Over his 17 years as a magistrate judge for the U.S. District Court for the Southern District of New York, Judge Frank Maas earned a nationwide reputation as a leading jurist in the area of eDiscovery law, a tech-savvy judge unafraid to dive into the intricacies of the discovery process. Following his retirement from the bench last September, Judge Maas has continued his leadership in the area, frequently speaking on discovery issues.
Logikcull’s Robert Hilson recently sat down with Judge Maas to discuss everything from the costs of discovery, to the role of judges in the discovery process, courts’ inherent authority to sanction discovery wrongdoing, and his approach to new discovery technology. A transcript of our discussion follows, lightly edited for clarity and concision.
Logikcull: It is pretty well-known that the number of trials in federal courts has been declining and continues to decline. I think I saw a stat somewhere that said less than two percent of all cases end up going to trial. So, this is a two-part question.
The first part: I think there is now an assumption that discovery is the de-facto dispute resolution mechanism in federal courts. Do you agree with that?
The second question: As technology proliferates, and as the information related to some of these disputes proliferates, to what extent do you think our discovery system is capable of withstanding this technological onslaught?
Hon. Frank Maas: Those are good questions. As to the first of them, in the federal system, the number of trials is down to around one percent of cases filed. That is unfortunate. If you put aside cost for the moment, I think our discovery system does enable parties to gather the information (sometimes, perhaps, too much information) necessary to be able to make intelligent decisions about the disposition of their cases, either through motion practice or settlement.
Faced with the dearth of trials, the increased use of techniques such as phased discovery could help ensure that the cost of discovery does not become disproportionate to the amount in controversy in a case. In my view, the number of cases in which discovery is disproportionate is actually small. I’m sure there are cases in which the cost of discovery makes it difficult — perhaps even impossible — to prosecute or defend a case successfully. But, even in a jurisdiction like the Southern District of New York, where I sat, most cases are relatively small, and don’t involve grossly disproportionate discovery. When discovery requests are abusive, judges can and, of course, should step in to restore some rationality. On balance, however, I think the discovery system works reasonably well for most litigants most of the time.
“When discovery requests are abusive, judges can and, of course, should step in to restore some rationality. On balance, however, I think the discovery system works reasonably well for most litigants most of the time.”
Logikcull: To your point about judges stepping in when there are disproportionate discovery requests, there was a small change to Rule 1 that basically suggested that judges should have more of a hands-on approach in terms of case management and so forth. Have you seen anything since those changes went through to indicate that judges are in fact taking a more hands-on approach to making sure that discovery is proportional to what it’s seeking?
Maas: Well, the Rule 1 change was really to emphasize that the parties also have a role in ensuring that each dispute is resolved in a just, speedy, and inexpensive way. There were other changes to the Federal Rules, such as the changes to Rule 16, intended to ensure that the date when a judge becomes involved in a case moves up. I do think judges are becoming involved sooner in their cases, and that there’s an increased focus on proportionality among not only the bar but also the bench as well.
“I do think judges are becoming involved sooner in their cases, and that there’s an increased focus on proportionality among not only the bar but also the bench as well.”
Logikcull: The changes to Rule 26 — a lot of people have said proportionality was obviously already in the rule, which it was, but it’s now been moved up. So, you think that has affected the way that judges and parties consider these issues?
Maas: I sat on a special committee of the New York State Bar Association a number of years ago which was tasked with making recommendations for improvements in federal practice. It overlapped somewhat with the period in which the rules changes were being considered, but the committee had a slightly different focus. The membership of the committee, in addition to me as a judge, included litigators from every walk of life – big firms, small firms, plaintiff’s-side, and defense-side. Much to my amazement, the recommendation about which the committee members had the greatest degree of agreement was that judges needed to become involved in cases early on. They therefore wanted to cut, as the amendment eventually did, the amount of time it took before you got before a judge to talk about things like discovery.
