Republished online with permission from the Fall 2007 issue of the Dispute Resolution Alert (JAMS)
If humans were completely rational and shared equal information, their legal decisions might resemble rational economic choices. But we know better than to assume that humans are completely rational and our experience with legal disputes is evidence of consistent information gaps. In this short piece (reprinted with permission from the Fall 2007 issue of the Dispute Resolution Alert (JAMS) (http://jamsadr.com/images/PDF/DRA-2007-11.pdf) and excerpted from a longer forthcoming piece in Vol. XIII of Harvard Negotiation Law Review), I describe a few ways that disputing parties and their lawyers systematically depart from rational decision making. Along the way, I offer tips on how to get productive settlement discussions back on track after being derailed by some pitfalls that are part of our all-too-human psychology.
A. Risk Tolerance and Loss Aversion
Some people are risk-takers and others are risk-averse. Filing and defending lawsuits is inherently risky, but having a higher risk tolerance than one’s opponent may be advantageous in negotiations. Differences in risk tolerance are a source of value creation – the party more willing to bear risk should get some benefit in the negotiation. If mediators knew more about parties’ risk attitudes, they could help them craft more successful settlement offers. Just what is known about decisions under conditions of risk?
Nobel Laureates Daniel Kahneman, Amos Tversky, and others have done important work in the areas of adaptive thinking and bounded rationality. While it is difficult to determine exactly how much more risk-seeking or risk-adverse a party to a particular suit is at a given point, research has uncovered important generalities.
Risk attitudes are dependant, in part, on whether the party faces a gain or loss. Plaintiffs generally seek recoveries that their defendants resist paying. Those roles change the lens through which each views potential outcomes. Unless they have high sunk costs or face fee shifting provisions, plaintiffs face a sure gain in settlement or the possibility of a larger gain at trial. In the absence of counterclaims or offers of judgment, defendants look through the other end of the telescope – they face a sure loss by settling or the potential of a bigger loss at trial. In experiments, Tversky and Kahneman found that a large majority of subjects facing gains preferred a certain $240 to a 25 percent chance of $1,000 (worth on average $250). On the other hand, when facing a loss, the same group preferred a 75 percent chance of loss of $1,000 (worth $750) to a sure loss of $750.
People tend to make risk-averse choices when facing a gain; that is, they prefer certain gains over larger but riskier gains. People facing losses, however, tend to make risk-seeking choices; that is, they prefer riskier outcomes to sure losses. Kahneman emphasizes the point by contemplating salary offers of $40,000 and $45,000 to people making $35,000 and $50,000. He then notes that the “psychological differences between the alternatives is likely to be greater in the latter case.”
In another experiment, groups of students were given the chore of buying and selling coffee mugs. With exactly the same information, the median price set by the sellers was $7.12, but the median buyer was only willing to pay $2.88 for the same mug. So it is with lawsuits. Well-intentioned parties and lawyers arrive at different valuations of the same outcomes not because of forces of good or ill, but differences in assigned position. Negotiators recognize that sellers and plaintiffs will usually assign higher values to a negotiated item than buyers and defendants.
Partisan aspirations are tempered, we hope, by the expert opinions of the lawyers. If those aspirations are aggressive, the chances of impasse naturally increase. To the extent that aggressive aspirations are the product of a risk-seeking attitude coupled with incomplete information, a mediator may unearth that reality and help the parties adjust their tactics and attitudes. At the very least, we would do well to recognize that not everyone views risk from the same perspective.
B. Optimistic Overconfidence
Life would be dull without optimists, but excessive optimism increases the odds of impasse. Faced with a nasty lawsuit, litigants want lawyers to be their champions, to hold an optimistic view of their chances. Most clients do not like their champions to be poking holes in their case. They expect mediators and judges to do that. But whoever does it, parties are unlikely to settle cases unless they perceive the negotiated outcome to be more attractive than their alternatives. While a cold water evaluation may dampen overconfident expectations, economic analyses iteratively guide litigants through probability-adjusted outcomes without turning them off by telling them they are wrong.
Overconfidence leads people to discount small probabilities and luck, and overestimate unattractive consequences. It is human nature to place more emphasis on facts that support desired outcomes and to make self-serving assessments of one’s own ability. More than 80 percent of interviewed entrepreneurs described their chances of success as 70 percent or better, and 33 percent described them as “certain.” That compares with a five-year survival rate for new firms in the 33 percent range. Couples about to be married estimated their chances of later divorcing at zero, even though most knew that the divorce rate is between 40 and 50 percent. Negotiators in final arbitrations overestimated the chance that their offer would be chosen by 15 percent.
Although most negotiators believe that they are more “fair” than average, in specific mediations they tend to overestimate their trial alternatives. People focus attention on assets while under-appreciating the issues on which their claim is weaker. A myopic focus on a case strength blurs focus on less favorable points. Focusing tightly on case merits runs the risk of undervaluing the transaction costs of continuing to trial.
