The Social-Welfare Implications of the Extent of Liability
From an economic standpoint, the extent to which economic actors should bear liability for the negative effects of their negligence has social-welfare implications because the extent of liability affects incentives for future behavior. Steven Shavell, Liability for Accidents, in 1 Handbook of Law and Economics ch. 2 (A.M. Polinksy & S. Shavell eds., 2007).The purpose of liability rules is not merely to punish negligence or make victims whole, but, from an economic standpoint, the social interest in liability rules is in encouraging diligence toward and investment in prevention of future harm. At the same time, imposing liability has social costs if and to the extent that the liability rules discourage economic agents from engaging in desirable market activities, or encourage investment in accident avoidance that is excessive relative to the social benefits. In the context of environmental liability, laws and regulations should encourage firms that engage in activities with high potential environmental risk—such as oil exploration and production, mining, farming, electricity generation, and heavy industry—to exhibit diligence in avoiding harmful environmental outcomes and in promoting safe operations, taking into account the costs to society of encouraging excessive risk avoidance. Our premise, and what we believe is a fundamental premise of the economic approach to liability analysis, is that liability rules and limitations should, to the extent possible, maximize expected social welfare by establishing incentives that optimally balance the costs to society of insufficient diligence against the costs to society of imposing excessive risk bearing.
Society has to be concerned generally about two opposing adverse effects of establishing improper liability rules. On one hand, if the rules tend to impose inadequate liability, companies will undertake inadequate efforts (i.e., will incur insufficient costs) to ensure that employees engage in safety-compliant behavior. If, on the other hand, the rules impose excessive risk of liability or penalty, companies will either engage in excessive efforts to ensure that employees engage in safety-complaint behavior, or they will shut down certain operations entirely, even if those operations are social-welfare enhancing despite the risk of environmental accidents.
The notion of socially optimal incentive creation derives from the economics literature on what is known as the “principal-agent problem.” Sanford J. Grossman & Oliver D. Hart, An Analysis of the Principal-Agent Problem, Econometrica 51 (Jan. 1983). The principal-agent problem is that of establishing incentives for an economic agent (the standard example in economics literature is the CEO of a firm) who is hired by the economic principal (the firm’s owners or shareholders) that maximize the welfare of the principal. This is done by aligning the agent’s incentives with those of the principal as closely as possible, taking into account that the agent’s actions cannot be perfectly observed, and the effects of an agent’s actions are inevitably influenced by factors that are outside the control of the agent. One efficient means of establishing incentives for appropriately diligent behavior is for the agent to bear financial consequences correlated with the effects of his behavior. This is why, for example, efficient compensation systems for CEOs of publicly traded companies typically involve the CEO holding the company’s stock (or options on it). The company’s owners know that tying the CEO’s compensation to the performance of the company encourages the CEO to manage the company with the diligence and care that the principals themselves would apply. See id.
In our context, the “agent” is the party, such as an oil company, electric utility, or other entity whose actions might cause environmental harm; whose behavior would presumably be influenced by standards of liability (or what we refer to in this article as “liability rules”); and the relevant “principal” is society as a whole. We assume that, as a matter of jurisprudence, one would want to establish liability rules that maximize society’s welfare.
As a first principle, one might think that the most efficient liability rules for encouraging socially beneficial behavior would be those that impose on the actor all consequences of his or her actions, no matter how indirect or distant in time or space. Indeed, in a world of perfect certainty and perfect information, in which all consequences were known and predictable, and in which an actor could be rewarded for all positive consequences as well as penalized for all negative consequences of his or her actions, this might be true. But from a social perspective, the lack of perfect information and perfect certainty generally would render imposition of unlimited liability on an economic actor suboptimal because of its chilling effect on economic activity, and liability rules that establish some level of limitation are not only more consistent with the intuitive notion of fairness, but are generally consistent with economically optimal incentive creation.
As implied by the foregoing discussion, optimal incentives must take into account two opposing forces—the harms of inadequate incentives for diligences that would result from insufficient liability, and the harms of excessive risk imposed on economic agents, and the associated costs to consumers that would result from excessive liability. This can be illustrated with reference to the example of the CEO and shareholders mentioned earlier. On one hand, a failure to punish and reward a CEO in accordance with his firm’s profitability would be likely to result in an inadequate level of managerial effort and diligence relative to that desired by the shareholders. Similarly, a company that bore no liability for harms to the environment or harms to non-customers would be expected to provide inadequate incentives to establish protocols and procedures to protect against such harms.