In that regard, I frankly didn’t view moving the proportionality standard from Rule 26(b)(2)(C)(iii) to Rule 26(b)(1) as a terribly significant change. But it did have the effect of once again focusing judges’ and parties’ attention on the need for discovery to be proportional. So, the benefit was that it reminded everyone that the requirement of proportionality is a very important aspect of the rules. Does that answer your question?
Logikcull: Absolutely. We actually spoke to Judge Facciola, probably 18 months ago at this point. One of the things that he was saying was that the cost of the federal cost system — specifically he was talking about eDiscovery — in his words has “driven an entire economic class out of federal court.” There’s often this perception, whether right or wrong, that there are problems on both sides. For plaintiffs, it can be the case that the cost of discovery can be so overwhelming that it might act as some kind of “chilling effect” where they might not even bring their claims in the first place for fear that they’re not going to be able to bring their dispute through to resolution. On the other side, for producing parties, it can be the case on occasion that the cost of discovery is so much that it might make sense just to settle the claims regardless of their legitimacy. Do you generally agree with that perception? Is that accurate? Is that unfounded? And which way do you see the trend going?
Maas: Well, Judge Facciola is very passionate about this issue, and I have heard him speak about it. There certainly are cases in which he’s correct that the cost of discovery can be a significant factor in the parties’ decision making.
But the cost of discovery has always been a cost of doing business in this country. To a certain extent, obviously, it’s become more expensive with the advent of electronic documents and electronic discovery. Still, I don’t think corporations are inclined to go back to conducting their business with paper files. So, it’s just something we’re faced with.
When I speak to industry groups, I advocate against retaining data that serves no continuing business purpose. Obviously, if litigation is reasonably anticipated or there is some regulatory requirement, information has to be preserved. But if not, corporations should regularly be purging their data files pursuant to some clearly-established policy. Unfortunately, most corporations, like most people (certainly me), are pack-rats. The more you have available to produce, the worst the problem becomes.
“When I speak to industry groups, I advocate against retaining data that serves no continuing business purpose… The more you have available to produce, the worst the problem becomes.”
Over time, the narrower scope of discovery in amended Rule 26(b)(1), which restricts discovery to matters relevant to claims or defenses, should help address the problem. When that change first was proposed, I didn’t think it was tremendously significant because, in 17 years on the bench, I never heard anybody object that the discovery that somebody was seeking as of right was related to the general subject matter of the case, rather than a specific claim or defense, and therefore couldn’t be obtained without the court saying that that discovery should go forth. That just never happened. I also asked a lot of my fellow judges, and none of them ever had somebody make that argument before them. But I do think that, over time, the defense bar will be raising more of those objections. If they’re granted, and clearly that was the intent of the drafters of the revised rule, that should help to some extent in narrowing the scope, and therefore the burden, of discovery.
Logikcull: Do you have any thoughts about whether eDiscovery costs should be awarded to prevailing parties?
Maas: Well, the only thing I would say is that as of when I left the bench, the case law was all over the place on that.
Logikcull: And it still is.
Maas: It would be nice to have some clarification. I’d like to tell you that all of the decisions I issued with respect to costs were wholly consistent over my years on the bench, but I’m not even certain that that’s accurate. It’s an area that would benefit from clarification.
Logikcull: You mentioned in a recent Legaltech News article that the changes to Rule 37(e) are among the most potentially misunderstood of the recent amendments. Where do you think people are getting that rule wrong?
Maas: Well, I said that probably six months ago because there were decisions from some judges who seemed to be treating either negligence or gross negligence as the manifestation of a party’s intent to deprive another party of information that should have been made available. They then improperly resorted to one of the extreme remedies in amended Rule 37(e)(2).
But I also said at the time that most judges were getting it right. It really was a concern that arose out of a small number of decisions. In the intervening months, I think judges are becoming increasingly more familiar with the amended rule. At this stage, I’d say my concern still exists, but not nearly to the extent that it did when I made that statement.
