priority for the guidance of employees at the lower levels of responsibility. No one was willing to prepare a set of planning objectives that went beyond financial returns to include those legitimate interests for the further guidance of managers and workers alike. No one, in short, was willing to challenge and then change a culture focused solely upon profit as summarized by the chairman, Lawrence Rawl: I'm bottom line oriented. I look at the revenues, and I look at everything that comes in between. When I find something that looks a little bit soft, I take a hard look. When the good times are rolling, you can ignore some of that stuff. But when times get difficult, you've got to do something. In fact, you should do it anyway. That's management. That's what shareholders pay us for.9
The most basic lesson in accident prevention that can be drawn from the wreck of the Exxon Valdez is that management is much more than just looking at revenues, costs, and profits. Management requires the imagination to understand the full mixture of potential benefits and harms generated by the operations of the firm, the empathy to consider the full range of legitimate interests represented by the constituencies of the firm, and the courage to act when some of the harms are not certain and many of the constituencies are not powerful. The lack of imagination, empathy, and courage at the most senior levels of the company was the true cause of the wreck of the Exxon Valdez.[Footnote]