The important and beneficial effects of rules that impose liability for harms caused are obvious and intuitive, and it is apparent that imposing insufficient liability for harmful events would have the undesirable effect of creating insufficient incentives to take due care. On the other hand, however, the harmful effects of excessive liability are equally implied by economic principles, though they are perhaps less obvious. For example, it is well understood as a matter of economics that an optimal compensation system for a CEO would typically not compensate him or her entirely on the basis of the stock value of the firm, because the latter is affected, both positively and negatively, by factors outside the CEO’s control. Imposing on the CEO the risk of extreme outcomes that are beyond the CEO’s control requires incurring additional costs to compensate the CEO for bearing that risk, and/or induces excessively risk-averse behavior by the CEO. See Paul Milgrom & John Roberts, Economics, Organization and Management 206–247 (1992).
Excessively risk-averse behavior might take the form, for example, of failure to engage in innovation or reduced introduction of promising new products, because the CEO recognizes that such projects may fail despite his or her diligent and skillful efforts. While undertaking these projects might be in the shareholders’ interests, imposing excessive risk on a CEO (essentially, punishing the CEO for the failure of a project, say, despite the fact that the failure may have occurred despite his or her extraordinary skills and best efforts) might stifle risk taking at all. Similarly, imposing excessive liability on a company via overly extensive liability rules would be harmful to society by inducing excessive risk avoidance. Optimal liability rules balance the benefits of incentives for diligent behavior against the social harms of imposing excessive risk bearing. See Shavell, supra.
Three Principles of Welfare-Maximizing Liability Rules
We identify three principles of efficient liability rules relevant to the issue of causation and proximity in major environmental cases that are based on general economic insights about risk bearing and incentives deriving from the results of the principal-agent paradigm that we just outlined. Before doing so, however, it is important to recognize that establishing liability rules for companies, as opposed to individuals, involves an additional nuance that is sometimes forgotten in such discussions. In general, when a catastrophic environmental event occurs, such as a large oil spill, massive forest fire, massive flood, or other such occurrence, the event may be the result of the behavior of individuals, but the liable party may be their employer. Companies are not individuals, and they may be able to influence but cannot control behavior of their employees. Companies can and do influence employees’ behavior by establishing rules, practices, procedures, protocols, incentives, penalties (including loss of job), monitoring, training, and reporting requirements; but they cannot fully control or determine what employees will or will not do. Hence, the meaning of negligence by a company is different from the meaning of negligence by an individual. An individual may be negligent by ignoring the company’s rules, while the company may have been fully diligent in establishing sound rules, procedures, and so forth. That is, failures can happen in at least two ways: (1) The company could have failed to establish rules, processes, incentives, and penalties that would be adequate reasonably to ensure compliant behavior by employees; and/or (2) even if a company did establish adequate rules, processes, incentives, and penalties, a given employee at a given time might not follow them and could still cause a failure.
If a company does not bear any liability for its employees’ negligence, the company will generally have inadequate incentive to establish properly aggressive rules and procedures to protect against employee negligence. See A. Mitchell Polinsky, An Introduction to Law and Economics 125–34 (3rd ed. 2003). But a company with optimal rules will not be able to prevent all adverse events unless it refrains from doing business at all, and will not be able to enforce its rules 100 percent of the time even under the best circumstances. Hence, the optimal liability rules—how much liability a company should bear as a matter of principle—certainly vary as a matter of economics, depending on whether the company was negligent in establishing and enforcing its rules, or whether it established appropriate rules and enforced them reasonably, but the event occurred as a result of employee negligence despite the rules and other controls in place.
Aside from these considerations, we identify the following factors as relevant to determining a socially efficient (i.e., welfare maximizing) set of liability rules:
1. the principle of control
2. the principle of causation
3. the principle of balance
The Principle of Control
The principle of control is that to create efficient incentives, the extent of liability should be linked with the extent of control. The harm from any given failure can vary for reasons that are not predictable and are outside the control of the firm and its employees. Consider, for example, a power company whose equipment failure causes a forest fire. The extent of the fire and damage caused will depend on a variety of factors unrelated to the equipment failure itself, such as the weather conditions before, during, and after the failure, and the speed and proficiency of the fire-response personnel. The power company may, through its rules and procedures, be able to influence the probability that equipment fails and sparks a fire, but it may have much more limited control or no control at all over the spread of the fire once it occurs. In such a circumstance, imposing liability for extreme outcomes may impose costs upon the company (and, therefore, on consumers) that outweigh the risk of catastrophic fire that society is willing to bear, because the company’s only ability to respond to full liability for catastrophic fires may be to decline to operate in certain fire-sensitive areas at all. Such an outcome may be contrary to the public interest.
One way to limit liability that is responsive to the principle of control (although not the only reasonable way) might be to limit the utility’s liability only to damage incurred within a certain geographic proximity to the origin of the fire. The issue is not one of causation per se—if an equipment malfunction sparks a fire and a residential neighborhood 50 miles away from the origin burns down because unusually dry conditions and high winds spread the fire an exceptionally long distance, it is clear that the equipment failure was (in interaction with the weather conditions) a causal factor of the lost property. However, the extreme level of damages may not have been under the control of the utility beyond its efforts to minimize the probability of any fire at all. Hence, a limitation of liability based on geographic proximity to the origin is not based on the logic of causation but on the logic of control.
