Locking the doors to discovery? Assessing the effects of Twombly and Iqbal on access to discovery.

• Author: Gelbach, Jonah B.

Many observers believe the Supreme Court's Twombly and Iqbal opinions have

curtailed access to civil justice. But previous empirical studies looking only at Rule 12(b) (6) grant

rates have failed to capture the full effect of these cases because they have not accounted for party

selection--changes in party behavior that can be expected following changes in pleading

standards. In this Note, I show how party selection can be expected to undermine the empirical

usefulness of simple grant-rate comparisons. I then use a conceptual model of party behavior

that allows me to derive an adjusted measure of Twombly/Iqbal's impact and show how to

estimate a lower bound on this measure using data from recent studies by the Federal Judicial

Center. My empirical results suggest that, depending on the nature of the s uit in question,

Twombly and Iqbal have negatively affected plaintiffs in at least 15% to 21% of cases that faced

Rule 12(b)(6) motions in the post-Iqbal data window. Again depending on the nature of the suit,

these figures represent between one-fourth and two-fifths of the cases that fail to reach discovery

on at least some claims in the post-Iqbal data window.

NOTE CONTENTS INTRODUCTION I. TWOMBLY, IOBAL, AND ACCESS TO DISCOVERY A. Pleading Standards: Rule 8(a), Conley, Twornbly, and Iqbal B. The Crucial Role of Discovery Access in Twombly and Iqbal C. Previous Empirical Literature II. METHODOLOGICAL UNDERPINNINGS: POTENTIAL OUTCOMES AND JUDICIAL BEHAVIOR EFFECTS A. Developing a Taxonomy for Cases' Pre-Discovery Posture B. Judicial Behavior Effects III. PARTY SELECTION IN AN ECONOMIC MODEL OF LITIGATION A. A Simple Economic Model of Litigation B. Extending the Model To Account for Pre-Trial Process C. Party Selection Effects in the Economic Model of Litigation IV. WHAT COMPARING GRANT RATES ACROSS PLEADING REGIMES MISSES V. WHAT CAN BE MEASURED: BOUNDS ON THE NEGATIVELY AFFECTED SHARE OF TWOMBLY/IOBAL MTD CASES A. The Negatively Affected Share Among Twombly/Iqbal MTD Cases B. Numerical Examples Illustrating the Negatively Affected Share C. The Relationship Between the Discovery-Prevented Share and the Negatively Affected Share D. Relating My Lower Bound to the Difference in Grant Rates Across Pleading Regimes VI. EMPIRICAL RESULTS A. The FJC Data 1. The FJC's Filing Data Set 2. The FJC's Grants Data Set 3. Calculating the Lower Bound B. Potential Threats to the Validity of My Empirical Results 1. Instability or Growth in the Set of Underlying Controversies 2. Are the FJC's Data Useful? VII. QUESTIONS FOR FUTURE INVESTIGATION CONCLUSION APPENDIX A: RAW CASE COUNTS APPENDIX B: ESTIMATED STANDARD ERRORS FOR THE ESTIMATED LOWER BOUNDS * INTRODUCTION

Rule 8 [of the Federal Rules of Civil Procedure] marks a notable and generous departure from the hyper-technical, code-pleading regime of a prior era, but it does not unlock the doors of discovery for a plaintiff armed with nothing more than conclusions. Ashcroft v. Iqbal, 129 S. Ct. 1937, 1950 (2009).

Bell Atlantic v. Twombly (1) and Ashcroft v. Iqbal (2) have sparked a broad debate concerning federal pleading standards. In the parallel-conduct antitrust context, Twombly overruled the Conley v. Gibson (3) standard that a complaint attacked by a Rule 12(b)(6) motion to dismiss for failure to state a claim should be dismissed only if no set of facts exists under which the complaint's claims could entitle the plaintiff to relief. (4) Iqbal explicitly broadened Twombly to "all civil actions." (5)

A number of empirical studies have appeared concerning Twombly and Iqbal's impact on motion-to-dismiss dispositions. (6) Opinions seem to be split about the implications of these studies, with some commentators suggesting that the body of evidence signals a major and ominous impact of heightened pleading's advent, (7) especially in discrimination and civil rights cases. (8) Others argue either that it is a good thing that Twombly and Iqbal elevate pleading standards, or that they do not actually change prior pleading doctrine. (9)

