Tuesday, October 10, 2006

Abusive Removals (updated)

Trevor Morrison (Cornell) has an interesting piece on Findlaw's Writ feature previewing an empirical look at removals.  The core conclusions:

First, over the last decade, tort filings in state court are down [around four to five percent].

Second, despite the shrinking pool of state-court filings, diversity-based tort removals from state to federal court have not commensurately declined over the last decade. One would think fewer cases means fewer removals. Instead, defendants are removing nearly the same number of diversity tort cases despite shrinking state court tort dockets.

Third, cases removed from state to federal court account for an increasing proportion of the federal courts' docket. . . .

Finally, and most importantly, remand rates are increasing over time. . . . In recent years, more than 20% of diversity tort cases removed to federal court were remanded to state court. That's a substantial increase over the remand rate in the early 1990s.

It's an important point -- plaintiffs are of course not alone in abusive practices.  That said, I wonder if some portion of the increasing rate in removals has to do with what are at least questionable attempts to defeat diversity in the first place (that might be increasing).  Who really believes that in a products case the local sales representative is going to be in the case by the time it gets to trial?  Those cases will often be removed, even with a pretty strong chance of remand.  It would be interesting to see some more details on at least a sampling of the cases analyzed

But that's not to take away from the overall point of the column, which is a reminder that attempts to eliminate abusive and wasteful practices need to be party-neutral.  I'll look forward to the full work to be published in the Journal of Empirical Legal Studies.  Update: In an e-mail, Trevor points out that the article actually came out last November, and that the column came from the summer of 2005.  He reports that the full article does explore Alabama cases in a somewhat more detailed way but that the analysis is not definitive.


Legislation, Reforms, & Political News | Permalink

TrackBack URL for this entry:


Listed below are links to weblogs that reference Abusive Removals (updated):

» "The Growth of Erroneous Removal"? from PointOfLaw Forum
On Bill Childs's TortsProf Blog, I've been debating Trevor Morrison in the comments section about his paper with Theodore Eisenberg, "Overlooked in the Tort Reform Debate: The Growth of Erroneous Removal," 2 J. Empirical Leg. Stud. 551 (2005):... [Read More]

Tracked on Oct 11, 2006 5:55:46 AM


I'm pretty sure I already read it on SSRN. The paper makes no attempt to distinguish legitimately controverted removals from illegitimate ones, treats illegitimate remands identically to legitimate ones, and fails to acknowledge the extensive degree to which federal jurisdiction has been in flux over this time. The view of courts regarding fraudulent joinder, in particular, is substantially different now than it was ten years ago, and one risks being accused of malpractice not to make the attempt.

Posted by: Ted | Oct 10, 2006 12:56:21 PM

I'll have to look on SSRN to see if it's up there when I get a chance - thanks. As my comments in the post suggest, I have my suspicions that some portion of the removals and remands are based on facially questionable efforts to avoid diversity jurisdiction. (I've been on the other side of some laughable such attempts, including in Alabama.)

The question, indeed, is whether the anecdotal evidence (of clearly problematic removals) matches up with the broader empirical evidence in more than just a surface way. Given the journal in which the article is being published, I'm hopeful that the final article will address that question.

Posted by: Bill Childs | Oct 10, 2006 2:10:05 PM

I don't doubt that there are illegitimate removals.

At the same time, I was on a case where we removed twice, and both removals were legitimate: the first time, the court remanded because the plaintiffs didn't specifically say they were asking for more than $75,000, which is the wrong standard unless the plaintiffs explicitly waive that right. (This would have been irrelevant if the court had correctly anticipated the Supreme Court's ruling in Allapattah.) Sure enough, one year passed and they amended their complaint to seek punitive damages over the jurisdictional limit, just as we warned the court would happen. Our second removal on the new facts was remanded because the court took the position that the Eighth Circuit hadn't yet explicitly adopted the Fifth Circuit's fraudulent-complaint exception to the one-year limit. (How the Eighth Circuit is supposed to ever get around to doing that when a remand is unappealable is beyond me.) At least once the removal should have stuck.

