« Is "Fine-Grained" Always Better? | Main | Scholarship, Relevance, and Division of Labor »

Thursday, June 04, 2009

Still more on Iqbal and Twombly

The prolific folks at Drug and Device Law responded to my post, arguing why Iqbal/Twombly is the right approach to pleading. I would respond by pointing to two points on which actually, somewhat, agree.

First, they criticize Conley for allowing plaintiffs to get away with complaints that say "Defendants violated my rights" or "Defendants injured me" or claims "an FDA violation." I agree that such complaints are insufficient. For one thing, I do not believe that a proper reading of Conley allows such complaints. And while I have no doubt that such complaints are filed all the time and some may even get passed some judges, I question whether they are the norm or whether going to the opposite extreme is the solution to such abuses. For another, there was far more to the complaints in both Twombly and (especially) Iqbal and it is not clear in that sort of case what the plaintiff possibly can do to get into discovery.

Second, they close their post with the following:

Carl Sagan once said that extraordinary claims require extraordinary proof. He should have been a lawyer. In today's legal environment, given the huge expense of discovery as a means of generating the necessary proof, at the outset extraordinary claims properly demand more rigorous pleading.

Again, I probably agree. But if we set too high a pleading hurdle, there never will be an opportunity to actually prove extraordinary (or even ordinary) claims when the proof is available only through some formal discovery process from which plaintiffs are barred when they are unable to allege facts that they cannot know or prove without discovery.

Ultimately, I think we may just have to agree to disagree as to whether the system should be skewed in favor of Type II errors (meritorious claims dismissed at the outset) over Type I error (non-meritorious claims getting into discovery). That is the bottom-line policy dispute at work here.

Posted by Howard Wasserman on June 4, 2009 at 05:01 PM in Civil Procedure, Howard Wasserman | Permalink


TrackBack URL for this entry:

Listed below are links to weblogs that reference Still more on Iqbal and Twombly:


Hi Howard,

I just noted that the DDL authors counter-commented to some posts of mine. I would like to respond here. Thanks ---

Thanks to the DD&L commentators for your comments on Iqbal and your perspective.

I continue to maintain, however, that the Twombly/Iqbal "fix" comes at a higher cost than "fixing" discovery.

First, I hate to quibble, but I did not posit that all complaints must go to a jury, as you claim. The piece you quote is a newspaper op/ed. It necessarily is limited in space. So while I was not as nuanced as I would have liked, I didn't make that strong claim. In the sentence preceding the one you quote, I wrote: "Our system of justice, however, is set up to reserve such on-the-merits decisions for later stages of trials, after gathering evidence, and ultimately for juries." I was referencing summary judgment procedures to a non-law audience.

In any event, I would like to speak briefly about discovery, costs, and substantive law.

I am a former civil defense attorney, I am sympathetic to the high costs of discovery. Indeed, one study of nearly 1000 civil cases found discovery responsible for, on average, half of the total cost of litigation. See Thomas E. Willging et al., Fed. Judicial Ctr., Discovery and Disclosure Practice, Problems, and Proposals for Change 15 (1997). Of course, defense counsel "weaponize" discovery too, especially when they believe the defense can outspend the plaintiff, which is very often the case (at least in my experience). So I am not sure that weaponzied discovery is a one-way injury.

Further this big tab is not consistent across all types of suits. For example, follow up suits in mass torts cost very little in terms of discovery, because information from the lead case is reused. Simple tort cases are cheap on discovery. Patent litigation, on the other hand, tends to be expensive. Contract cases tend to have inexpensive discovery. And on and on and on.

All this is to say, a simple "discovery is too expense" response may be true --- but it is not universally true. Thus, a universal "fix" seems inappropriate. Now discovery might be too expense in certain practices, say patent law or drug mass torts, but we should prove that. Iqbal does not even pretend to make such a proof.

All these questions cry out for a solid base of statistically data that I am unqualified to provide. But even if discovery is 50% of the cost of every case, that doesn't necessarily prove that the cost it too high. That stat., includes plaintiff wins, defense wins at trial (by definition a closer case), defense wins at summary judgment, etc. If we are to do a cost-benefit analysis of tighter pleading (which is what we are up to, clearly textual analysis and precedent are not driving the bus) then we need some evidence to show that defendants spend $X in defending fishing-expedition suits (with some definition of fishing-expedition that is neutral). Then we could compare the potential lost plaintiffs' claims to that amount.

But what does seem to be clear is that higher discover costs maps on to the substantive law. That is, discovery is cheap in a breach of contract case or a car accident because the elements of the claim are easy to prove.Design defect is tough to prove and thus expensive. Even worse for discovery costs are intentional discrimination claims.