However, while overconfidence is prevalent, it is not universal. A mediator cannot simply discount the value of each sides’ offers by the same amount or proportion. One side might be well calibrated while the other is far off. What they can do is prepare alternative scenarios looking through different ends of the same telescope. Some scenarios will be rosy and others thorny, but together they are more likely to cover the range of potential outcomes – worst case to best. Requiring specific explanations for various outcomes can break single-minded focus on one. In the process, overconfidence can be reduced.
C. Perfect Information
Lawsuits sound better to lawyers and judges when they only hear one side. As information improves, the bloom may fade. People undervalue aspects of the situation of which they are relatively ignorant. As an example, in one study, subjects given only half of the evidence in a case predicted the jury’s decision with greater confidence than those who were given all of it. Not only were they more confident than those who were better informed, they were not able to adequately compensate when told that their evidence was lopsided.
While people want all available information before making decisions, experts and seasoned executives are accustomed to making decisions under uncertainty. Shell executives made billion-dollar investment decisions based on Joe Jaworski’s hypothetical scenarios for the price of crude oil in 30 years. Business clients routinely take risks with limited information.
Lawyers are held to a different standard. Sixty percent certainty that a new product launch will be successful is considered great information. But missing 40 percent of the possible information in discovery may lead to a malpractice suit for a losing lawyer. Part of any litigator’s analysis should include the amount her side is willing to spend to find out additional information. Since price and risk are inversely correlated, if one accepts the risk of limited information by adjusting price downward, he should balance the risk portion of the equation too. The alternative puts the lawyer in the uncomfortable position of leaving rocks unturned when dealing with a client who is comfortable navigating risk with less information.
Decision trees can help determine how much to pay to close an informational gap. As one would expect, it has everything to do with the spread between the decision points (“litigate” v. “settle”). Assume a hypothetical driver involved in an automobile accident. Her (overly simplified) legal analysis tells us that her principal claim is negligence and that the range of remedies is $0 to $100,000. The probabilities are 50:50 for each outcome.
A settlement offer is outstanding for the expected value: $50,000. Graphically, the decision looks like this.
The economic analysis reflects the simplicity of the hypothetical – the plaintiff should be indifferent to the two options since they both equal $50,000. But the gap is wide. So expected value may not be as helpful as improving the information she has available to make a dichotomous choice.
While decision points are rarely this elementary, plaintiff’s decision is whether to accept the $50,000 offer or spend more money discovering additional information that may improve her odds – and the offer. Since she stands to double her money, she may seek more information than she might want in a closer call. But how much will she and her lawyer spend to take a swing at the $100,000 outcome?
Once plaintiff has a $50,000 settlement offer (or reasonably expects one in that range), she is bracketed by a choice between a 50 percent chance of recovering $100,000 and a sure $50,000 settlement. Since the offer comes early, she must make that choice with less than perfect information. Of course, if she knew the jury was coming back in an hour with $100,000 award, she would not settle (“win” fork). If she knew the jury was going to zero her out, she would take the offer. But her choices come in the real world. The amount she is willing to spend turns out to be half the spread between outcomes. We take the probabilities (50:50) and solve for the difference between the outcomes by examining each scenario. That means we set the “litigate” probabilities on the “win” and “lose” forks to 100:0. The “win” outcome is swinging for $100,000 at trial and the “lose” outcome prefers to “settle” at $50,000. The expected value for the new “win”/”lose” fork is $75,000. Therefore, plaintiff should be unwilling to spend more than $25,000 to obtain additional information to decide between a $50,000 settlement offer and the chance of 100,000 at trial. Of course, the information she discovers could also be damaging and push her closer to $0.
People often face these choices irrationally. Many will spend more to “increase the probability of a desirable outcome from 0.99 to 1 than from 0.80 to 0.85.” But that decision should be made wide-eyed. We all make decisions with less than perfect information. In litigation, we do well to balance price and risk.
D. Attribution Errors and Anger
The same psychological lenses that give us confidence also color our perception of others’ conduct. The likelihood of settling a lawsuit is impacted by the parties’ attitudes toward one another.
In his best-selling book Blink: The Power of Thinking Without Thinking, Malcolm Gladwell notes that “there are highly skilled doctors who get sued a lot and doctors who make lots of mistakes and never get sued.” The differentiator is not shoddy medical care, it’s “something else” – “patients say that they were rushed or ignored or treated poorly” and it made them mad. “‘People just don’t sue doctors they like,’ is how Alice Burkin, a leading medical malpractice lawyer, puts it.” Medical schools teach bedside manners and “[i]nsurers list a good bedside manner and a willingness to answer patient questions as effective ways to reduce the odds of facing a malpractice suit.
Trial lawyers are equipped as repeat players to help clients factor this bias into their analyses too. Mediators can help by probing alternative explanations for conduct in an effort to debias models, if not actually reduce anger. In the absence of such explanations, parties fill in the blanks – and make attribution errors in the process. In certain carefully chosen circumstances, apologies have been shown to reduce anger and increase the likelihood that a party will accept a settlement offer, but apology comes with risks.
