Rethinking Civil Settlement

J.J. Prescott and Kathryn E. Spier, A Comprehensive Theory of Civil Settlement, N.Y.U. L. Rev. (forthcoming 2016), available at SSRN.

How should we understand settlement in civil litigation? In A Comprehensive Theory of Civil Settlement, J.J. Prescott and Kathryn Spier rethink civil settlement and take a significant step forward in the scholarly conversation about this topic. Generally, settlement has been understood as an alternative to a judicial disposition in the case. In this view, settlement is a zero-sum event from a systemic standpoint. (Of course, the parties negotiating a settlement may split the differences between them and both receive gains, but I am speaking of the court’s perspective here.) Prescott and Spier provide a new way of thinking about settlement as a continuum rather than as an either/or event that ends the dispute. At one end of the continuum is litigation according to the default procedural and substantive rules. On the other end is the termination of the dispute with an agreement. In between are many smaller agreements that parties can, and do, reach in moving toward resolution of their dispute. In explaining the implications of this insight, this article ties together disparate scholarship in a neat way. They support their argument with evidence from a sample of cases in New York’s summary jury trial docket.

The first step in their analysis is to define a settlement. Prescott and Spier define settlement as any agreement between the parties that improves their position in the litigation by some combination of (1) reducing adjudication costs, (2) mitigating losses due to risk, or (3) maximizing ex ante returns. This agreement need not end the litigation. A high-low agreement, for example, sets a range for the outcome of the case because the parties agree that regardless of what the adjudicator decides, they will set a cap and a floor to the damages. Still, the case goes to trial. In the binary view, a high-low agreement does not count as a settlement, but it is an agreement between the parties that mitigates losses due to risk. Prescott and Spier categorize partial settlements (that is, settlements that are on the continuum between no agreement and resolution) into three types: award-modification agreements, issue-modification agreements, and procedure-modification agreements.

An example of award-modification agreements, taken from the authors’ prior work and likely the inspiration for this article, are high-low agreements. Such agreements can have several consequences for litigation. For example, they might affect the level of investment in the suit by narrowing the range of possible outcomes. They might also allow parties to continue with the litigation while controlling risk leading to a resolution that they prefer to a complete settlement of the dispute. For example, the authors cite to agreements where different allocations of fault by the jury will result in different damages awards that have already been agreed upon by the parties. Thus when the parties cannot agree on how to allocate fault but can agree on the amount of damages they can still reach an agreement. Especially interesting to me, the authors show how these agreements can be structured as continuums (“smooth award-modification agreements”) rather than tiers (“kinked” agreements). This is an analysis I haven’t seen before and one that is very intriguing because, as the authors note, such a structure can increase the flexibility of the parties in negotiation. A linear settlement continuum might even be possible to institute in small cases, although other factors (such as entrenched practices) may stand in the way. And one can imagine a “smooth” award-modification agreement being very useful in larger cases.

Some of the implications of this analysis are not worked through in the paper but are quite important and open avenues for future work. That is another great contribution of this piece – once the conception of settlement is opened up, there are many opportunities to generate new ideas and analysis. For example, trials are very rare in the American court system. If the reason for the decrease in trials is due to risk adverseness, the high low agreement can cabin this risk and increase reward for both parties. Can this lead to more trials? The answer is a qualified yes, as we shall see in a moment. Another potential implication of such partial settlements is along the lines of Marc Galanter’s famous thesis of how repeat play can distort the legal system. Award-modification agreements may give repeat players who can obtain a high-low from their opponent added incentives to go to trial in cases that will benefit them in the long term. By mitigating risk, high-low agreements can create incentives for plaintiffs to take low value cases to trial where they would not in a binary settlement world.   A low verdict, in turn, will improve defendant’s leverage in negotiating settlements in future cases, because cases are usually settled by reference to other settlements or to trial outcomes. Whether such distortion in fact occurs could be a subject of future inquiry.