But it seems to me, that if the substantive law makes defendant liable for conduct X, and information about that conduct X is necessarily only in the possession of defendant, pre-discovery dismissals based upon the veracity of plaintiff's claim must be forgiving. There is simply no other source for the relevant information. One could well run into due process arguments, if this is not the case. It seems that the Iqbal position is as much an attack on substantive intentional discrimination law (the Court has no love for Bivens claims) as it is a procedural opinion. Perhaps intentional discrimination laws and design defect laws cost more than they are worth ... I would like to see the cost-benefit analysis, however, before making that conclusion. Assuming that they are a net benefit, then procedure should not be used to gut them sub rosa.

This does not mean that institutional defendants just have to pay. Short of purging the substantive law of such subjective elements, there are many discovery-based solutions worth considering. Many states, via their version of Rule 27, allow pre-suit discovery. More aggressive use of Rule 11 for fishing expedition suits. Moving certain claims to a Rule 9 standard. Having an initial brief discovery period, before making plaintiff pass a plausibility standard. Etc.

In short, I have yet to see the evidence that discovery is too expensive all the time and that no modification to discovery can rein in the problem.


Posted by: Lumen Mulligan | Jun 8, 2009 3:55:16 PM


I must not have conveyed my proposal very well. Iqbal and Twombly are still going to allow for a lot of cases that ultimately fail on the merits. I did not propose loser-pays for those claims.

Rather, for those claims that meet Conley but not the recently articulated standard, their claims should be dismissed, unless they consent to a loser-pays standard. This doesn't "conflate" anything, it simply shifts allocates some of the defendant's litigation expenses to the plaintiff as a risk of loss.

Applying loser-pays to the narrow set of cases where the plaintiff's success is possible but not plausible seems appropriate. I would guess that in this category, there is at least a 95% chance that the plaintiff is wasting the time and money of the defendant and the court. The plaintiff gets a chance at discovery, but only if he is willing to pay for it.

I think the fundamental question is this: Are we using limited public resources wisely when we allow people to use the judicial system to discover whether there are facts that might support completely speculative theories of wrongdoing? I think a lot of tort-reform advocates would say no. My answer is, 'Sometimes, but not often.' I don't think anyone can defend an unqualified 'Yes.'

Posted by: JP | Jun 8, 2009 11:11:27 AM

JP: The problem with your lose-pays proposal is it conflates successful pleading and an opportunity for discovery with ultimate success on the merits--those two things never have been thought to mean the same thing. As for the rest of your point, yes, the claims are based on speculation precisely because the facts (people, places, date, actions) are unknowable to the plaintiff without discovery.

Anon: I am not sure what good FOIA would be for a lot of the "speculative" facts that are going to be difficult for plaintiffs. I need to know what the supervisory officer knew about the abuse on the ground and when he knew it, which means I need to speak with witnesses--FOIA is not going to get me that. FOIA may get me a paper trail, if any. Plus, the scope of what is discoverable under FRCP 26(a) is much broader than what is accessible through FOIA.

Posted by: Howard Wasserman | Jun 7, 2009 6:32:14 AM

Skewing is the wrong way to look at it. The court is trying to strike the correct balance.

Moreover, as against the govt., it is reasonable that plaintiffs(lawyers) go through the freedom of information request process before the formal discovery process.

Posted by: anon | Jun 6, 2009 11:49:40 AM

Iqbal and Twombly were both, in essence, conspiracy theories. The allegations were grounded in speculation--not fact. An innumerable combination of facts (people, places, dates, actions) could have supported their theories. But, because the claims are speculative, the plaintiffs don't know which, if any, of those sets of facts exist, and so were unable to allege a specific set of facts. I think the Court was right to conclude that these types of pleadings don't demonstrate entitlement to relief.

In principle I'm not opposed to the use of litigation to explore conspiracy theories (particularly against the government). I would propose a new rule for plaintiffs who can state a claim under Conley, but not Iqbal and Twombly (i.e., they might be entitled to relief under some possible set of facts, but the existence of those facts is speculative or implausible). Those plaintiffs should have the option of proceeding if they are willing to pay the defendants fees if they lose.

Posted by: JP | Jun 5, 2009 11:01:21 AM

Howard, I think you're exactly right. Isn't it clear -- indeed, hasn't it been clear since the days of Charles Clark -- that there is no way to be perfectly neutral ex ante between plaintiffs and defendants, and that one *must* inevitably err on the side of letting too many cases get to discovery or of prematurely ending many meritorious suits? In short, pleading is political, and the law has tilt.

The Rules drafters made a policy choice; they believed that the greater harm occurs when claims where the evidence is most likely in the defendant's file cabinets can't be brought because the pleading rules require the plaintiff to have her ducks lined up at the outset of the case. Without that choice, many types of claims, such as antitrust and employment discrimination, would largely not be viable. It's one thing to oppose these bodies of law; it's quite another to render them inert with harsh pleading rules.

And it's something else altogether, from a legal process perspective, for the Supreme Court to reverse the presumption in Rule 8 on its own.

Posted by: Vladimir | Jun 4, 2009 5:31:59 PM

The comments to this entry are closed.