Stepwise liability: Between the preponderance rule and proportional liability

Abstract

There are two familiar decision rules: the binary, preponderance of the evidence and the continuous, proportional liability rule. This article proposes a thought experiment. Instead of all-or-nothing or a continuous rule, the law can utilise a middle ground—assigning liability stepwise, according to the procedural progression of the case—stepwise liability. Stepwise liability relies on the gradual design of civil procedure. Under the current system, the plaintiff has to pass several procedural thresholds with increasing evidentiary requirements in order to proceed to trial. Examples are a motion to dismiss and then a summary judgment. I propose that, corresponding to the procedural progression of the case, after surviving each step the plaintiff will be entitled to a gradually increasing share of the damages. Stepwise liability offers several advantages relative to the traditional rules. It provides partial compensation where the defendant's liability falls short of the 50% threshold, hence restoring incentives to take care. Unlike the proportional rule, this outcome can be achieved without major modifications to the existing decision rules. Unlike both rules, the proposal enables plaintiffs to cash in with some award before trial. I analyse the foregoing advantages together with the potential pitfalls, such as over-deterrence, larger legal expenses, and the day-in-court ideal.

Introduction

There is a long-standing debate in the literature about the preferable decision rule: the discontinuous, preponderance of the evidence rule or the ‘smooth’, proportional liability rule. The former, preponderance rule assigns the victory to the plaintiff only in case it shows that her case is more probable than her rival's (e.g., Mueller and Kirkpatrick, 2013: § 3.5). This is the default rule, at least in Anglo-American legal systems. In probabilistic terms, it is commonly thought that the plaintiff should win according to this rule if there is more than a 50% chance that her account is correct (e.g., Glöckner and Engel, 2013: 242). It thus presents an all-or-nothing decision rule. The latter, proportional rule allows for a partial remedy, based on the likelihood that the evidence justifies relief. Thus, a plaintiff who proves her case at a level of x% is entitled to that proportion of the damages, whether or not the evidence surpasses the 50% threshold. ¹ The proportional rule thus presents a continuous decision rule, which can be contrasted with the preponderance rule (e.g., Abramowicz, 2001: 235–236).

Each rule presents, of course, advantages and disadvantages. The most salient advantage of the proportional rule concerns ex-ante deterrence, as the proportional rule, by and large, better incentivises wrongdoers to take care, ex-ante (Pelled, 2021). To demonstrate, a wrongdoer whose responsibility, in probabilistic terms, is always 40% will never pay for her wrongdoing under the preponderance rule—an outcome that undermines the goal of deterrence. Likewise, a wrongdoer who is responsible for 70% will always be liable under the preponderance rule, and thus be incentivised to take excessive care. The proportional rule avoids these pitfalls. By contrast, the preponderance rule minimises, under certain assumptions, the erroneous allocation of remedies between the two parties, in that specific case (Kaye, 1982: 507–508). More generally, the preponderance rule is ‘our current default burden of proof rule in civil cases’ (Spottswood, 2021a: 202). It better fits the current legal climate, where the judge has to choose between two versions: the plaintiff's or the defendant's (Pelled, 2021). Indeed, while the proportional rule allegedly performs better than the preponderance rule ‘for all kinds of cases’, it faces the ‘innate conservatism of lawyers when considering reforms to trial procedure’ (Spottswood, 2021b: 824, 827).

This article contributes to the literature by presenting a thought experiment—levelling the choice between the discontinuous, preponderance rule and the continuous, proportional rule. Instead of all-or-nothing or a continuous rule, the law can utilise a middle ground—assigning liability stepwise, that is, a stepwise liability rule. In particular, the proposed middle rule is tied to the procedural progression of the case. To demonstrate, with each procedural step that the plaintiff passes—say, motion to dismiss and then summary judgment—the defendant's liability ‘jumps’ such that the plaintiff will be able to drop and cash in with a higher share. Thus, the proposal can correct some of the difficulties with the current rules. In particular, stepwise liability enables partial compensation where the defendant's liability falls short of the 50% threshold, hence restoring incentives to take care (think about cases in which plaintiffs survive a motion to dismiss but should fail at trial). Unlike the proportional rule (and hybrid rules that the literature has proposed), this outcome could be achieved without requiring the fact-finder to announce a probabilistic number. Thus, it better fits the role of judges in our system and necessitates little modifications in the existing legal landscape.

Stepwise liability, then, harnesses the gradual design of the legal process to achieve an outcome that resembles in some respects the proportional rule. Nevertheless, stepwise liability's effects are by no means identical to the proportional rule. It does not merely suggest ‘jumps’ in liability. As stepwise liability is tied to the procedural progression of the case, it enables the plaintiff to receive partial compensation relatively quickly, by dropping and cashing in after surviving the first steps. This feature should be contrasted with the proportional rule—which also grants partial compensation, but only after a complete trial. This characteristic facilitates access to justice to financially-constrained plaintiffs.

The proposal, of course, also entails potential pitfalls, which largely mirror the foregoing advantages. As a pro-plaintiff regime, in which plaintiffs with relatively weak cases acquire partial award, it could result in over-deterrence and a flood of petty cases. The ‘stepisation’ of the legal process presumably implicates legal expenses. Quick resolution might be inaccurate. And awarding (partial) compensation without the option to a full trial might infringe the day-in-court ideal. I discuss these and other issues below and suggest possible ways to mitigate the concerns.

This article is by no means the first to suggest a middle decision rule, between the two familiar decision rules. Indeed, a few articles have suggested some variation of a mixed decision rule, which borrows elements from both extreme decision rules. Abramowicz (2001), for instance, suggests a ‘compromise approach’ where in relatively clear cases the fact-finder would return an all-or-nothing decision; and in closer cases the proportional rule would prevail. Along somewhat similar lines, Spottswood (2021b) mathematically formulates a ‘logistic’ burden of proof, which resembles the continuous, proportional rule around the 50% rule, and the traditional all-or-nothing rule at the extremes. Such hybrid approaches to the legal decision rule can achieve the benefits of both. Abramowicz highlights the gains in risk reduction and fairness where the proportional rule applies in close cases. Spottswood stresses that a hybrid rule can strike a better balance between the goals that each traditional rule achieves (that is, deterrence and reducing error rates). Relatedly, and in the spirit of this article, Picinali (2022) suggests an intermediate criminal verdict—a step between the two binary options, guilty/non-guilty, which requires a lower threshold but attaches a weaker sanction. Picinali shows that such intermediate criminal verdicts have historical roots in continental legal traditions; and proposes the intermediate verdict of conditional acquittal. ² Other relevant works discussed the two familiar rules—the preponderance rule and proportional liability—and suggested that each rule can better fit different areas (Pelled, 2021). Along these lines, Kaplow (2012) conceptualises the burden of proof as maximising welfare—including the benefits of deterrence and the costs of chilling benign behaviour. Kaplow thus suggests that the optimal rule should differ across contexts.