I've been very disappointed in empirical legal work. They usually consist of a study that performs technically accurate counting of statistics, and then wild jumps to conclusions that coincidentally correspond to the authors' biases without acknowledgement of the limits of the data. For example, ending the complete diversity requirement and permitting any out-of-state defendant to remove with or without consent would virtually end wrongful removal and the problem of fraudulent joinder (while at the same time being more consistent with the original intent, see Federalist No. 80), yet that option is ignored in this paper in favor of one-sided fee-shifting for remands--despite the lack of any evidence that the Seventh Circuit's adoption of a per se fee-shifting rule had any effect on remand rates. But perhaps this is a problem with all academic empirical work, and I'm only noticing it in the legal setting.

Posted by: Ted | Oct 10, 2006 3:03:41 PM


It's not clear from your comment whether you've actually read the full article that Ted Eisenberg and I wrote (as opposed to the short piece for Findlaw, which really isn't a suitable forum for comprehensive academic research). I'm assuming you haven't, because otherwise I don't think you'd suggest that we had failed to acknowledge the limits of our data, or that we had hastily jumped to any firm policy recommendations.

In fact, we are quite careful in the article to discuss the limitations of our data, and we avoid drawing any definitive conclusions about the reason for the rise in erroneous removal. We note that it might be on account of abuse, but we also note that it might be on account of legitimate efforts by defense counsel to respond to genuine legal or factual uncertainty. And we explicitly say that our data can't tell us which is right.

As for policy recommendations, we mention the possibility of mandatory fee shifting at the very end of the piece as a possibility that Congress might want to consider. We certainly don't "jump" to this option as necessarily best. Indeed, the main aims of the piece are descriptive, not normative. We seek to identify a previously under-recognized systemic phenomenon. What, if anything, to do about that phenomenon is evidently something about which you have strong feelings, but it's not the focus of our article.

Posted by: Trevor Morrison | Oct 10, 2006 5:55:40 PM

Professor Morrison: I've read your article, and discussed it with Professor Eisenberg when we appeared on a panel together in May; Professor Eisenberg made me a generous offer to provide me the underlying data that I have not yet had the opportunity to follow up upon.

There are a number of places where the article makes sweeping claims unsupported by the data.

For example, your article concludes, and I quote, "Substantial empirical evidence establishes that, especially among state cases removed on diversity grounds, erroneous removal is a significant and growing phenomenon." Your empirical evidence doesn't establish this for the reasons I state, unless one defines "erroneous removal" to mean something other than erroneous removal. Given the existence of the word "remand" to accurately convey what your data is actually counting, it's hard to see how you're not asking the reader to draw conclusions by using the non-neutral adjective "erroneous."

Your paper also says: "That is, the high rate of erroneous removal in Alabama may be best attributed to knowingly wrongful removal, not simply to honest uncertainty about the law." Again, the data doesn't support this, and, though your data permits an alternative test that would be sufficient (if not necessary) to demonstrate that the problem is one of legal uncertainty, it doesn't test it. (It was this test I discussed with Professor Eisenberg.) It was an offensive statement that could have been avoided had the authors spoken to any attorney who has done at least a dozen removal motions and accounted for what's actually happening on the ground.

Not once does the paper recognize the limitation on your data that remands are frequently erroneous (and thus characterizing a remand as an "erroneous removal" begs the question); not once does the paper recognize the growing problem of fraudulent joinder and the controversy in the federal courts on how to address it, much less the other half-dozen-or-so controverted questions in federal jurisdiction that the Supreme Court has addressed. The one policy recommendation the paper makes, for fee-shifting, was possible to be tested for efficacy, yet it wasn't tested; other policy possibilities (such as giving courts more explicit power to scrutinize fraudulent joinder or re-establishing minimum diversity to simplify the federal jurisdiction question) are ignored entirely with the exception of language in the CAFA Senate Report that was mocked without addressing the underlying argument. So I stand by my earlier disappointment that the paper is so unjustifiably anti-reform, and that its authors have been using data that doesn't support their conclusions to argue against needed reforms and bad-mouth defense counsel.