The second category of partial settlements that Prescott and Spier address is issue-modification agreements. They analytically separate different types of issues that might be more or less amenable to partial settlements, such as legal issues that are divisible (they offer the example of comparative fault) as opposed to legal issues that are not (for example, state of mind). They rightly point out that there may be some issues with respect to which the party may take unilateral action – by not disputing the issue at trial, for example. While such a unilateral decision cannot be considered a settlement, the defendant and plaintiff may each agree to one “free” issue – an issue that each concedes to the other. Jury instructions, they argue, are also an important moment for such issue-modification agreements to take place, yet as they point out much of the cost of litigation has already been incurred at that point. This opens the possibility for parties to consider entering into such agreements early on. Considering the implications of such developments is a further line of inquiry that is suggested by this piece. Should the substantive law change to accommodate partial issue settlements? Should the procedural law be structured to make such settlements easier to reach? We already permit bifurcation and trifurcation of trials; perhaps there are other mechanisms that could be considered.

The third type of partial settlements are procedure-modification agreements. This type of agreement is often discussed in the procedural literature under the term “contract procedure” but the authors helpfully fold it in to the discussion of settlements as well. Such agreements might include waiving the jury right, altering the timing of litigation, evidence rules, secrecy, bellwether trials, and appeals. In some ways, a significant amount of real-life litigation process is negotiated, so perhaps calling agreements for extension of time and discovery schedules “settlements” goes too far and erodes the category of settlement so much that it becomes incoherent. Still, there are ways in which altering the procedure can look like a settlement, especially when it relates to appeals or extrapolation from bellwether trials. And of course these different strands may be combined. In the GM bellwether trials now ongoing in the Southern District of New York, the first bellwether trial included a secret high-low agreement. How would such an agreement affect party investment in the case and what would this mean for the extrapolated cases? The insights from this article can help answer such questions.

The authors also discuss the interesting implications of attorneys’ fees, which can be categorized as both award and procedure modifications. These have been addressed in the law and economics literature but this paper connects them to the broader question of settlement. An agreement to apply the British rule, for example, can increase risk, causing parties to invest more in the litigation. The authors demonstrate how “doubling down” in this way can increase welfare for both parties in some cases. The authors do not mention Rule 68, but that rule provides the opposite scenario: a switch from a default one-way fee shift (a modified British rule) to the American rule. An initial offer of settlement on the lower end can limit the defendant’s exposure in terms of attorneys’ fees, altering the incentive to invest in the litigation. Here, too, future work is opened up. For example, what would be the effect of a “smooth” curve in shifting attorneys’ fees, along the lines that the authors suggest in award modifications? (That is, instead of shifting fees entirely parties might agree on incremental shifts). This might be a useful way of thinking about fees in large cases such as class actions or MDLs. The next step, indeed, is to consider the ideas presented in this article for multi-party and complex litigation where more strategic behavior is likely to occur.

At the end of the article, Prescott and Spier analyze a set of data from the New York court system’s summary jury trial program to demonstrate the use of partial settlements. As they explain, the sample of cases that enters into that program is skewed. The data are still very interesting, however, and the authors hint that they may do more with it. For now, they have found two things. First, they find that settling key issues such liability is positively associated with award modification agreements and that settling damages only is negatively associated with high-low agreements (as the low is set effectively at zero with a finding of no liability). Second, they find that parties who have partially settled their cases are more likely to continue through trial rather than fully settle their cases. Overall, the New York data show that cases with high-low agreements settle about 20% of the time. Accordingly, some types of systems that encourage partial settlement may also encourage trials, with the caveat that choosing the summary jury trial process is probably a choice in favor of trial anyway. It would be interesting to test this finding on different types of cases, especially higher stakes cases. In any event, this finding provides food for thought for those civil procedure scholars who favor trials.

A Comprehensive Theory of Civil Settlement broadens our conception of what a settlement is, leading to a number of implications and further areas of study. The authors of this article are both economists, and there were some things that I would have suggested they change because of the way I know procedure to work. For example, some procedural rules, such as those on amending pleadings, are so forgiving that parties are unlikely to settle around them. Such changes would not alter the analysis overall, but my reaction reminded me how much potential there is for economists and procedural scholars to collaborate or provide input on work that approaches similar sets of questions from different angles.

Cite as: Alexandra D. Lahav, Rethinking Civil Settlement, JOTWELL (April 14, 2016) (reviewing J.J. Prescott and Kathryn E. Spier, A Comprehensive Theory of Civil Settlement, N.Y.U. L. Rev. (forthcoming 2016), available at SSRN),