This article follows this literature. In the spirit of the works of Spottswood (2021b) and Abramowicz (2001), I seek to introduce a middle-ground rule between the two familiar decision rules, the preponderance rule and proportional liability. Similarly to Picinali (2022), I also suggest intermediate steps, which require weaker evidentiary thresholds but also attach a milder outcome. Unlike previous literature, though, the stepwise liability proposal does not require full trial. It contributes to previous literature by harnessing the gradual design of the legal process and linking the outcome—stepwise liability—to the procedural progression of the case. This characteristic enables, in a sense, an expediting process in which the plaintiff could terminate her case and receive partial award after passing each step. This partial-realisation highlights the trade-off between the day-in-court ideal and facilitated access to justice.

The article proceeds as follows. The first section introduces stepwise liability by laying out the gradual design of current legal procedure and clarifying the possible design choices for the proposed regime. The second section analyses stepwise liability along several dimensions such as the role of judges, deterrence, settlements, legal expenses, etc. The third section concludes, and the Appendix contains technical discussion regarding the effects of stepwise liability on deterrence and settlements. ³

The proposed regime

This section presents the proposed regime. To do so, I will first depict the current, piecemeal design of legal proceedings, which underlies the stepwise liability proposal. Against the backdrop of alternative designs choices, I will present two variations of the proposed rule.

The piecemeal design of current legal proceedings

To understand how stepwise liability works one has first to lay out the piecemeal nature of the legal process. In a current civil lawsuit there are typically several stages that the plaintiff has to pass. These steps are crucial to the implementation of stepwise liability. The following will describe in more details the two common steps that a plaintiff typically has to pass before trial—the first is a motion to dismiss, and the second is summary judgment.

The defendant can move first to dismiss the plaintiff at the beginning, based on the plaintiff's pleading. According to prevailing precedents, the judge will dismiss the case if at that preliminary phase the plaintiff does not present ‘enough facts to state a claim to relief that is plausible on its face’. ⁴ Motions to dismiss are typically considered at an early stage, before discovery. Indeed, although ‘[t]he weight of evidence required to overcome a motion to dismiss is unclear’ (Bambauer and Day, 2011: 1090, n. 264), this evidentiary standard is apparently interpreted as a relatively low one, and certainly below the 50% threshold required to win the case (Herman and Sacher, 2015). To the extent the plaintiff survives dismissal (or the defendant refrains from moving to dismiss) the case proceeds to discovery. Importantly, this standard was interpreted as an evidentiary standard concerning the strength of the evidence at that stage, namely, factually thin cases can be dismissed. ⁵ I discuss below this point to a further extent.

Later on in the life of a case, and after pre-trial discovery, each party can move for ‘summary judgment’. ⁶ A case can be resolved at the summary judgment stage, without moving forward, ‘if the movant shows that there is no genuine dispute as to any material fact’ (Federal Rules of Civil Procedure, Rule 56). Note that unlike motions to dismiss, both the plaintiff and the defendant can move to show that the rival party's raises no ‘genuine dispute’. Again, this standard expresses the assessed strength of the case at that stage. The precise evidentiary requirement to survive summary judgment, that is, the probabilistic translation of the ‘genuine dispute’ standard, is likewise unclear; but is certainly higher than the one required to survive a motion to dismiss, and lower than the 50% threshold.

Finally, then, if the court denies a summary judgment motion (or the parties refrain from moving for summary judgment ⁷ ), the case should proceed to trial (bench or jury). The plaintiff in a civil case wins only when she meets the well-known ‘preponderance of the evidence’ standard, i.e., if the plaintiff's case is more likely than not (Zamir and Ritov, 2012). This evidentiary standard is commonly interpreted as a 50% threshold (Zamir and Ritov, 2012; Glöckner and Engel, 2013).

The stepwise liability proposal harnesses the foregoing piecemeal nature of the legal process. As we have seen, a typical case proceeds gradually. At each phase the fact-finder possesses more information—at the motion to dismiss stage, the fact-finder solely decides based on the pleading. At the summary judgment stage, the judge typically acquires more information—as the parties have already exchanged information through discovery proceedings. At the final, trial phase, the fact-finder also has the benefit of direct testimonies.

Relatedly and importantly, these thresholds could be interpreted in probabilistic terms, that is, as reflecting the strength of the case on a 0–100% scale (Lavie et al., 2020). ⁸ Then, at each phase the evidentiary threshold is higher than the former. While the ultimate standard to win at trial, the proverbial more-likely-than-not standard, is commonly interpreted as a 50% threshold, the standard to survive earlier phases is lower. As we have seen, these pre-trial standards do not have a clear numerical value. ⁹ But obviously the standards progress, that is, it is harder to survive a summary judgment than a motion to dismiss, and it is harder to win at trial than surviving a summary judgment (hence the pre-trial standards should be set somewhere between 0% and 50%).

One can liken, then, the legal process to an imaginary ladder which the plaintiff has to climb. The first, motion to dismiss step requires, say, a 20% chance that the plaintiff's account is true. The second, summary judgment step equals, in probabilistic terms, to a showing of, say, 35% on the part of the plaintiff. The final, trial step is the familiar 50% probability. The precise probabilistic threshold of each step, as well as the number of steps in each case, can of course vary. In certain areas there are more steps to climb (think of certification of class actions); and in actuality the parties do not always trigger the steps. But the general picture is clear. Currently, a civil case proceeds gradually, with several steps that the plaintiff has to pass, whose evidentiary thresholds gradually rise.

A qualifying note on probability thresholds

The foregoing takes a somewhat simplistic approach to procedural thresholds, which merits a qualifying note. ¹⁰ I interpret the current procedural thresholds as reflecting the overall probability—assessed at that early point—that the defendant is liable. ¹¹ This might fit an ordinary case, in which the two parties present competing stories, and more evidence is presented as the case proceeds. However, this interpretation could also raise difficulties, which stem from the fact that at the first stages the court accommodates a narrow body of evidence. ¹² To demonstrate, take a case of two steps, a preliminary dismissal step, with a concomitant standard that is translated as 30%; and a trial phase, which requires the familiar 50% threshold. Suppose that all possible evidence is presented at the first step, and the judge likewise believes that the defendant is liable with 35%. However, the judge also knows that no other evidence could be presented later. Under the proposed, stepwise regime, the plaintiff should pass the first step—but the judge should dismiss the case immediately thereafter (as there is no new evidence, and the case is certain to fail at trial). Under the current law, though, the judge would probably dismiss the case at the first stage. This example suggests that judges do not simply apply probability thresholds in preliminary stages. ¹³

This qualification notwithstanding, I believe that the interpretation of procedural standards as probability thresholds (concerning the likelihood that the defendant is liable) reflects the essence of judicial decision-making in many cases. This interpretation indeed acknowledges that a case will pass the first step but fail the next one. And even more so: due to the new information, the judge may reveal at the second step that the case should not have passed even the first step. ¹⁴

The remainder of this article, then, takes the approach that judges interpret (or should interpret) procedural standards as probability thresholds, concerning the likelihood that the defendant will be found liable.

Design choices

With the procedural progression of the case in mind, we can now better sketch the mechanics of the stepwise liability regime. The following demonstrates two implementation versions and discusses other relevant practical considerations.

