Introduction
An important decision in litigation—one no doubt informed by an understanding of the evidence and relevant legal issues—is how to evaluate the settlement value of a case. This decision will generally come up first in evaluating a case and will typically become more prominent when a party considers making a settlement offer or deciding whether to accept or reject an offer from an opposing party. There may also be additional instances in which the size of a settlement will be considered, such as when a potential acquirer evaluates the purchase of a company with outstanding legal issues, or if the size of a proposed settlement leads to a coverage dispute between the insured and its carriers. Consequently, many parties have a desire for a methodology for estimating settlement values that is both accurate and convincing.
Approaches to Estimating and Evaluating Settlements
Approaches to estimating and evaluating settlements, or settlement offers, can range from subjective to objective. At one extreme is the purely subjective analysis that may be conducted by an experienced professional, such as a lawyer, judge, or mediator. At the other is the objective quantitative analysis performed by a statistician or econometrician. As discussed below, there are benefits and drawbacks to both approaches that can vary depending on the circumstances in which they are used.
Subjective Approaches
A subjective analysis may involve an experienced professional reviewing the key documents in a case and evaluating the value of a case based on that information. One clear benefit from this approach is that the professional can do his or her best to incorporate all kinds of information, including information that does not lend itself to quantification or information that is unique to the case at hand, meaning that there are no comparable situations to use as bases for evaluating the effects of that information on the value of the case. Because this type of evaluation relies heavily on the competence and fairness of the professional, it will be most useful in internal analyses by a party that knows and trusts that professional. The professional will often be one of the litigators involved in the case, using his or her experience to evaluate settlement proposals as the case progresses.
In contrast, this type of subjective analysis will be less useful in areas where one may wish to convince the opposing party, who can have their own professional presenting an honest but different valuation of the case. Similarly, in a coverage dispute related to whether a settlement was reasonable, while professionals can use their expertise to highlight relevant features of a case that should push the value of a settlement up or down, there may be little basis for them to convincingly argue that they have quantified either the baseline value of the case before those features or the effect of those features on the settlement value.
Objective Approaches
At the other extreme are objective measures of settlement size that are based on statistical analyses of similar cases that have settled. Certain benefits of these analyses are generally obvious: they rely on "hard" data that may allow them to pass a Daubert1 analysis, whether that is formally necessary or just shorthand for some of the characteristics of an analysis that tends to make it more convincing. Similarly, some of the drawbacks are also well known: the statistical results necessarily depend only on the information used in the analysis, and any "soft" evidence that is not quantifiable will be excluded.
Among the less obvious benefits of an objective measure is that it provides not just a statement about the settlement value of a case, but also can provide data to allow for a meaningful comparison of any proposed or disputed settlement with the statistical prediction. For example, the figure below shows the relationship between actual and predicted settlement amounts for approximately 1,000 securities class actions settled since 1996. This prediction model is based on a number of quantifiable variables and discussed in more detail in a prior publication.2 Each point corresponds to an individual case, and the horizontal axis shows the predicted settlement and the vertical axis shows the actual settlement. Superimposed on these points is a line showing where all the points would lie if the actual settlements equalled their predicted values. Points above the line represent cases that settled for more than their predicted value while points below the line settled for less than their predicted value.