Logikcull: Where do you stand on the issue of the court retaining its inherent authority as it relates to Rule 37(e)? I know it seems like the line of reasoning on this was fairly consistent until Judge Francis said that maybe people are thinking about it wrong. What are your thoughts?
[In CAT3 v. Black Lineage, U.S. Magistrate Judge James C. Francis defended courts’ ability to bring sanctions for spoliation under their inherent authority, writing: “Where exercise of inherent power is necessary to remedy abuse of the judicial process, it matters not whether there might be another source of authority that could address the same issue.” Logikcull covers the case in more depth here.]
Maas: Well, the advisory committee that drafted the recent changes included a lot of smart people. After considerable study, they concluded that they could, through rulemaking, prevent courts from relying on inherent authority as a basis for sanctioning parties for the loss of ESI. And then, as you know, my former colleague Judge Francis reached a seemingly contrary conclusion in the CAT3 v. Black Lineage case. You may be aware that he, along with Eric Mandel, also wrote an article for the Sedona Conference Journal that reached that conclusion.
What I didn’t see discussed very much in the newspaper articles and blogs that came out after CAT3 was the fact that what Judge Francis said concerning inherent authority was dictum, since he also found that he had the authority to act as he did under Rule 37.
Wholly apart from Rule 37, there also are other rules that expressly empower judges to deal with discovery misconduct.
I guess the bottom line is, I’m far less sure than Judge Francis is that inherent authority lives on in cases that fall within the four corners of Rule 37(e). I think it’s going to be interesting to see how the case law develops in this area.
In particular, it will be interesting to see whether other judges believe that the amended rule provides them with too few tools to deal with egregious discovery abuses without resorting to inherent authority as a basis for imposing sanctions. If I had to guess, I tend to doubt that will prove true. I think people will find sufficient authority in the rules to deal with the cases and litigants before them.
“I’m far less sure than Judge Francis is that inherent authority lives on in cases that fall within the four corners of Rule 37(e). I think it’s going to be interesting to see how the case law develops in this area.”
Logikcull: With regard to eDiscovery, in my view, having never been in your shoes obviously, one of the biggest challenges related to making proportionality decisions is the limited visibility into the actual costs of eDiscovery. That’s not limited to judges. In many cases, the other side in the dispute does not have access to what these things actually cost. Just in general, vendor pricing is very hard to come by.
Then there’s limited visibility as well into how various eDiscovery technologies actually work. I’m thinking specifically of predictive coding, but there are any number of other technologies out there. So, the upshot of that might be that judges may face situations where a producing party claims they can review and produce a large amount of data in X number of days, but it’s going to cost quite a bit of money and it’s going to have to be done through some kind of proprietary technology. I’d be interested to know if what I’ve described is an accurate assessment. The specific question is: How do courts ensure the information on which they’re basing these proportionality decisions is accurate?
Maas: It’s interesting that you ask me this because I was asked a similar question at a conference in the Midwest last week, so, it is obviously an area of increasing concern.
One tactic that won’t work, but is often resorted to, is simply telling a judge or a discovery master that certain discovery is prohibitively expensive. You need to be able to provide metrics, and there has to be some knowledgeable person willing to stick his or her neck out to attest to the accuracy of the numbers being provided to the court. I don’t mean to say that necessarily this has to be done in affidavit form, but there has to be a presentation that seems credible to the court. Once you have the information before the court, lawyers and judges are trained to probe factual assertions and identify and test assumptions – and what they’re called upon to do in this area of ESI is really no different.
“You need to be able to provide metrics, and there has to be some knowledgeable person willing to stick his or her neck out to attest to the accuracy of the numbers being provided to the court.”
I don’t think that full-blown evidentiary hearings are necessary to resolve these sorts of disputes, but a detailed, unrushed discovery conference, and sometimes more than one of those conferences, should enable a judge (or discovery master) to reach conclusions, based on information that they have confidence in, regarding the true costs of the discovery that has been requested, whether it will be allowed, and to what extent.