Posted by: Ted | Oct 11, 2006 5:48:42 AM

Well, Ted, I guess I'm glad you've read the paper, but (unsurprisingly, I suppose) it looks like we're going to have to disagree as to its merits. Two responses, and then I'll relent:

1. Your first criticism, if I understand it correctly, is that we shouldn't use the term "erroneous" to refer to removals that are deemed incorrect by the district court and then remanded. Your point, I take it, is that you or others would deem some of the remands themselves to be erroneous, based on your own analysis of the cases.

I agree that we do indeed use "erroneous" as you describe -- that is, we treat a remand as dispositive evidence that the removal was erroneous. But I guess I don't view that as a problematic terminological choice. As you know, the system as it is currently devised treats district court remand decisions as essentially unreviewable. So the district court has the last word. Our terminology defers to that last word. If a district court deems a removal to have been in error, the system as it's currently constructed deems it to be in error, and so we adopt that determination in our terminology.

Now, you and I may or may not agree with particular remand decisions, but that's really a separate point. The system as it's currently constructed has a mechanism for reviewing removals and correcting those that the final decision maker determines were wrong. By calling those removals erroneous, we're simply taking the system's outputs as they are.

Justice Jackson once famously quipped that the Supreme Court is not final because it is infallible; it is infallible because it is final. Finality, in other words, makes a tribunal's determination decisive. You or I can question the "correctness" of that determination from our points of view, but to do that would be to import our own substantive views and ideological predispositions. That's not what Ted Eisenberg and I tried to do. Instead, we use "erroneous removal" simply to refer to removals that the system has finally determined were in error.

Note also that the *only* removals we deem erroneous are those that are remanded. If we were to make our own determination about the propriety of every removal according to our view of the law, we might well fairly determine that the district court erred in some cases by *not* remanding the case. But we don't count those cases as involving "erroneous removals" for the same basic reason: the system has not deemed them erroneous.

Again, I am not trying to argue that it would be illegitimate to sift through every removal in the country and determine for oneself whether it was erroneous according to one's own view of the law. That would be a perfectly fine project, I'm sure. But that wasn't our project, and I just can't see how it was illegitimate or offensive for us to use "erroneous removal" to denote those removals, and only those removals, finally deemed erroneous by the system as it is currently structured.

2. Your other criticism is about what we say about Alabama. You quote one possible explanation that we offer, which is that the very high rate of erroneous removal in Alabama may be due to defense side abuse. But then you fail to quote the other explanation we offer, which is that changes in the available information between the time of removal and the time of remand may help explain the difference in error rates we're discussing in that part of the paper. In other words, we ourselves posit different possible explanations -- one of which would involve defense side abuse and one of which would not -- and we say that our data as we've analyzed it don't tell us which explanation is correct.

Again, you seem quite certain about the correct explanation, and I have no doubt that you are sincere in your beliefs. Undoubtedly, participation in "at least a dozen removal motions" would give one an anecdotal basis for thinking one knows "what's actually happening on the ground." But others with comparable experience on the plaintiff side have different views, and would say that abusive removal goes on all the time in Alabama and elsewhere. We don't reach a conclusion about which side is right, because we don't think our data reliably permit such a conclusion. You, on the other hand, evidently believe you know which side is right. Is there a chance that this belief is the product of your own biases?

Posted by: Trevor Morrison | Oct 11, 2006 6:59:00 AM

Professor Morrison,

I suggest that if one starts from the premise that the current system correctly decides jurisdictional questions, one is likely to come to a different conclusion about where problems reside than someone who is seeking to analyze whether the current system is correctly deciding jurisdictional questions. Much like other Eisenberg papers, the question-framing assumes away the complained-about problem, and then makes announcements that reformers are wrong. It would be one thing if the errors of this sort were randomly distributed, but it starts to be suspicious when they all go in one direction. It would be posible to completely rewrite this paper using the same data and charts to argue that reformers are right about a variety of issues; instead, one sees only the opposite hypotheses for interpretation of the data, and, at best, the occasional "Well, it's possible there's a harmless explanation."

"Erroneous removal" is a term your paper invented, and you admit that it's a synonym for "remand," an admission omitted from the paper itself, permitting the equivocation that readers would interpret it in the pejorative sense. I find your explanation of why your paper required the use of a misleading neologism instead of the standard legal terminology less than persuasive.

