The Effects of Twombly and Iqbal

• Published date: 01 September 2017
• DOI: http://doi.org/10.1111/jels.12153
• Date: 01 September 2017

The Effects of Twombly and Iqbal

William H. J. Hubbard*

Ever since Twombly and Iqbal introduced the doctrine of plausibility pleading, a cottage

industry of legal scholars (including myself) has undertaken to detect the effects of Twombly

and Iqbal on litigants and case outcomes. Results so far have been equivocal, and it has been

hard to make sense of the disparate methodologies and findings. In this article, I develop a

comprehensive yet nontechnical framework for empirically testing the effects of Twombly

and Iqbal on lower courts and litigants, taking into account a wide range of confounding

factors and the numerous ways in which Twombly and Iqbal may have indirectly affected

litigant behavior. Using this framework, I test for effects of Twombly and Iqbal on district

court and litigant behavior using two datasets---one of administrative data covering over

700,000 cases, and one of detailed, hand- and machine-coded docket and complaint data

covering a representative sample of nearly 2,000 cases. I also review existing findings. I find

only limited evidence that Twombly and Iqbal, the two most important pleading cases in 50

years, have had a major effect on the behavior of lawyers and judges across all cases. For

represented plaintiffs, rates of dismissal with prejudice have held steady, motions to dismiss

remain uncommon, and settlement and filing patterns have not changed appreciably in the

wake of Twombly and Iqbal. There is, however, some evidence of effects---potentially major

effects---on pro se plaintiffs. Further, while case outcomes for represented plaintiffs have

been largely unaffected by Twombly and Iqbal, there is evidence that lawyers changed their

pleading and motion practice in the wake of those cases.

I. Introduction

Rule 8(a)(2) requires that a plaintiff’s complaint contain “a short and plain statement

of the claim showing the pleader is entitled to relief.” As aff‌irmed in the seminal case

Conley v. Gibson, this “notice pleading” standard required only that the pleading give the

defendant notice of the plaintiff’s grievance.

1

A line from Conley became the mantra for

this approach: “a complaint should not be dismissed for failure to state a claim unless it

*Professor of Law and Ronald H. Coase Teaching Scholar, University of Chicago Law School, 1111 E. 60th St.,

Chicago, IL 60637; email: whubbard@uchicago.edu. I am grateful for comments from participants at the Chicago

Junior Faculty Workshop, the Boston University Law School Law and Economics Workshop, the University of

Michigan Law School Law and Economics Workshop, the Conference on Empirical Legal Studies 2015, the Har-

vard Law School Faculty Workshop, the University of Virginia Law School Faculty Workshop, and two anonymous

referees. Alex Foster, Kira Fujibiyashi, Kevin Jiang, Matt Ladew, Dan Marcin, Ray Mao, Neha Nigam, Layne

Novak, Rafeh Qureshi, Adam Sudit, and Robert Zhou provided valuable research assistance. I thank the Coase-

Sandor Institute for Law & Economics and the Paul H. Leffman Fund for research support.

1

355 U.S. 41, 45 (1957).

474

Journal of Empirical Legal Studies

Volume 14, Issue 3, 474–526, September 2017

appears beyond doubt that the plaintiff can prove no set of facts in support of his claim

which would entitle him to relief.”

2

When Bell Atlantic Corp. v. Twombly

3

“retired” this language from Conley, the result

was “shockwaves through the legal community—for academics, practitioners, and judges

alike.”

4

Twombly was on its way to becoming one of the most cited cases of all time,

“notice pleading” was now “plausibility pleading,” and the academic reaction to Twombly

ref‌lected a sense that a revolution in pleading and court practice was underway.

5

Two years later, Ashcroft v. Iqbal elaborated on Twombly and reiterated the rule that

“only a complaint that states a plausible claim for relief survives a motion to dismiss.”

6

With Iqbal, the controversy over pleading standards intensif‌ied. Law review articles and

congressional testimony predicted “devastating” consequences to civil litigants, especially

employment discrimination plaintiffs, who often lack direct evidence of the defendant’s

motives at the outset of litigation and thus may be unable to plead suff‌icient factual

detail to reach discovery.

7

This raises the question: After Twombly and Iqbal, how much changed? We have

the benef‌it of a large body of empirical work inspired by Twombly and Iqbal (herein occa-

sionally, Twiqbal). What have we learned?

Surprisingly, given the strong academic reaction in the immediate wake of those

cases, the empirical evidence to date has been inconclusive. Indeed, every study of the

rates at which motions to dismiss

8

are granted with prejudice

9

has found no statistically

signif‌icant change in grant rates after either Twombly or Iqbal, even for employment dis-

crimination cases.