One of the inherent strengths of economic analysis is that it focuses the parties on the component parts of the problem at hand. That is not to suggest that there is not an important and cathartic role for emotions and venting in negotiation, even in commercial disputes. There certainly are. It is to say that when deciding to pass up an opportunity to negotiate an alternative to litigation, the parties should objectively evaluate the price they put on those emotions. “[G]ive me liberty or give me death!” carried a lot of meaning. It also clarified the price one patriot was willing to pay for his alternative. While the alternatives to inevitable human conflict are usually less stark, it is important for our analyses to contemplate the attributions we are likely making about our opponent, and the ones they are surely making to us.
As we move from dispute analysis to negotiation planning, we are often faced with making the first offer or awaiting one from the other side. That decision turns on a number of variables. Because the car dealer knows its real costs, it posts a sticker price that is intended to push negotiations above those costs. With less information, we may await our opponent’s move. Their offer may telegraph informational asymmetries or align with our expectations. It may reflect overconfidence borne of ignorance and it might just be a strategic move. No matter whether they are rooted in reason or something else, first offers have power.
Psychologists call the phenomenon “anchoring” and have studied its influence on opening offers and demands, insurance policy caps, statutory damage caps, negotiator aspirations, and other “first numbers.” And while training and information asymmetry certainly limit the impact of anchors, even “real estate agents’ judgments about the market price of homes were influenced by manipulations of the list prices.” Anchors function much like our “gut” reactions to the value of an object or lawsuit. The more relevant information our analytical mind has, the less we are swayed by an unreasonable anchor. Mistaken or misguided anchors can increase the odds of impasse and have unintended consequences.
Information quality and symmetry can have a clear impact on the weight of an anchor. Our legal and economic analyses increase our confidence in our valuations and thus the offers we make. These analyses place us in a better position to influence the negotiations by dropping an anchor or disregarding an unreasonable attempt to anchor by another.
F. Reactive Devaluation
There are certain things we just do not want to hear from our adversaries. In fact, the perceived source of a message has a lot to do with our perception of it. We discount whatever the other side offers, even if it’s favorable (“They wouldn’t have offered those terms if those terms strengthened our position relative to theirs.”). We also tend to reject or devalue whatever is freely available and strive for whatever is denied – the “grass is always greener on the other side of the fence.” Student assistants were given the option of cash or authorship credit by a professor writing an article. The students who were offered cash expressed a desire for authorship credit. Those offered authorship credit wanted cash.
A Cold War experiment quantified the magnitude of this bias. Soviet leader Gorbachev made a proposal to reduce nuclear warheads by one-half, followed by further reductions over time. Researchers attributed the proposal to President Reagan, a group of unknown strategists, and to Gorbachev himself. The surprise was not that the group reacted differently to the same proposal depending on its source, but the wide range of difference. When attributed to the U.S. President, 90 percent reacted favorably. That dropped marginally when attributed to the third-party (80 percent), but in half (44 percent) when attributed to the Soviet leader. It also comes as no surprise that the responsiveness of Israeli student subjects to a proposed peace agreement between Israel and the Palestinians depends on whether they perceive the proposal as emanating from the Israeli government or the Palestinian Authority.
If we know that our proposals could be discounted by half just because of their source, we should consider the source in scenario planning. The arms control proposal from “unknown strategists” was viewed almost as favorably as the same one coming from the home team – nearly twice as favorably as when it came from the opponent. A mediator can accept one side’s demonization of the other and gently reframe the underlying issue as the parties work through various outcome scenarios.
G. Other Factors – And There Are Always Other Factors
There are always other factors impacting case valuation. One litigant may want to avoid the market or bankruptcy effects of an adverse verdict, the risk of a no-liability finding, or the distraction of litigation on management. Another may want to set precedent or ward off future claims with a consistent litigation strategy. Others may be intent on legislation or appellate decisions that change their opponents’ alternatives. We all use “rules of thumb” to short-circuit decisions. Sometimes they work, but if we overpay for something, we experience the “winner’s curse.” We perceive whatever we are selling to have a higher value than the buyer appreciates – the endowment effect.
Decision-makers allocate resources based on anticipated returns. Once we have thoroughly analyzed a case (or series of cases) from different perspectives, a decision-maker can better decide how much time and money she is willing to spend to make those points or avoid those costs. A hard-fought principle may be at stake – at least until an objective analysis places a dollar price tag on it. The existence of these and other psychological impediments to successful resolution call for objective models to test party aspirations. Mediators are well-positioned to check many of these biases as they nudge the focus back to future outcomes. The proper use of objective tools that continually redirect litigants to the problem rather than the personalities should only work to increase effectiveness.
O.K., the subject line was meant to shock you and to draw criticism for what I will admit is my greatest unresolved prejudicial default -- that white men over 65 who didn't participate in...By Victoria Pynchon
The following dispute could happen anywhere, as it deals with raw emotions, a long term friendship that went sour and the mediation process. The plaintiff and defendant were long term...By Jim Hildreth
The California Law Revision Commission has issued its draft recommendations for an amendment to the Evidence Code carving out a new exception to mediation confidentiality. This project has been about...By Joe Markowitz