I don't claim to know which side is right (though I have my guesses); I just know that the broader anti-reform statements made in your paper (and in your op-ed) do not ineluctably follow from the data collected in that paper. I have two sets of hypotheses consistent with my empirical experience that I look forward to testing early next year.

Posted by: Ted | Oct 11, 2006 3:13:11 PM

To clarify, I don't object to the idea of performing a study that simply analyzes remand rates. I object to drawing conclusions from that study that its original circumscribing no longer permit it to draw.

Posted by: Ted | Oct 12, 2006 7:11:12 AM

Ted, I had planned my previous comment to be my last in this thread, but I'm afraid that your latest comment contains some serious accusations and inaccuracies that demand response. So I'll make these my last words in this thread:

1. Unless I'm misreading your comment (and I hope I am), you appear now to be accusing Ted Eisenberg and me of academic dishonesty. You're suggesting, I think, that our paper (and others that Ted has written on his own or with other co-authors) intentionally employs a misleading research design and misleading terminology, in order to fool the reader into buying some spurious claims. I don't know about the world of Washington thinktanks, but in the academic world that's a serious accusation. You offer no proof to support the accusation, nor could you: it is categorically false, and I deny it categorically. Surely it's possible to debate ideas without hurling around this kind of unsupported invective.

2. Now to more specific inaccuracies in your comment. You say that "'[e]rroneous removal' is a term [my] paper invented, and [I] admit that it's a synonym for 'remand,' an admission omitted from the paper itself, permitting the equivocation that readers would interpret it in the pejorative sense." I have a number of responses here.

First, I certainly don't think of our paper as having invented the term "erroneous removal." There must surely be many, many district court remand opinions that characterize the removals they're reversing as "in error" or "erroneous." I can't believe you think otherwise. As I have already explained, we treat as erroneous all those removals -- and only those removals -- deemed erroneous by the federal courts. So again, far from inventing a category or a term, we're just taking the federal judiciary's work product as it is.

Second, you say that I "admit" in this thread that we equate "erroneous removal" with "removal reversed by a remand." I may be misreading you, but it seems you've chosen the word "admit" here to suggest that I'm recognizing some failing or culpability in our terminological choice, or at least that it was some sort of secret whose cover has now been blown. But that's not so. It's true that we equate removal error with removal followed by remand, but I don't think of that as an "admission" or concession because, as I say, I think there's nothing wrong with our treatment.

Third, and most seriously, you say that our paper does not include the above "admission," which you say "permit[s] the equivocation that readers would interpret it in the pejorative sense." The premise of this claim is totally false. Again, although I don't think of it as an "admission," the fact that we use "erroneous removals" to refer to removals reversed by remands is plain from our paper. It is plain, for example, on page 556, when we first summarize our findings:

"First, the annuanl number of cases removed to federal court on the basis of diversity generally increased between 1979 and 2002. In addition, the proportion of the federal diversity docket that originated in federal court by way of removal rose steadily from 1979 to 2003. Finally, the percentage of removals ultimately remanded to state court increased significantly over this same time period. That is, just as more cases purporting to satisfy diversity jurisdiction were being removed to federal court, and just as removals were occupying an ever-larger part of teh federal docket, more removed cases were being remanded to state court. In short, erroneous removal is a growing phenomenon."

Note that in this summary, the term erroneous removal does not appear until the final sentence, which (as the words "in short" make clear) gives a shorthand restatement of the findings spelled out in the previous several sentences. And as those previous several sentences make clear, the only "error" we're talking about when we say "erroneous removal" is a removal later reversed by a remand. This same sentence pattern appears in the abstract of the piece. Thus, I'm confident that any reasonable reader who actually read the paper (or even only the abstract) would have immediately grasped what we mean by "erroneous removal." The paper is perfectly clear on this point, and your accusation that we somehow hid this fact in order to lead readers to think we mean something more "pejorative" by the term is inaccurate, unfounded, and inappropriate.

3. The next sentence in your comment characterizes my earlier comment as having claimed that "[our] paper required the use of a misleading neologism." I've already addressed the "misleading neologism" point -- I'm sure it's not a neologism, and I'm even more sure that, as we've deployed it, it is not misleading. I won't say more on that score. Instead I want to clarify that I have never claimed that our paper "required" that we use the term "erroneous removal" as we have.

