Table of Contents Introduction I. Twombly, Iqbal, and What We Can Learn from Summary Judgment Results A. The Evolution of Pleading Standards and How They Relate to Discovery B. Defining "Merit" Postdiscovery C. Understanding the Summary Judgment Link when Only Judges' Behavior Changes Following Twombly and Iqbal D. Party Selection Effects of Twombly and Iqbal 1. Party selection effects that involve filtering out at the Rule 12(b)(6) stage 2. A party selection effect that involves selection into summary judgment II. Data III. Employment Discrimination Cases A. Basic Characteristics of the Employment Discrimination Sample B. Summary Judgment Adjudication Results for Employment Discrimination Cases 1. The percentage of cases in which plaintiffs win as to all challenged claims or issues 2. The percentage of cases in which plaintiffs win on at least one challenged claim 3. Measuring the quality-filtering effect using changes in the number of employment discrimination cases facing summary judgment motions IV. Contract Cases A. Basic Characteristics of the Contracts Sample B. Summary Judgment Adjudication Results for Contract Cases 1. The percentage of cases in which plaintiffs win as to all aspects raised by defense summary judgment motions 2. The percentage of cases in which plaintiffs win on at least one claim challenged via defense summary judgment motion 3. Challenges in measuring the quality-filtering effect using changes in the number of contract cases facing summary judgment motions V. Further Issues A. Why It Would Be Problematic to Focus Only on Cases in Which a Rule 12(b)(6) Motion Was Filed B. Threats to the Validity of the Empirical Approach 1. The recession 2. Ledbetter v. Goodyear Tire & Rubber Co. and the Lilly Ledbetter Fair Pay Act 3. Scott v. Harris 4. Amendments to Rule 56 5. Changes in primary behavior 6. Other factors that might lead to invalid estimates Conclusion Introduction
Much controversy has surrounded the Supreme Court's opinions in Bell Atlantic Corp. v. Twombly (1) and Ashcroft v. Iqbal. (2) Critics of the plausibility standard introduced through these cases argue that it will reduce access to the federal courts for meritorious suits. (3) In some disputes, the critics argue, the defendant controls the information that would be necessary to plead in sufficient factual detail to meet the plausibility standard. Under Conley v. Gibsons now-retired "no set of facts" pleading standard, (4) plaintiffs could allege wrongdoing generally and then rely on discovery to unearth the facts necessary to establish the elements of such causes of action. By requiring plaintiffs to allege such facts before Rule 12(b)(6) adjudication, (5) critics argue, Twiqbal's plausibility standard establishes a catch-22: pleading sufficiently to reach discovery requires access to information that is available only through discovery. (6)
On the other side of the "Twiqbal" debate, supporters of the plausibility standard argue that too many plaintiffs intentionally bring low-merit lawsuits for settlement value only. According to this view, defendants must agree to pay off plaintiffs in such cases because the burden of discovery is greater for defendants than for plaintiffs. Thus, plaintiffs in such suits have little disincentive to proceed through discovery, leaving defendants to choose between either settling beforehand or bearing the high discovery costs that precede the summary judgment phase where the defendants are very likely to win. Supporters of the switch to Twiqbal's plausibility standard believe it will help eliminate low-merit cases, whose plaintiffs, they believe, will be unable to plead with sufficient factual detail before discovery costs mount.
Both the critics and the supporters hold theoretically coherent views: logically, both views could be correct. Further, it is possible that there are sizable numbers of both meritorious cases likely to face a catch-22 problem under the plausibility standard and strike suits likely to be filtered out by the plausibility standard. Therefore, the question as to which effect predominates ultimately is an empirical one.
While there has been no shortage of empirical work on Twiqbal, it has overwhelmingly focused on the question whether judges have indeed applied a higher standard when adjudicating Rule 12(b)(6) motions. (7) But almost no work besides the present Article has even attempted to measure the extent to which Twiqbal has affected the merit of cases that proceed beyond the Rule 12(b)(6) stage. (8)
This Article takes up that task, using outcomes of defense summary judgment motions to attempt to measure Twiqbal's effects on case quality. More specifically, the analysis defines a case as having no merit postdiscovery if a reasonable factfinder could not find for the plaintiff on the basis of the postdiscovery record. (9) Defendants in such cases can be expected to file and win summary judgment motions. Assuming that the plausibility pleading standard is effective at snagging low-merit cases at the Rule 12(b)(6) stage, such no-merit cases will be eliminated from the set of cases that make it to summary judgment in the post-Iqbal period. Thus, (i) there should be fewer cases in the summary judgment population post-Iqbal, and (ii) those cases should be more likely to survive summary judgment. Consequently, the plaintiff's win rate against defense summary judgment motions should be greater following Twiqbal than before. On the other hand, if Twiqbal's critics are correct, then the plaintiff's win rate will have either dropped (indicating that Twiqbal filters out higher-quality, rather than lower-quality, cases) or stayed the same (indicating that Twiqbal filters equal proportions of high- and low-quality cases).
In reality, things are more complicated. There are good reasons to believe that parties' case strategy and Rule 12(b)(6)-motion-filing behavior, as well as their inclination to settle, will change following a perceived change in the pleading standard. (10) Accordingly, Part I.D also discusses the role such party selection effects play in the empirical analysis.
A critical question that arises in this discussion is whether there are any disputes that "select into summary judgment," in the sense that they would develop into lawsuits that ultimately face summary judgment motions under Twiqbal, but not under the pre- Twiqbal pleading standard. While counterintuitive, the possibility that some such lawsuits exist is real. (11) This possibility is important because the empirical approach adopted in this Article to measure the quality-filtering effect of Twiqbal requires the assumption that there are no such cases--an underlying premise subsequently referred to as the "no-selection-into-summary-judgment" assumption. This assumption is strong in that it amounts to a restriction on the scope of party selection effects. Without the assumption, differences in the plaintiff's win rate at summary judgment involve not only the quality-filtering effect discussed above, but also a component related to the quality of cases that are selected into summary judgment as a result of Twiqbal.
If there were such a mixture of quality filtering among some cases and selection among others, it would be impossible to use empirical evidence to isolate Twiqbal's quality-filtering effects. For example, a finding of no change in the plaintiff's win rate at summary judgment would be consistent with either no quality-filtering effect and no selection effect or a large quality-filtering effect that is exactly counteracted by a large selection effect operating in the opposite direction. In such a situation, neither possibility can be ruled out. Consequently, it would be impossible to assess commentators' claims about Twiqbal's quality-filtering effects in particular. Thus, the no-selection-into-summary-judgment assumption is necessary to assess commentators' claims about Twiqbal's effects on the mix of cases that make it past the Rule 12(b)(6) stage. (12)
Finally, as to the no-selection-into-summary-judgment assumption, I note that it is partially testable. If this assumption is correct, then Twiqbal would have eliminated some cases from the set facing summary judgment motions while not causing any new cases to face summary judgment motions. Thus, Twiqbal would have to reduce the share of filed cases that face summary judgment motions, all else being equal. If the no-selection-into-summary- judgment assumption is correct, then there should have been a drop in the share of cases facing summary judgment motions following Twiqbal.
The empirical work in this Article is based on a random sample of cases in which defendants filed Rule 56 summary judgment motions. This sample was created using a unique database of docket reports. (13) To construct the dataset, all civil cases filed in the federal district courts in the periods of October 1, 2005 to June 30, 2006 (the pre- Twombly period) and October 1, 2009 to June 30, 2010 (the post-Iqbal period) were combed to find those which had at least one motion for summary judgment filed. (14) After restricting attention to cases with a PACER code indicating that the nature of the suit...