Sunk steps and disappearing steps

Take a case in which the plaintiff has to pass two steps: a motion to dismiss, with an evidentiary threshold of 20%, and a summary judgment, with an evidentiary threshold of 35%. Suppose also that the damages are not disputed. In that case, stepwise liability should simply assign gradually increasing award to the plaintiff, based on the procedural stage that she crossed. Suppose, for simplicity, that the ‘price’ for surviving a motion to dismiss (summary judgment) is 20% (35%). The following uses this numerical example to elaborate on two design choices—that I refer to as ‘sunk steps’ and ‘disappearing steps’—considering the possibility that the plaintiff will pass one step but fail the next one. ¹⁵

Sunk steps

Under this alternative, the plaintiff who passes each step is guaranteed a partial award. That is, even if she loses later her gain is already sunk, and she ‘falls back’ on the previous award. Practically, after passing each step the plaintiff can choose whether to quit or proceed, without affecting her partial award. In our numerical example, to the extent the plaintiff survives a motion to dismiss she will be guaranteed a sum of 20% percent of the damages; and if she passes the summary judgment stage she is guaranteed to get 35% of the damages. If the plaintiff wins at trial she receives the entire pie of damages. ¹⁶

Disappearing steps

Under this version, a partial award disappears upon moving forward to the next stage. Hence, this version places a burden on the plaintiff—whether to quit and gain the partial award or proceed to the next step and risk losing that award. To illustrate, in our previous example a plaintiff who won on summary judgment can choose whether to cash in with 35% of the damages; or to proceed to trial, in which she receives nothing if she loses and the entire pie if she wins. Note that plaintiffs should prefer sunk steps to disappearing steps, as the latter exposes a plaintiff who chooses to move forward to a loss of her previous gains.

Additional procedural issues

Before proceeding to the merits of the proposal, the following briefly discusses additional procedural issues.

Setting the relevant numbers

Who sets the relevant figure at each step, e.g., that a motion to dismiss is ‘worth’ 20%, etc.? I envision a regime in which these ‘prices’ for each procedural step are pre-determined by policymakers: by the Supreme Court's precedents or through the Rules of Civil Procedure. In the following I will accordingly assume a given set of prices. There are, of course, alternative design options, e.g., letting the judge decide these numbers on a case-by-case basis, ¹⁷ or a mutually agreed set of numbers. ¹⁸ These design options seem inferior, and, at any rate, lie beyond the scope of this article.

Should the numbers correspond to the probabilistic threshold? In the numerical example above, the price (20% and 35% of the damages) fits the evidentiary threshold that is required to survive each step. But this need not be the case. An alternative approach is setting the price (in numerical terms) independently of the verbal standard.

Although the former approach might have some advantages, ¹⁹ in this article I take the latter avenue. In particular, I see the main advantage of the latter approach in the flexibility it allows in setting the relevant numbers—for instance, to the extent stepwise liability creates over-deterrence, policymakers could lower the ‘price’ for surviving each step. ²⁰

Disputed damages

Thus far I have assumed that damages are not disputed. What happens if damages are disputed? In that case it is impossible to award the plaintiff her proportion of the damages before trial.

The answer is two-fold. First, I envision stepwise liability as subject to the discretion of the presiding judge. A case in which there is a serious dispute with regard to the damages possibly justifies giving up the stepwise liability option. There are many cases in which the issue of damages appears relatively simple—statutory damages lawsuits, for instance, seem highly suitable for stepwise liability. Second and relatedly, where she deems it appropriate, the judge can simply split the damages from the determination of liability. ²¹ Indeed, in actuality it is common for judges to split the adjudication of liability and damages, where liability is determined first (Gensler, 2000). ²² As a side note, although practical difficulties abound, the stepwise approach could theoretically be administered in property cases too, e.g., by allocating the plaintiff a gradual share in the property as the case proceeds. ²³

Further ‘stepisation’ of the legal process

The proposed regime envisions gradually increasing steps—in our example, in order to reach trial, the plaintiff has first to pass a motion to dismiss and then to survive summary judgment. However, in the current practice these steps do not always materialise. Rather, they are triggered at the parties’ discretion. Only the defendant can trigger a motion to dismiss the plaintiff's case; and each party can force her rival to survive summary judgment. ²⁴ Importantly, the plaintiff cannot compel the judge to examine whether her own claims cross one of these thresholds. ²⁵ It is the defendant who possesses that power, and she may abstain from so doing for various reasons.

Stepwise liability, then, necessitates a more complete ‘stepisation’. One approach is to mandate the plaintiff to pass certain procedural stages. In several contexts the law does erect mandatory steps, that is, it requires judicial approval in order to proceed. A notable example is class actions—the law requires ‘certification’ of a class action before proceeding to trial. ²⁶ This approach, though, largely deviates from the current adversarial climate. A milder approach, which I endorse in this article, is to give the plaintiff the power to move the judge to decide whether she meets the relevant legal threshold (such that her own case is terminated if she fails to pass that threshold). To heed the importance of such a power, suppose that a plaintiff (knowingly) files a questionable case. ²⁷ Such a plaintiff may wish the court to rule, as soon as possible, on a motion to dismiss. This could be useful in order to clear the cloud over her own claims, signal the judge's position and avoid a losing course of action. Under the current law the plaintiff cannot invite such an early assessment of her case, and she has to wait for the defendant to move to dismiss. However, defendants do not always move to dismiss even if they apparently have sufficiently strong claims. ²⁸ The power to self-terminate her case should she fail to meet the relevant legal threshold, helps, therefore, to implement stepwise liability.

The case for stepwise liability

We can now turn our look to examine the major benefits and flaws of the proposed regime, relative to the more familiar ones—the preponderance rule and proportional liability. This section discusses stepwise liability, based on our recurrent example and the foregoing two versions, sunk and disappearing steps, in light of several important considerations, e.g., deterrence, settlements, etc. Mostly, the proposed regime can achieve better incentives to take care than the preponderance rule; and is easier to implement than the proportional rule. It can also ease the burden from financially-constrained plaintiffs and can affect settlements in various, subtle ways.

Probabilistic adjudication and the role of verdicts

An important benefit of stepwise liability over the proportional rule is the former's ease of implementation within the current system. Proportional liability (and other, hybrid solutions that rely on some form thereof) forces the fact-finder—a judge or a jury—to announce a probabilistic number at the end of a trial. The preponderance rule, by contrast, solely requires the fact-finder to determine the winner based on a given legal standard. As will be explained below, stepwise liability resembles the preponderance rule in this respect.

The issue of probabilistic adjudication provides various reasons to prefer the all-or-nothing, preponderance rule over proportional liability (Abramowicz, 2001: 250–55). For instance, fact-finders may find it ‘difficult to translate subjective estimates of probabilities into numbers’ (Abramowicz, 2001: 250–51). Similarly, one can argue that fact-finders ‘simply think better with standards than with numbers’ (Abramowicz, 2001: 252). ²⁹

A related line of argument concerns the acceptability of legal verdicts and the public message sent thereby. Allegedly, the fact-finder should provide in its verdict ‘a statement about what happened’ rather than ‘a statement about the evidence presented at trial’ (Nesson, 1985: 1358). Such a declaration is, according to some views, necessary in order for the public to accept the verdict and ‘affirm[] [the] behavioral message’ of the legal rule. ³⁰ The preponderance rule conforms to this idea. Regardless of the strength of the evidence, the fact-finder cuts the case, for the plaintiff (if the evidence crosses the 50% threshold) or the defendant (otherwise). This binary approach holds even in close cases—‘dura lex, sed lex’. ³¹ By contrast, the proportional rule fails in this respect. ‘Either the plaintiff suffered a wrongful loss and should be compensated for it, or she did not’ (Pelled, 2021: 193).