10

Not all the evidence indicates no effect, however. Gelbach (2012) uses data from

Cecil et al. (2011a) and f‌inds an effect in the form of higher rates at which defendants

f‌ile motions to dismiss, though courts grant the motions at the same rate before and

2

355 U.S. at 45--46.

3

550 U.S. 544 (2007).

4

Steinman (2010:1305).

5

Hoffman (2008:1235); Steinman (2010:1310); Miller (2010:28).

6

556 U.S. 662, 679 (2009).

7

Reinert (2011:123); see also Spencer (2008), Cooper (2011:960), Davis (2009), and Rubin (2009).

8

Throughout this article, “motions to dismiss” and “dismissal” refer to motions to dismiss and dismissals for fail-

ure to state a claim upon which relief can be granted. Nothing in this article speaks to motions to dismiss on

jurisdictional grounds, or for failure to join an indispensable party, for example.

9

Unless specifically noted, “dismissal” refers to dismissal with prejudice, including both (1) judicial orders foreclos-

ing the possibility of amendment and (2) dismissals with leave to amend, when the plaintiff is unable to replead

in a way that meets the pleading standard announced by the court.

10

See Section II. As I discuss below, in my own earlier study of Twombly (I did not have data on Iqbal at the time),

I was able to reach a fairly precise estimate of zero for the effect of Twombly on dismissals.

475The Effects of Twombly and Iqbal

after Twiqbal. Even here, though, the estimated lower bound of the effect (because of

the methodology, it was not possible for Gelbach to estimate an upper bound) is about

1 percent of all cases, even for employment discrimination cases.

11

The reason for this

surprisingly small lower bound is that motions to dismiss for failure to state a claim

upon which relief can be granted are rare, even after Iqbal. Cecil et al. (2011a,2011b)

found that for cases f‌iled in 2010, during the f‌irst six months of the case, a motions to

dismiss was f‌iled only 6 percent of the time.

12

In short, the literature to date provides little evidence of a seismic shift in plead-

ing practice or a differential impact on employment discrimination cases after Twiqbal.

However, the evidence does not close the door to the possibility of a large effect, either.

The literature has been plagued by serious methodological limitations, including small

sample sizes, nonrepresentative samples, failure to account for selection effects, and use

of coarse data or proxies for the true outcomes of interest. Thus, no single study has

been able to prove or disprove a substantial impact of Twombly and Iqbal on the behavior

of courts and, in turn, on the fortunes of plaintiffs and defendants.

Now is the time to take stock of the effect of Twiqbal in a more comprehensive

way. Earlier studies have each contributed important pieces of the larger puzzle, but

large gaps remain. In this article, I aim to f‌ill the major gaps in the literature and begin

to patch together a mosaic of results that allow us to see, in a more holistic way, the

extent to which Twombly and Iqbal have changed civil litigation—not just the granting

and denial of motions to dismiss, but settlement patterns, the content of pleadings, and

the willingness of plaintiffs to f‌ile suit in the f‌irst place. While the mosaic of empirical

results on this question will never be complete—there is always more work to be done—

my goal in this article is to organize enough fragments that the big picture becomes

intelligible.

At this point, I must make clear exactly what this article seeks to measure. What

does it mean to say that Twombly or Iqbal had “an effect”? Given that Twombly and Iqbal

def‌ine the standard by which a motion to dismiss is granted or denied, their only direct

effect on behavior will be on the propensity of district judges to grant motions to dis-

miss. Any other changes that emanate from these decisions will be the indirect conse-

quence of this direct effect. For example, deterrence of plaintiffs from f‌iling suits in

federal court would stem from the fact that a greater likelihood of dismissal may reduce

a plaintiff’s chances of reaching discovery—and therefore her chances of a favorable set-

tlement or trial verdict. Thus, holding all else equal,ifTwombly and Iqbal have had any

effect at all, they must have changed the tendency of district court judges to grant

motions to dismiss. Of course, all else may not be equal. If plaintiffs’ lawyers, recogniz-

ing an increased chance of dismissal, respond to Twiqbal by f‌iling fewer cases or drafting

longer and more detailed complaints, then Twiqbal will have had effects, even if we see

no rise in dismissals. As Dodson (2015) notes, strategic responses by actors in the

11

See Section II.

12

See also Willging (1989) (finding motion to dismiss filings in 6 to 12 percent of a sample of cases from 1988).

476 Hubbard