For discussion purposes, one point is highlighted in red in Figure 1.3 As can be seen, the point in red is well above the prediction line. In fact, one can estimate that, based on the statistical model, there was only a 4% chance that the case represented by the red point would have been that far from the prediction line if the case was "average" in terms of the factors not accounted for in the statistical model. Alternatively, one could say that this case would fit the prediction if it was in the outer 4% of cases in terms of the unusualness of the non-statistical factors, or, more specifically, in the 2% of cases for which those factors most favored plaintiffs (with the other 2% of unusual cases being those in which those factors most favored defendants). While this statistical analysis would not prove that the settlement was inexplicably high, it does provide some information about how unusually strong plaintiffs' case would have to be to explain the size of the settlement.4 And while it still may be difficult to assess how unusual a case is in terms of the factors not in the model, the use of a statistical benchmark would allow one to have a baseline to which one can apply those subjective assessments.5
Among the sources for statistical predictions of settlements are databases of actual settlements and objective factors about a case. For example, several consulting firms regularly analyze the relationship between announced securities fraud class action settlements and publicly available features of a case, such as the trading volume and stock price movements over the class period. Because class action settlements and trading data are publicly available, it is possible to construct such databases. One additional benefit of the regular publication of data regarding these models is that a model from a previous publication can be used, forestalling a claim that the model was designed to produce a certain result in the case being analyzed.
For an analysis that is more directly tied to the magnitude of claims in a case, one can compare settlements to the damages claim made by plaintiffs' expert in different cases. Because not all cases reach the point where such claims are made, it is important to note that the applicability of a prediction for an ongoing case may depend on what stage the case has reached. For example, the assessment of damages in ongoing cases in a formal expert report may be different from the damages assessment provided for mediation. Similarly, if plaintiffs or defendants get to selectively choose which settlements to announce, then any database of publicly announced settlements may not be reflective of all cases.
An In-Between Approach
One other potentially useful source for a settlement prediction that combines subjective and objective analyses is the use of mock juries. However, unless both parties to the litigation are participating in the mock jury process, the evidence and theories of the case presented may not reflect what would likely be presented at trial. In addition, because mock jury sessions will tend to last for a shorter time period than a trial, they may not give an accurate read on the relative strengths of simple and complex arguments. Ideally, a law firm or firm engaged in creating mock juries would have a database of decisions reached by a mock jury and later settlements or trial verdicts from which one could examine how well the mock juries predict actual outcomes and if the predictions are biased in any systematic manner.6
Conclusion
Determining whether a settlement value is reasonable for a case often involves both subjective and objective factors. It may seem trite to note that each type of method has its benefits and pitfalls, and even that the two can be combined in many instances. More interesting is the fact that these benefits and pitfalls affect the usefulness of the methods in different circumstances, with the more subjective analyses having a comparative advantage in cases involving highly idiosyncratic features and when a party is simply trying to make an internal valuation, and objective analyses having a comparative advantage in cases that are generally similar to others and when trying to convince another party.
Footnotes
1 Daubert v. Merrell Dow Pharms., Inc., 509 U.S. 579.
2 "Trends 2010 Year-End Update: Securities Class Action Filings Accelerate in Second Half of 2010; Median Settlement Value at an All-Time High," available at www.nera.com..
3 This is not an actual case, but was added to the data for purposes of this discussion.
4 One should note that not every settlement above the prediction line should be treated as materially abnormally large. Under such reasoning, any settlement not exactly on the line would be considered either abnormally large (and therefore represent an overpayment potentially forced on insurers) or abnormally small (and therefore represent an underpayment to the class). Statisticians typically consider the outer five percent of cases relative to a prediction to be unusual, and refer to those as statistically significantly different from the prediction at the 5% level, the standard level of statistical significance. Thus, in the absence of any other explanation, this case would be considered to be different and above the prediction by a statistically significant amount.
5 It is worth repeating that such subjective assessments must be about how unusual a case is. All cases have some strong and weak points, so the question is whether those points are more or less important than in a typical case. For example, if defense counsel believed that they could cut out fifty percent of the damages claim in a case, this would not say whether that position was unusually strong or weak for defendants unless one knew how much defense counsel typically thought they could reduce a damages claim.
6 Again, care is needed in making such extrapolations. For example, if a defense firm relied heavily on mock juries in deciding whether to settle or go to trial, it may be the case that when it accidentally gets a favorable mock jury result (i.e., a result due to circumstances not related to the merits of the case), it rejects reasonable settlement offers and proceeds to trial only to be met with a more typical result.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