While one can debate the wisdom of the foregoing, this reasoning is definitely a major consideration in the current debate over the proportional rule. Indeed, courts have allegedly hesitated to implement the proportional rule for these very reasons (Nesson, 1985; but see Abramowicz, 2001). Corrective justice theorists resist the proportional rule and support the more traditional, preponderance rule on somewhat similar grounds (Pelled, 2021: 193–97).

Against this backdrop, it is easy to see that stepwise liability is considerably easier to implement in the current legal landscape than the proportional rule. At no time does the fact-finder need to declare a number that corresponds to her assessment of the evidence. ³² Rather, under the proposed stepwise rule, the fact-finder's role remains the same—deciding whether the plaintiff's passed the relevant legal threshold (motion to dismiss, summary judgment and/or trial). As judges routinely determine whether a specific set of facts meets a legal standard, in this sense the proposal does not carry any change. Accordingly, the decisions under the proposed rule could emit the same public message that is currently sent by courts. Although the distinction between a judge who declares a probabilistic liability (under the proportional rule) and a judge who translates the procedural standard to probability threshold may seem artificial (cf. Tuzet and Esposito, 2023: 325–329), it is based on existing practices. ³³

In sum, then, stepwise liability fits the existing role of courts and is thus easier to implement within the current legal framework than the proportional rule. In this sense, the proposal mimics one of the major advantages of the traditional, preponderance rule. This advantage of stepwise liability equally pertains to both the sunk and the disappearing steps versions.

Deterrence

We have seen that the proportional rule is rejected, at least in part, because it deviates from the current practice of judges. This concern notwithstanding, a major advantage of the proportional rule is creating better incentives for wrongdoers. Where wrongdoers calculate their actions in a state of uncertainty, proportional liability—rather than the preponderance rule—pushes them to fully internalise the harm they create (e.g., Pelled, 2021: 173–78). ³⁴ To take a trivial example, a physician whose chances of being found liable are 70% will always pay the entire damages under the preponderance rule (over-deterrence); and vice versa, a 30% liability will be translated to no liability under the preponderance rule (under-deterrence).

In general, stepwise liability rectifies at least some of the difficulties that are associated with the preponderance rule. In under-deterrence situations, under the 50% threshold, the plaintiff is entitled to at least part of the damages should she manage to pass pre-trial stages. In addition to imposing some liability below the 50% threshold, the proposal also facilitates quick pay, before reaching a full trial, and thus eases access to justice for financially-constrained victims (relative to the proportional rule). ³⁵ However, the proposed regime will not rectify over-deterrence situations under the preponderance rule. Beyond the 50% threshold the plaintiff wins the entire pie under both the proposed regime and the preponderance rule. ³⁶ By contrast, the proportional rule rectifies both under- and over-deterrence.

As a generalisation, then, stepwise liability resembles proportional liability in consistent under-deterrence situations (liability < 50%), and it is identical to the preponderance rule in consistent over-deterrence situations (liability > 50%). Subsections A and B to the Appendix analyse more complicated settings to better understand the effects of the proposed regime on deterrence where defendants calculate their behaviour ex-ante are under uncertainty with regard to the future realisation of their acts. The main points that emerge are as follows.

First, there might be important differences between the sunk and the disappearing steps versions. By and large, the disappearing steps version is closer to the preponderance rule. To demonstrate, if we think that the plaintiff is always optimistic—and she always proceeds to trial—then stepwise liability is identical to the preponderance rule. More generally, the more we think that the plaintiff suffers from the tendency to mistakenly proceed to trial, the more the disappearing steps version approaches the preponderance rule.

Second, it is clear that stepwise liability regimes result in over-deterrence relative to the preponderance rule—after all, stepwise liability grants partial compensation below the 50% threshold and full compensation thereafter. However, the precise comparison between stepwise liability and the preponderance rule is situation-dependent. Sometimes the preponderance rule is optimal and stepwise liability over-deters. ³⁷ Sometimes both the preponderance rule and stepwise liability over-deter, where the latter over-deters to a larger extent. ³⁸ Sometimes both rules under-deter. ³⁹ And sometimes the preponderance rule under-deters and stepwise liability over-deters. ⁴⁰

Third and relatedly, while stepwise liability would always generate a higher liability than the preponderance rule, by decreasing the ‘price’ for passing each step one could mitigate this outcome. The fact that the expected liability under the preponderance rule serves as a lower bound for the outcome under stepwise liability leads to several inferences. Subsection C to the Appendix accordingly generalises: where the preponderance rule over-deters, stepwise liability perforce over-deters. Where the preponderance rule under-deters, and stepwise liability over-deters, policymakers could reduce the price for each step such that the expected liability under stepwise liability is optimal.

* * *

Deterrence is a main advantage of proportional liability over the preponderance rule, and it is important to understand the precise effects of stepwise liability in this respect. Relative to the preponderance rule, the proposed regime increases the defendant's expected liability, hence it can address under-deterrence situations (for example, negligence that results in evidentiary weak cases). However, stepwise liability implicates risks of over-deterrence. While these risks do exist, below are a list of reasons for which the case for stepwise liability, in my view, holds.

First, it is true that stepwise liability increases deterrence. But the scope of over-deterrence varies, and is situation-dependent; sometimes both stepwise liability and the preponderance rule under-deter. More generally, in over-deterrence situations policymakers can adjust the price for each step downward—a lower price mitigates or even eliminates the over-deterrence problem. ⁴¹

Second, similarly to the proportional rule, stepwise liability might be implemented selectively. ⁴² It is tempting to utilise the proposed regime in areas in which we think that there is under-deterrence, and we would like victims, even with weak cases, to receive partial (and quick) compensation.

Third, there are various other mitigating factors. In general, for instance, the defendants are better off under the disappearing steps option (as in the sunk steps option the plaintiff could proceed without risking her partial award). This feature reduces over-deterrence concerns and renders the disappearing steps option more palatable. In particular, where the plaintiff is overly optimistic or misinformed, the disappearing steps version brings the proposed regime closer to the preponderance rule. Furthermore, under each regime the parties can well settle—where they typically compromise their positions and expected gains, and, by extension, reduce the over-deterrence problem.

Finally, while this issue is hotly debated, many believe that, as a rough generalisation, under-deterrence is a more pressing problem than over-deterrence in the current climate. In that case, the proposed rule is an improvement relative to the default preponderance rule. Legislative amendments and new precedents have made it harder for plaintiffs to file and litigate a lawsuit in the last decades in the U.S. (e.g., Moore (2015); Bell Atl. Corp. v Twombly, 550 U.S. 544 (2007)). Legal expenses and informational gaps hinder plaintiffs from filing claims, and skew the distribution of cases in courts. ⁴³ Furthermore, forward-looking wrongdoers, who can control and predict the strength of the evidence in a prospective case, can avail themselves of the preponderance rule. Under the preponderance rule a sophisticated wrongdoer can in theory freely commit a tort as long as it keeps the evidence against it below the 50% threshold. ⁴⁴ While this risk also exists under stepwise liability—e.g., the wrongdoer can keep the strength of the evidence at the 19% level, to avoid the first ‘step’—its scope is narrower.