I think my previous comment was quite clear on this point, but let me clarify just in case: Although I think it is perfectly legitimate to measure removal error as we have (and to refer to it as such), there are undoubtedly other legitimate ways to assess removal error as well. For example, as I noted in my previous comment, one could go through every removal in the country and assess for oneself whether each one satisfies one's own understanding of the criteria for removal. (Note that, to be evenhanded, that sort of study should not just look at every remand, but should look at every removal since in some cases the district court's error might lie in its refusal to remand.) That too would be a legitimate research project. But it's not ours. I have not claimed that we were "required" to choose the research project we chose, only that it was perfectly legitimate for us to make the choice.

To underscore the legitimacy of our choice, let me offer an analogy. I think it would be perfectly legitimate for someone to undertake a study called "An Empirical Study of Unconstitutional State Laws Passed in the Last 20 Years," where the study equated "unconstitutional" with "held unconstitutional by the United States Supreme Court." Is this the only possible understanding of "unconstitutional"? Of course not. Another study might look at all state laws held unconstitutional by state supreme courts or lower federal courts. And another might make assessments of unconstitutionality depend on the author's own substantive views about what is constitutional and what is not, irrespective of what the courts say. Each of those studies would implement a permissible definition of unconstitutionality. The same is true in the case of removal and remand. Our treatment of "erroneous removal" parallels my hypothetical study's treatment of "unconstitutionality": it equates it with final judgments made by the highest level of the federal judiciary authorized to make such judgments. That's not the only possible treatment, but surely it's a legitimate one.

Moreover, and critically, our treatment is a reasonably neutral one. Plaintiffs' lawyers would undoubtedly like us to include cases that they think should have been remanded, but that district courts "erroneously" refused to remand. You and other defense-side folks would evidently like us to exclude remands that you think were erroneous. Our measure does neither, and has the virtue of not skewing to either your or the plaintiffs' view of the underlying law.

4. Lastly, a few quick words about the following sentence in your comment: "Much like other Eisenberg papers, the question-framing assumes away the complained-about problem, and then makes announcements that reformers are wrong." I won't take time here to defend Ted Eisenberg against your criticism of his other work, as he hardly needs defending. Ted is widely recognized by academics from across the political spectrum as one of the very best, and most careful, empirical legal scholars in the business. Instead of talking more about his fine work, I want to note two things.

First, your use in this sentence and elsewhere of the "reform" label is quite ironic. While criticizing us for using terms that you say are misleading, you persist in using a term that everyone recognizes as loaded. Distinterested third parties recognize "tort reform" as a euphemism for "changes in the law that are helpful to tort and other civil defendants." You yourself appear to use the term that way. Thus, you treat our brief mention of the possibility of mandatory fee-shifting as evidence that we are "unjustifiably anti-reform." A neutral definition of "reform" would treat the term as encompassing *any* proposed change in the law whose proponents think it would improve the law. Individual commentators could then argue that some reform proposals were better or worse than others, but the label "reform" wouldn't be restricted only to reforms favored by a particular side. Under this more neutral understanding of "reform," a proposal to adopt mandatory fee-shifting would qualify as a proposed "reform." You don't appear to use the term this way, though. Instead, you treat even our modest, brief support for *considering* mandatory fee-shifting as evidence that we are "anti-reform." In this way, you perpetuate a loaded, non-neutral usage of the reform label.

Second, it's simply inaccurate to describe our paper as announcing that "reformers are wrong," if by that you mean that we oppose various other "tort reform" proposals favored by you and others on your side. The thrust of our paper is that the increase in removal error has been largely overlooked up to now, and that it merits consideration by those contemplating other reforms in this area of civil litigation. We don't argue that other reforms are necessarily wrong or unnecessary; we argue that this problem deserves consideration too.

I've gone on too long, but the inaccuracies in your last comment (some of which, I'm afraid, are really quite inflammatory) are sufficiently troubling that I thought I needed to set the record straight. That said, I will rest here and provide you the last word if you want it, hoping that word will look more like responsible debate than overheated rhetoric.