In March 2011, the Federal Judicial Center (FJC) released a detailed report concerning Rule 12(b)(6) practice and adjudication, hereinafter the "original FJC report." (10) The FJC released another report in November 2011, hereinafter the "updated FJC report." (11) The authors of these FJC reports characterize their results as suggesting that there is little reason to believe that Twombly and Iqbal have importantly affected Rule 12(b)(6) adjudication. (12) Others have gone further, like prominent Supreme Court litigator Andrew Pincus, who stated in congressional testimony that the original report has "proven wrong" prior "speculation" that Iqbal would greatly restrict civil justice access. (13) Not surprisingly given the stakes, critics of Twombly and Iqbal have raised objections to the original report. (14)

In this Note, I make several contributions related to the measurement of Twombly/Iqbal's effects. First, I show that simply comparing the Rule 12(b)(6) motion-to-dismiss grant rate under Conley with the Rule 12(b)(6) motion-to-dismiss grant rate under Twombly/Iqbal is of limited use in evaluating whether pleading standards have changed. An increase in pleading standards can change the number of granted Rule 12(b)(6) motions to dismiss (MTDs) through any or all of the following four channels:

(i) Judicial behavior effects: More MTDs might be granted in cases that would have MTDs filed under either pleading regime.

(ii) Defendant selection effects (15): Defendants might file MTDs in cases that they would have answered under Conley, and some of these new MTDs will be granted. (16)

(iii) Plaintiff selection effects: Plaintiffs might choose not to file some cases that they think will be either more expensive to litigate or less likely to get to discovery. Since some of the cases plaintiffs choose not to file as a result of increased pleading standards would have faced MTDs under Conley and some of those MTDs would have been granted, plaintiff selection will tend to reduce the number of granted MTDs. (17)

(iv) Settlement selection effects: The parties' perceptions of the gains and costs from litigation might either create or eliminate the possibility of settlements where settlement surplus was either negative or positive under Conley. (18) Thus, the number of settlements might rise or fall, and the number of MTDs granted might change as a result.

A critical consequence of party selection is that differences between the Conley-era and post-Twombly or post-Iqbal MTD grant rates mix together two kinds of effects: changes in judicial behavior among cases that would have MTDs filed under either pleading regime, and selection-induced compositional differences in the sets of cases that actually do face MTDs under the Conley and Twombly/Iqbal regimes. Previous empirical studies have not taken account of such potentially important changes in behavior. These studies typically subtract the observed MTD grant rate under Conley from the observed MTD grant rate under Twombly/Iqbal. But as I argue in Part W, any observed difference in grant rates across pleading regimes is consistent with large, small, or even zero judicial behavior effects. Thus, the existing empirical literature cannot settle disagreements over the effects of switching from Conley to Twombly/Iqbal.

My second contribution relates to the fact that changes in pleading standards affect litigants in ways other than through changes in judicial behavior. For example, the discussion of defendant selection above explains that if defendants file additional MTDs after a switch to the Twombly/Iqbal pleading standard, more MTDs likely will be granted as a result. This effect will harm the plaintiffs in these cases, in that the switch causes these cases not to reach discovery. Conversely, the switch in pleading standards will help defendants, for the same reason. We should want to measure such effects.

In Part V, I derive a measure of Twombly/Iqbal's impact among a particular set of cases, those in which MTDs are actually filed under Twombly/Iqbal. (19) Among these cases, switching to heightened pleading causes two types of effects. First, there are discovery-prevented cases-those that would reach discovery under Conley but do not as a result of switching to Twombly/Iqbal. Second, there are settlement-prevented cases--those that would be settled under Conley but that, because of the switch to Twombly/Iqbal, plaintiffs will file and defendants will challenge via successful MTDs. Even though these settlement-prevented cases would not reach discovery under either pleading regime, plaintiffs in these cases are worse off as a result of switching from Conley to Twombly/Iqbal, since they do not receive a settlement under Twombly/Iqbal and must also litigate MTDs that they wind up losing.

Taken together, discovery-prevented cases and settlement-prevented cases constitute the set of what I call "negatively affected cases," because these are cases whose disposition leads to worse results for the plaintiffs who bring suit. (20) The negatively affected share is the ratio of (i) the number of negatively affected cases to (ii) the number of cases that face MTDs under Twombly/Iqbal. For selection-related reasons that I explain in Part V, the negatively affected share generally cannot be calculated with real-world data. However, I show how to use available data to measure a lower bound on it. (21) This is an important contribution, because it allows me to provide meaningful empirical evidence concerning the effects on the parties of switching pleading regimes, among cases that have MTDs filed under Twombly/Iqbal.

My final contribution comes in Part VI, where I use data from the FJC's two reports to estimate this lower bound separately among three categories of cases. My results suggest that switching pleading standards affected plaintiffs negatively in a sizable share of those cases that faced MTDs in the Iqbal...