Settlements

It is well-known that most cases settle. Settlement save legal expenses, for both the parties and the legal system. However, settlements can compromise various other goals, e.g., deterrence (cf. Fiss, 1984). Regardless of the social value of settlements, the stepwise liability regime affects settlements, and these effects can be subtle and intricate (cf. Spottswood, 2021a). I consider in this subsection two influential settlement theories. The first maintains that settlements fail due to the parties’ divergent expectations, in the form of mutual optimism with regard to the outcome. The second attributes settlement failures to asymmetric information, e.g., the defendant knows whether she is liable or not, whereas the plaintiff only knows the general probability that the defendant will be found liable (and not whether that specific defendant is negligent). ⁴⁵ The analysis concerning the two versions of stepwise liability is, by and large, similar. The following first considers, in detail, the sunk steps version; then, I will briefly highlight the unique features of the disappearing steps regime.

Sunk steps

We will first discuss the decision to move forward or drop the case after each step. Recall that under the sunk steps version the plaintiff, after passing each step, is guaranteed the price for that step regardless of the outcome at later stages. The plaintiff under the sunk steps regime will move forward if the legal expenses, from that point, justify the remaining expected award. The following illustrates:

Illustration (a). Suppose that the damages are 200, and the plaintiff expects to win at trial with probability 60%. The expected judgment, from the plaintiff's perspective, is thus 120. Suppose also that the plaintiff has passed the summary judgment stage, and that the concomitant pre-determined ‘price’ is 35% of the damages. Therefore, the plaintiff now receives regardless of the outcome at trial 35%*200 = 70. If she moves forward and wins, she receives the remainder of the damages, 130. Assuming that the plaintiff expects to win at trial with the same probability, 60%, the expected value of moving forward after surviving summary judgment is 60%*130 = 78. The plaintiff will thus move forward and bear the remaining legal expenses if the costs of doing so are smaller or equal to 78.

Observe that the sunk steps version makes trial less attractive relative to the current regime, precisely because some of the award has been sunk. Specifically, under the current regime the plaintiff in our example will move forward if the remaining legal expenses are lower than 60%*200 = 120. Hence, if the remaining legal expenses are in the range 78–120, the plaintiff will drop under the sunk steps version but move forward under the current regime. However, legal expenses decrease as the case proceeds. Hence, unless trial is particularly costly relative to pre-trial phases, we should expect the plaintiff to proceed under the sunk steps regime.

Divergent expectations

The following will analyse the decision to settle under the sunk steps regime, assuming the divergent expectations theory of settlements.

Lower stakes. In general, the lower the expected outcome at trial the more likely the parties to settle. The sunk steps version provides a guaranteed amount after passing each step. Hence, after each step the stakes are lower and, other things being equal, settlements are more likely. Intuitively, the parties have fewer incentives to fight where part of the compensation is guaranteed, and the remaining pie is thus smaller.

Lower legal expenses. The lower the legal expenses, other things being equal, settlements are less likely. Intuitively, parties that need to spend less on litigation are more inclined to go to court. In the context of stepwise liability, after each step some of the legal expenses have already been spent, such that the remaining legal expenses are lower. Hence, settling after passing each step is less likely, other things being equal.

We showed that, under the sunk steps regime after each step the stakes are lower (more settlements) but the remaining legal expenses are also smaller (more litigation). What is the net effect of these contradictory factors? Appendix D shows that settlements are more likely after each step to the extent the legal expenses of both sides decrease at a lower rate than the reduction in the remaining stakes, and vice versa. The precise effect of stepwise liability in this respect is, therefore, hard to predict. ⁴⁶

Judicial signal and mutual optimism. The divergent expectations model assumes that the parties differ with respect to their (subjective) expectations from trial. This gap in expectations hinders settlements, as it is harder for optimistic parties to settle. Thus far I have assumed that passing each step does not affect the parties’ expectations from trial. However, in actuality each ‘step’ provides the judge with the opportunity to express her opinion on the case. This information presumably narrows the expectation gap between the parties and facilitates settlements. ⁴⁷

In this sense, then, the parties are more likely to settle after each step. Indeed, examples could be drawn from other contexts. Preliminary injunction decisions, for instance, implicate the merits of the case—hence they inform the parties about the judge's position and facilitate settlements (Jeitschko and Kim, 2013). Another example for this reasoning could be drawn from the certification step in class actions—a fact-intensive procedure that conveys to the parties information regarding the likely result at trial. Indeed, it is well-known that the parties often settle immediately after (and if) the class manages to navigate the certification step (e.g., Hart, 2005: 780). There are other examples that share a similar spirit. ⁴⁸ In sum, the signal from the judge under each step seems as an important contribution to settlements under the proposed regime.

One can argue, though, that these incentives to settle after passing each step exist already, without rewarding the plaintiff for passing each step. However, as I have shown above, the proposed regime provides a more comprehensive ‘stepisation’ of the legal process. ⁴⁹ Hence, it should be useful in further reducing the gap in the parties’ (subjective) beliefs regarding the outcome, encouraging settlements.

Additional factors. The foregoing implicitly assumed that parties are risk-neutral. However, for risk-averse parties, who desire to avoid the risk of trial, settlements are relatively more attractive (Prescott and Spier, 2016: 73–75). The sunk steps version guarantees a certain amount regardless of the outcome at trial. Thus, it ‘hedges’ the plaintiff's side and reduces the overall risk from litigating, particularly after passing each step. In a sunk steps regime, then, and other things being equal, risk-averse parties may be more likely to go to trial. ⁵⁰

Stepwise liability could affect the indirect costs of litigation in other ways. The sunk steps version, in particular, enables quick payment for financially-constrained parties. This effect, in essence, reduce legal expenses and may again discourage settlements.

Asymmetric information

The foregoing results by and large hold under the asymmetric information theory of settlements.

Lower stakes and legal expenses. In general, as before, the lower the stakes of a trial, and the larger the legal expenses, the parties are more likely to settle under the asymmetric information theory (see generally Bebchuk, 1984: 409–10). These intuitions work in contradictory directions—after each step the stakes are lower but the legal expenses are lower too. Hence, the precise effect of stepwise liability is situation-dependent. Appendix D shows that, under certain assumptions, and contrary to the foregoing analysis, if the ratio of remaining legal expenses to damages remains the same after each step, settlements are less likely; settlements can be more likely only if the decrease in costs after each step is smaller than the reduction in the expected damages.

Additional information. The asymmetric information theory maintains that settlements fail due to informational gaps. To the extent that each step provides information, settlements are, of course, more likely under stepwise liability. Indeed, interim steps often implicate the exchange of private information: affidavits, depositions, discovery and even the mere filing of motions and answers to these motions. As before, the more complete ‘stepisation’ of the legal process under the proposed regime should thus encourage transmission of information and induce settlements relative to the current regime. To the extent that the parties anticipate that private information would later be revealed, they are also more likely to settle ex-ante.