Posted by: Trevor Morrison | Oct 12, 2006 8:37:57 AM

Morrison's response focuses mostly on my use of the phrase "misleading neologism." He says that this is equivalent to an accusation of academic dishonesty, an accusation I did not make. His rebuttal mostly hinges on the claim that "erroneous removal" is not a neologism; if true, then my criticism has been unfair and unjustified. But Morrison is making an empirical claim that is readily tested.

If one googles the phrase "erroneous removal," as of October 15, 2006, one finds 520 total entries on Google (compared to over 9000 for the phrase "fraudulent joinder", 49 thousand for the phrase "complete diversity", 1.5 million for the phrase "federal jurisdiction", 1.7 million for the phrase "tort reform", and 2.4 million for "remand +jurisdiction"). Six of the top seven are references to the Eisenberg/Morrison paper, and the seventh has nothing to do with federal jurisdiction. Out of the top twenty Google search results for the phrase, eight are references to the Eisenberg/Morrison paper, and twelve have nothing to do with federal jurisdiction.

The phrase appears precisely once in Supreme Court opinions in the last twenty years, in Caterpillar v. Lewis; the Supreme Court used it in the context of a case that was removed but should not have been removed, though it was never remanded. In other words, in a different sense than Eisenberg/Morrison use it, and closer to the sense I argue "erroneous removal" is better understood.

Only one paper on SSRN uses the phrase "erroneous removal" in a title or abstract: Eisenberg/Morrison.

A grand total of thirty documents in the Westlaw JLR database use the phrase "erroneous removal." Looking at the twenty most recent entries, four are citing to Eisenberg/Morrison; four (including one that cites Eisenberg/Morrison) use it in the same sense as or in quoting Caterpillar v. Lewis; eleven are not discussing federal jurisdiction at all but rather immigration or family law; one uses the phrase "patently erroneous removal" to distinguish it from an everyday controverted remand; and one, a 2003 Marquette L. Rev. article written by a Sidley & Austin junior associate, uses it in precisely the same sense as Eisenberg/Morrison.

Given that "erroneous removal" is a synonym for "remand" in Eisenberg/Morrison, and given that the only use of the phrase in the Supreme Court has a different meaning in the federal-jurisdiction context, I stand by my argument that the term is a neologism or, at the very least, an uncommon way to rephrase the simpler standard legal terminology "remand."

The paper omits numerous arguments for why the data support reform proposals; the paper may not be "anti-reform" in the sense of a position paper, but it certainly exhibits an anti-reform confirmation bias.

I apologize if you find my statements inflammatory; it was not my intent to be inflammatory; if it was inflammatory, it was considerably less inflammatory than your accusations of me of not reading your paper, that my arguments are the product of biases, and your implicit (and Professor Eisenberg's explicit) ad hominem. I've analyzed your arguments and your paper on their own terms. Please show me the same courtesy.


Separately, Professor Morrison is correct that the word "reform" has veered from its etymological roots when used to refer to liability reform and that I am using it in the descriptivist sense that a prescriptivist might plausibly have objected to thirty years ago. (Similarly, "anti-Semitism" refers to prejudice against Jews, rather than prejudice against Semites.) If noone before me used reform in the sense that I use it here, and if "reform" were not regularly used by both sides of the debate to describe measures intended to increase the accuracy and reduce the scope and expense of the liability system, his argument would be analogous to mine.


Separately, this isn't the place to do a point-by-point complaint of each of Ted Eisenberg's papers, but I refer one, for example, to Professor George Priest's refutation of the Eisenberg/Miller paper on class-action fees, where the problems of magnet jurisdictions and coupon settlements were effectively assumed away, and the results then used to argue against the adoption of the Class Action Fairness Act.

Posted by: Ted | Oct 15, 2006 8:55:49 AM

Cf. also Santamarina v. Sears, Roebuck & Co., No. 06-3054 (7th Cir. Oct. 19, 2006), which, without using the term "erroneous removal" refers to other situations where a case can be erroneously removed, but not remanded.

Posted by: Ted | Oct 23, 2006 3:30:37 AM

Post a comment