Logikcull: What about in situations where the technology at issue is not easily understood by either the opposing party or the judge? How do you work in a way so that both parties and the judge are comfortable with the science at work, the expected outcomes, and the validating of whatever algorithms are at the heart of the technology? Have you ever approached issues like that? What was the process there?
Maas: When I would wander from booth to booth at Legaltech, I sometimes would kid that most vendors seemed to be saying that they had a black box: you fed in information in at one end and the correct responsive data came out the other end, but you couldn’t be told what was inside the box — which, I guess is a little disconcerting.
My view is that, as a judge, or now as a discovery master, I don’t have to understand the technology or the algorithms that underlie a tool. All I need to find is that it works with a reasonable degree of accuracy. I guess I therefore have just a one-word answer to your question, which is: sampling.
Parties need to review their non-responsive documents to ensure that responsive documents are not being missed to an unacceptable extent. They also need to undertake a similar analysis with respect to documents that are being withheld based on attorney-client privilege or work-product protection, to make sure that they’re not concluding, simply because one of the persons copied on a memo was an attorney, that it necessarily is a privileged document.
Equally so, regardless of what tool is being used, it’s critical for counsel and their clients to ensure that the right sources of ESI have been accessed. You can have the best tool in the world, but if you haven’t used it to search the files of all the necessary custodians or data sources, you’re not going to find all of the responsive ESI.
“My view is that, as a judge, or now as a discovery master, I don’t have to understand the technology or the algorithms that underlie a tool. All I need to find is that it works with a reasonable degree of accuracy.”
So, I’m less concerned with the intricacies of how a tool works and more concerned with the degree of accuracy that it achieves in a particular case. I’m told that certain tools are better suited for certain types of cases than others; that’s also beyond my technological knowledge, but I think the question before a particular judge or discovery master is “has it worked in this case?”
It’s kind of interesting because, in the old days, nobody ever challenged the skill or intelligence of the young associate or paralegal who was reviewing paper files to locate responsive documents. I can understand why parties are more dubious of technology-assisted review, but all of the studies seem to show that, as a general matter, it’s less prone to error than human review. Given the age of ESI that we live in, and the fact that there are frequently numerous duplicate sources of information, for example for email, it seems less likely that hot ESI documents or files can successfully be buried than was true in the past in the paper world — where if you missed one file box, the hot document might never be found.
“As in most areas of discovery, it seems to me that cooperation is really the key.”
As in most areas of discovery, it seems to me that cooperation is really the key. If counsel sit down with each other and review what they’re doing and why, how their tools works, and what their quality control procedures are, there is less likely to be disagreement and distrust, and judges hopefully will less frequently need to be involved.
Logikcull: That makes sense. Going back to the paper analogy you just made, you probably could make the argument given the complexity of some of this technology that some of it is truly black-box/proprietary algorithms. And it could be the case that a producing party, and I’m not saying this happens, essentially manipulates the system in a way that is letting more hot documents go unproduced. If you were the requesting party or the judge, you might have less visibility into that process as a whole…
Maas: One thing that hasn’t changed are the ethics rules.
Logikcull: Great point.
Maas: You can’t intentionally fail to produce a requested document. Even when we were back in the days of keywords, rather than black boxes, there was always the concern that a plaintiff might wanted to search for the word “accident” in a particular case in which the defendant was using euphemisms like “unfortunate incident” to describe an event. Accordingly, if you didn’t have “incident /2 unfortunate” as a search phrase, you might not find the documents you wanted. But, generally speaking, people also describe what they want by subject. Faced with such requests, I don’t think in the days of keywords you could ethically escape turning over unfavorable documents simply by saying “these are the keywords we’ve searched,” and thereby ignore the fact that the keywords used would not produce all of the responsive documents. Similarly, now, if you know that you’re using technology in a way that will prevent the production of certain key documents, you still have an ethical problem.
This post was authored by Casey C. Sullivan, Esq., who leads education and awareness efforts at Logikcull. You can reach him at firstname.lastname@example.org.