Additional factors. As before, to the extent stepwise liability lowers related costs of litigation, e.g., the risk of trial or the need to finance litigation, settlements are less likely. Similarly, the higher expected payoff for the plaintiff under the proposal should raise the amount for which the parties settle.

The content of settlements

Finally, the content of settlements could also be affected by the stepwise regime, regardless of the theory of settlement. Specifically, the sunk steps regime improves the plaintiff's bargaining position. Consider the following:

Illustration (b). Consider the numerical example in illustration (a) above, that is, the damages are 200 and the plaintiff expects to win at trial with probability 60%. Suppose also that the legal expenditure for each party, plaintiff and defendant alike, are 30—10 before the summary judgment stage and 20 thereafter. Finally, assume the parties litigate under the American rule, that is, the loser does not reimburse the winner. Where the plaintiff proceeds after winning on summary judgment, she expects 200 with 60% (winning at trial); and 70 with 40% (losing at trial and keeping the fruits of surviving summary judgment). Net of legal expenses (assumed to be 20), then, the plaintiff expects at that point a payoff of 60%*200 + 40%*70–20 = 128. Thus, the plaintiff will not accept any settlement offer that falls short of 128.

Observe that the expected payoff of the plaintiff from proceeding is higher under the sunk steps regime than in the current regime, as in the current regime the plaintiff could only expect (net of legal expenses) 60%*200–20 = 100. Likewise, one can verify that, after summary judgment, the defendant expects to pay under the sunk steps version (including trial costs) 60%*200 + 40%*70 + 20 = 168. However, under the current regime the defendant expects to pay 140.

These differences allow the plaintiff to extract a better settlement. Suppose, for simplicity, that the parties have equal bargaining power, that is, they meet at the middle of the settlement range. In that case, the plaintiff settles at the summary judgment stage for 148 under the sunk steps regime, and for 120 under the current regime.

Disappearing steps

The foregoing analysis largely pertains to the disappearing steps version. There are, though, several important distinctions between the two regimes.

Dropping

Under the disappearing steps regime proceeding to trial is less attractive for the plaintiff, as there is no guaranteed amount in the event of a loss. Hence, the plaintiff is more likely to drop after passing each step. Consider the following:

Illustration (c). Consider the numerical example in illustration (a) above, and suppose that the plaintiff survives a summary judgment, with a pre-determined price of 35% of the damages, such that she can cash-in with 35%*200 = 70. If the plaintiff proceeds, she expects to gain 200 with probability 60%, or, 120. In that case, the plaintiff is willing to spend up to 120–70 = 50 in order to move forward.

Recall that in the sunk steps version the plaintiff was willing to spend up to 78 to proceed. ⁵¹ This suggests that when the legal expenses associated with moving from summary judgment to trial are in the range 50–78, the plaintiff drops under the disappearing steps rule and proceeds under the sunk steps regime (as well as under the current regime).

The inclination to drop also depends, of course, on various other factors. To the extent the plaintiff is risk-averse, trial becomes even less attractive, and cashing in is more likely. A higher pre-determined price makes trial less attractive, and can push plaintiffs to drop. Optimism makes a trial more attractive—and a super-optimistic plaintiff will always move forward.

Lower stakes

The disappearing steps version no longer reduces the stakes once the plaintiff decides to proceed after passing each phase. Observe that once the plaintiff moves forward, her previous gains are eliminated; hence, from that point, both parties again fight over the entire pie, whereas the remaining litigation expenses are lower (similarly to the current practice, in essence). ⁵² This characteristic neuters the advantage concerning settlements that the sunk steps regime possesses in this respect.

Risk

The sunk steps regime ‘hedges’ the plaintiff's bet. By contrast, the plaintiff who decides to proceed under the disappearing steps regime faces a riskier choice. This should make trial under the disappearing steps regime less attractive, inducing risk-averse parties to settle (or drop).

Content of settlements

The sunk steps option ‘hedges’ the plaintiff and thus leverages her bargaining position relative to the current regime. But this effect does not extend to the disappearing steps regime. In the disappearing steps version, a plaintiff who survives summary judgment under our recurrent numerical example (Illustration (b)) expects 60%*200 + 40%*0–20 = 100 from going forward (net of legal expenses). One can verify that the defendant expects to pay, inclusive of trial expense, 140. Therefore, the bargaining range is 100–140, identical to the bargaining range in the current regime. Although the disappearing steps option does not improve the plaintiff's bargaining position, observe that the plaintiff is better off under this option relative to the current regime. This advantage stems from the fact that in certain circumstances the plaintiff is better off dropping under the disappearing steps option (relative to the current regime). ⁵³

* * *

The proposed regime has various subtle effects on settlements. The most important insights are the following. First, the effect of surviving each step on settlements is ambiguous, as there are various moving forces in the background. An important issue in this respect is the information that the mere survival of each step conveys, as this information could streamline settlements. This factor seems most important where we attribute settlement failures to the parties’ optimism. Second, there are some differences between the two versions of stepwise liability. By and large, the riskier option, the disappearing steps regime, encourages dropping after each step; by contrast, if the plaintiff decides to proceed, the sunk steps version induces more settlements. Of course, the parties could ex-ante heed these effects, and settle accordingly at the beginning.

One could wonder whether the proposal is different from the current practice, in which parties typically settle after the intermediate steps. ⁵⁴ I argue that the proposed regime differs from the current practice in several respects. First, not all cases settle, let alone at the beginning of the case. Some cases make it to trial; others litigate through the first stages till a late settlement is reached. The proposal enables a middle-ground result, early on, in all cases. Second, in the current regime interim decisions indeed provide information, hence parties often settle thereafter. However, the proposal suggests a more complete ‘stepisation’, which should enhance the flow of information through judicial decisions and encourage settlements. Third, and perhaps most importantly, the proposal also affects the content of settlements relative to the existing framework. It sets a clear price for each step (which can be tweaked by policymakers), and, as it enables the option to cash in with a certain award, it improves the plaintiff's position (in settlement negotiations or through the option to drop).

Additional considerations

Legal expenses

Will stepwise liability reduce or increase litigation costs? There are reasons to think that the proposed regime will raise, overall, litigation costs. The following briefly highlights several effects.

Costs of stepisation

As elaborated above, stepwise liability adds more steps to the process relative to the current practice, hence it increases legal expenses to both the parties and the judge. ⁵⁵

Should we expect the parties to always trigger the steps? Recall that I envision a regime in which each party could trigger procedural steps. It is plausible to think that, by and large, at least one of them would wish to do so. It is true that in the current regime defendants sometimes avoid raising motions to dismiss. ⁵⁶ As plaintiffs are generally better off under stepwise liability, the defendants’ incentives to avoid each step are stronger. One could also think of situations in which the plaintiff wants to avoid the procedural steps, e.g., a plaintiff who possesses weak evidence but is optimistic regarding her odds a trial. To the extent there are instances in which both parties avoid triggering the procedural steps, the costs of stepisation are lower.