More generally, an implication of the principle of control is that when considering the proper extent of liability, the concept of “proximity” should be assessed with respect to proximity of control. The more proximate are the allegedly negligent party’s levers of control to the outcome, the more likely it is that the efficient sphere of liability should include that outcome.
The Principle of Causation
The principle of causation is the obvious principle that parties should not be liable for outcomes that are not effects of their actions. For example, suppose that a resort on the Gulf Coast claims that it lost business due to an oil spill. Suppose, specifically, that its business was down 30 percent in the summer of the oil spill relative to the previous summer. If, however, 50 percent of that decline in business would have occurred without the spill due to an overall decline in tourism resulting from overall economic downturn, the oil company—if found negligent—should be relieved of liability for (at least) the half of the decline that the oil spill did not cause. The fact that an oil spill would reasonably cause a decline in profits at resorts on beaches polluted by the spilled oil, and the fact that a beach resort experienced a decline in profits, do not together imply that the entire decline in profits of that beach resort was the result of the oil spill. Indeed, more generally, the decline in profits caused by the spill could be more or less than the resort’s observed decline in profits relative to the previous year. The damages caused by the spill are not the difference between the pre-spill and post-spill profitability, but rather the difference between the actual post-spill profitability and the profitability that would have been achieved in the but-for world in which the same conditions held except for the spill. This is well understood and we do not believe we are making any new or controversial point here. We do observe, however, that the practical challenges of identifying harm caused by an event are not only significant but are likely to be more significant as the proximity of the harmed party to the event decreases.
The Principle of Balance
The principle of balance relates to the extent to which liability rules may create an inherent bias that discourages risk taking. Consider again the scenario in which an oil spill in the Gulf of Mexico causes resorts on affected beaches to lose business. This may be considered a direct effect of the spill. In addition, suppose resorts on unaffected beaches in nearby East Coast states also experience a decline in business because potential customers change their vacations plans based on the incorrect belief that those beaches also were affected. Further, suppose resorts in California experience an upturn in business, because travelers substitute California vacations for East Coast vacations. If the harm to East Coast resorts is $x and the benefit from the increased business to California resorts is $y, and if these were the only indirect effects of the spill, the net indirect harm would be $z = $x ‑ $y. At best, thesocially optimal liability that the oil company should face from an incentive standpoint for the indirect effects of the spill would be $z, not $x. Hence, requiring the oil company to bear the indirect liability $x will be excessive from a social standpoint and cause the oil company to engage in excessive risk-shielding, because it is neither feasible nor consistent with general legal principles as we understand them to calculate the benefits to the California resort owners and offset them against the harms to the East Coast resort owners when calculating damages owed to the East Coast resort owners, let alone for the oil company to recover those benefits from the benefited parties.
The principle of balance, like the other principles, does not provide a bright-line test for identifying the socially efficient point of demarcation beyond which liability should not be applied. Indeed, it suggests, again, that the point of demarcation may not be best thought of as associated with the metaphor of “proximity.” It does suggest, however, that where the avenue of harm is such that the harmed parties may have suffered to the benefit of other parties, that avenue of harm may fall outside the efficient sphere of liability.
Determining the limits to the effects of a negligent action in a specific case can be informed by how the finding affects investment on a going-forward basis. Expansive liability rules that seek to compensate victims far and wide from a negligent action will have the effect of increasing investment in risk-reducing activities and reducing the investment in risk-embodying activities. In contrast, narrow liability rules would be expected to result in less risk-reducing investment and more risk-embodying investment. Neither type of investment has a monotonically positive or negative effect on social welfare, but instead would generally improve social welfare up to a point, after which incremental investment in risk-reducing activities or reduction in risk-taking activities reduces, rather than increases, social welfare. The principles of control, causation, and balance are aspects of limit setting that should be considered when determining where the boundary of liability should lie in a particular case. Reasonable application of these principles will help ensure that society’s resources are efficiently allocated so as to properly balance the social welfare interest in encouraging private entities to exercise care and diligence with the social welfare interest in encouraging private entities to undertake inherently risky productive activities.
Keywords: litigation, environmental litigation, BP, Macondo, gulf spill, principle of control, principle of causation, principle of balance, Palsgraf
Dr. Debra J. Aron is an adjunct associate professor at Northwestern University and managing director of Navigant Economics in Evanston, Illinois. Dr. Francis X. Pampush, CFA, is director of Navigant Economics in Evanston, Illinois. The opinions expressed in this article are those of the authors and do not represent the opinions of Navigant Economics, Navigant Consulting, or any other individuals associated with those institutions. This article was presented at the ABA Section of Litigation Annual Conference on April 15, 2011, in Miami Beach, Florida.