Filing of cases

The proposed regime raises the plaintiff's payoffs. Hence, it likely motivates victims to bring cases that are not filed under the current, preponderance liability regime. To illustrate, suppose that a plaintiff can survive a motion to dismiss but her odds at later stages are slim. The capacity to quickly cash in encourages filing of these cases. More filings of course generate more legal expenses. ⁵⁷

Dropped cases

The addition of steps could also reduce legal costs. First, cases that fail to survive the steps are dismissed, saving future legal costs. Second, plaintiffs that survive these steps could drop and cash in. As abovementioned, this effect is most conspicuous under the disappearing steps version. ⁵⁸ We should expect risk-averse and financially-constrained plaintiffs, for whom it is harder to proceed, to drop at a higher rate.

Investment in litigation

Another possible factor in favour of stepwise liability is the incentives to invest in litigation. Consider a case that goes to trial. As the stakes grow larger, parties likely have bigger incentives to invest in litigation—spending more lawyer-hours, expending on experts, etc. The all-or-nothing, preponderance rule raises the stakes, as the outcome is binary. It thus creates stronger incentives to invest in litigation than the continuous rule. This effect is the strongest in close cases, where each party's additional investment can relatively easily tip the scales. ⁵⁹ Stepwise liability lies between the preponderance and the continuous rules—some cases that would have ended in an all-or-nothing outcome under the preponderance rule would result in an intermediate award under the proposed regime. Hence, the proposal should decrease the incentives to invest in litigation relative to the default, preponderance rule. ⁶⁰ By the same logic, the incentives to spoliate evidence should also be lower under stepwise liability relative to the default, preponderance rule. ⁶¹

There may be, then, various effects of the proposed regime on legal expenses, and it is hard to accurately predict its precise outcome. Overall, though, one can conclude that a proposal that highlights the stepwise nature of the legal process is bound to add steps—and costs—to the legal process.

Allocation of errors

The literature that compares the preponderance and the proportional liability rules often highlights the different allocation of error costs between the parties under each decision rule—that is ‘the amount of dollars that are either wrongfully awarded to an undeserving plaintiff or withheld from a deserving plaintiff’ (Spottswood, 2021b: 792). As a generalisation, and assuming that we equally value errors against each side, one can show that the preponderance rule—as opposed to proportional liability—minimises the costs of errors of allocating the award between the two parties.

The idea of minimising the allocation of errors is not self-evident. ⁶² To the extent one deems it an important factor, how does stepwise liability fare in this respect? As a hybrid rule, stepwise liability's costs of error lie between the two familiar rules. ⁶³ To see this, observe that cases whose evidentiary strength crosses the 50% threshold should end in a complete win for the plaintiff under both stepwise liability and the preponderance rule (but in a partial award under the proportional rule). This result reduces error costs under stepwise liability relative to the proportional rule in this group of cases. ⁶⁴ By contrast, cases whose evidentiary strength is below the 50% line should end in a partial award under both the proportional rule and stepwise liability (but in a zero award under the preponderance rule). ⁶⁵ This result increases costs of error under stepwise liability relative to the preponderance rule.

Process accuracy

The foregoing discussed costs of allocating errors under a given decision-rule. A different issue is the accuracy of the implementation of each decision rule. Alternatively put, the previous subsection asked what is the erroneous allocation of remedies, given that the evidentiary strength is x%. The following discusses whether we can trust the evidentiary strength x% that we observe. ⁶⁶

Stepwise liability leads to less accurate decision-making precisely because it grants weight to intermediate procedural steps. As more evidence is proffered throughout the life of a case, the final outcome should be, by and large, more accurate than the outcome in earlier stages. Accuracy, in turn, has social value (e.g., Kaplow, 1994). ⁶⁷

To demonstrate this argument, take a case in which the plaintiff has survived a motion to dismiss. In the current regime, the plaintiff is entitled, at this step, to nothing. Under stepwise liability, by the mere virtue of surviving a motion to dismiss the plaintiff is entitled to part of the damages. Suppose that later, after additional evidence is presented, it turns out that the plaintiff should not have survived dismissal. In essence, the plaintiff was mistakenly entitled to a partial award (had the additional evidence been presented earlier).

Although stepwise liability entails a less accurate process, there are various mitigating factors. First, the problematic situations are those in which the assessment of the plaintiff's case materially changes throughout the case; however, the extent to which these situations are common is unclear. Moreover, even if the plaintiff has mistakenly survived a motion to dismiss, as in the example above, she might still be entitled to something under the proportional rule (e.g., ∼10% of the damages) had the case proceeded to trial. Second, under the disappearing steps version the problem seems less pressing—as plaintiffs who believe they have good cases move forward and then lose. Third, loss of accuracy can be compensated by other means. A broader right to appeal throughout the case—interlocutory appeals—can improve judicial decision-making.

Finally, the loss of accuracy should be compared to the current regime. It is well-known that most cases settle anyway. Apparently, current settlements do not fully reflect the expected judgment either. Furthermore, the settlements under the stepwise regime may or may not be more accurate. It has been shown earlier that the sunk steps regime improves the plaintiff's bargaining position—and to the extent the defendant has bargaining power, settlements under the sunk steps option better reflect the law. ⁶⁸

Equality

The proposed regime benefits plaintiffs, but it particularly helps poorer ones—to both file a case and go to trial (or, to credibly threaten to go to trial). These effects should also manifest themselves in the amount that poorer plaintiffs can extract through settlements.

First, unlike proportional liability and the preponderance rule, stepwise liability allows plaintiffs at the least the choice to drop the case and cash in after surviving each step. Moving forward to trial, even for relatively good claims, is costly. Hence, this choice—whether to quickly drop with a partial award or proceed to trial—should particularly benefit plaintiffs who face financial constraints and promote equality among plaintiffs. ⁶⁹ Second, the capacity to drop and cash in also reduces the overall risk from adjudication—as the plaintiff can expect, sometimes, an intermediate rather than all-or-nothing outcome. This again benefits the most the poorest plaintiffs, who are apparently the most risk-averse. These two effects, the capacity to enjoy a quick payment and the reduced risk from adjudication, are stronger for the sunk steps option, which can guarantee a certain amount regardless of the result at trial.

The day-in-court ideal

The stepwise proposal enables the plaintiff to receive compensation conditional on passing an intermediate procedural stage. However, in the current regime (and under proportional liability too), the court could award the plaintiff only after a judgment, and trial, on the merits. Moreover, under the stepwise approach a plaintiff could win a partial award even though she would have lost at trial. It is true that pre-trial settlements do exist in the current regime, but the parties’ consent is essential.

I believe that this line of criticism fades, to a large extent, when we better look at the nature of adjudication under stepwise liability together with the current practice. Stepwise liability does not eliminate the role of judges. Judges are expected to provide a reasoned decision at the preliminary steps. These decisions are based on the motion filed by one party—but also on the rival's response. Often, they are also based on some evidentiary background, e.g., documents. These decisions are also subject to (interlocutory) appeal. It is hard to say, then, that defendants under stepwise liability completely lose their right to a day in court. Indeed, important and influential decisions have been given by courts through preliminary steps.

This line is all the more stronger when one looks at the current realm of civil litigation. De-jure, each litigant is entitled to her day in court and a trial-based, judgment on the merits. Deviations from this day-in-court ideal could happen only through both parties’ consent. However, the notion that judges actually decide all cases seems in the current climate outdated, as virtually all cases settle and few reach trial. Furthermore, given the growing dockets most judges today largely view their role as actively advancing settlements. Indeed, it is common to describe the modern judge as ‘leading’ or ‘encouraging’ the parties to an irenic solution through aggressive pre-trial involvement, namely, ‘managerial judge’ (Resnik, 1982). This change in the traditional role of judges has received fierce criticism (ibid.; Galanter and Cahill, 1994; Fiss, 1984). Particularly, it infringes the core procedural rights of litigants without their consent, with little appellate review. Litigants often ‘capitulate to judicial pressure rather than risk the hostility of a judge who…has ongoing responsibility for the case’ (Resnik, 1982: 413).

It is hard, then, to argue today that litigants receive a meaningful day in court. They are often pressed to settle at early stages, under uncertainty. The stepwise approach moves little away from these practices. In essence, it formalises the reality in which the defendant pays a certain sum to the plaintiff, at early stages and under uncertainty.

The foregoing also reflects the historical evolution of the Federal Rules of Civil Procedure. The original, 1938 rules envisioned a single-step process—trial on the merits. However, the increasing burden of cases pushed courts to erect more demanding pre-trial phases. ⁷⁰ These additional steps hinder plaintiffs from proceeding and obviate the day-in-court ideal. Stepwise liability restores some of the spirit of the original rules, as the plaintiff could be entitled to some compensation through simple (and even single-phased) proceedings.

* * *

The idea that this article presents, granting liability as a proportion of the damages, according to the procedural stage the plaintiff survived, has notable pros and cons. On the one hand, stepwise liability offers partial compensation to those who are not able to meet the proverbial preponderance of the evidence standard. This feature is likely to raise deterrence, particularly for the group of wrongs that consistently suffer from weak evidentiary power. In addition, the link between stepwise liability and the current procedural design guarantees that it could be implemented with minimal changes to the existing framework—in essence, policymakers only need to set a price for the current standards. Moreover, the ability to achieve quick resolution with partial compensation, that is, to ‘cash in’ after passing the first steps, promotes equality among plaintiffs.

On the other hand, stepwise liability also presents some conspicuous pitfalls. The mirror image of quick (though partial) compensation in weak cases is concerns for over-deterrence, a flood of petty cases, larger legal expenses and inaccurate decision-making.

Are the advantages greater than the disadvantages? To further push the idea of stepwise liability I would like to offer several general directions for thought.

First, the balance between the advantages and disadvantages is, at the end, situation-dependent. To illustrate, as I elaborate under ‘Deterrence’, stepwise liability indeed generates greater deterrence relative to the preponderance rule. In case the preponderance rule under-deters, stepwise liability can achieve an optimal result (or at least move us in the desired direction). In case the preponderance rule over-deters, stepwise liability aggravates the problem.

Second, given the disadvantages of stepwise liability, one can think of an array mitigating steps. To illustrate, over-deterrence concerns could be mitigated by setting a lower price for passing each step. ⁷¹ Concerns for a flood of petty cases could be answered by demanding a higher bar for surviving the first steps. ⁷² Along these lines, courts could also be more aggressive in sanctioning abusive litigation. ⁷³ Accuracy concerns could be addressed by a broader right to interlocutory appeals.

Finally, the disappearing steps version seems more palatable in these respects. It is less favourable to plaintiffs, raises fewer concerns regarding over-deterrence, and, more generally, closer to the preponderance rule. Furthermore, recall that, as the case proceeds and new information arrives the judge may find at the second step that the case should even have failed the first step. Under the sunk steps regime, the plaintiff is entitled to partial compensation in this case—an uncomfortable outcome. By contrast, under disappearing steps the plaintiff in this case gains nothing. The fact that the disappearing steps option is more palatable has practical importance. One can think that judges may wish to sidestep stepwise liability in various ways, for instance, requiring a higher threshold to pass the initial phases. The disappearing steps option is less likely to be circumvented.

Against this backdrop, the most fruitful direction, in my view, is limiting the stepwise liability proposal to confined enclaves. This, in essence, is the way proportional liability is currently implemented—only in some, relatively narrow instances. ⁷⁴ What are, then, the most important contexts to use the stepwise liability regime? Areas that suffer from consistent under-enforcement, particularly due to pervasive evidentiary problems, seem appropriate for the stepwise liability proposal (as well as the proportional rule). In addition to remedying chronic evidentiary problems, stepwise liability allows quick, partial compensation and thus can be particularly suitable where victims endure financial constraints. Likewise, the proposal is beneficial where access to justice is costly, e.g., due to high filing fees, costly expert opinions, etc. ⁷⁵ Finally, stepwise liability suits cases in which damages are easy to calculate—e.g., statutory damages claims.

Practically, the presiding judge could decide, on a case-by-case basis, whether to mandate stepwise liability (somewhat similarly to the current use of the proportional rule). Alternatively, policymakers could define, in advance, the relevant areas for the implementation of stepwise liability. As stepwise liability fits the case of financially-constrained victims, one can think of ‘fast-track’ courts in specific areas, à-la small claims courts, in which victims would receive quick compensation according to their procedural progression. A possible example for such a fast-track may be medical malpractice cases. There seems to be chronic under-deterrence in this context, which is tied to plaintiffs’ inferior information. ⁷⁶ Likewise, victims of medical malpractice seem to suffer from financial constraints. On the other hand, there are reasons to believe that small claims may not be best-suited for the implementation of stepwise liability. Small claims litigation is typically short and simple, and it sometimes lacks the procedural steps that underlie the stepwise proposal. Likewise, small claims typically involve less sophisticated parties, with lower stakes. Small claims plaintiffs may be weaker at triggering the relevant steps themselves, and/or may be deterred by the additional expenses of litigating these steps. A fuller discussion of this point lies beyond the scope of this article.

Conclusion

This article proposes a thought experiment. Instead of an all-or-nothing or a continuous rule, the law can utilise a middle ground—assigning liability stepwise, according to the procedural progression of the case. The stepwise liability proposal has several notable benefits. It enhances deterrence, as it allows for partial compensation, like the proportional rule, for cases that are not sufficiently strong to win under the preponderance rule. Unlike the proportional rule, stepwise liability is embedded in the current design of civil procedure, and thus better fits the existing legal framework. Moreover, it allows for quick compensation for plaintiffs who manage to move through the initial procedural hurdles. The proposal has, of course, drawbacks. It attracts more cases into the legal system, and is expected to increase legal expenses. It might also create over-deterrence and infringe the day-in-court ideal. I suggest, then, that the proposal should be implemented in specific contexts, where a procedure that provides partial, quick compensation seems most desirable.

The foregoing has shown how a stepwise design could benefit current legal proceedings. The stepwise approach, though, could be useful in other contexts. At its core it trades off accurate decision-making for partial, quick award. It thus encourages right-holders to partially realise their rights where full realisation is risky and costly. I leave a more general account of the stepwise approach to future projects.