Spring 2014

The False Hope of Proposed Rule 37(e): Why It Will Not End Data Producers’ Over-Preservation Habits

The False Hope of Proposed Rule 37(e): Why It Will Not End Data Producers' Over-Preservation Habits

Saturday, April 19, 2014 - 11:45pm
Stanley Richards


Proposed Rule 37(e) of the Federal Rules of Civil Procedure is the latest effort to respond to the tendency of large data producers to over-preserve electronic information as a protection against the common law of spoliation sanctions and procedural rules regarding electronic discovery. The proposed rule’s failure to properly differentiate a breach of the duty to preserve from bad faith, however, threatens to negate its otherwise promising reforms.


Both the common law of spoliation sanctions and procedural rules have attempted to respond to the challenges posed by the growing volume of electronic data. The proposed Rule 37(e) of the Federal Rules of Civil Procedure is the latest effort to respond to these challenges, but, in its current form, it will not end the tendency of large data producers to over-preserve electronic information. Judges can easily interpret the rule to permit sanctions, even when a litigant does not act in bad faith or is merely negligent. This Remark shows how judges can interpret the rule in this way, and argues that a clear, bright-line rule would provide a more-helpful approach to discovery reform.

I.       The Problem Faced by Proposed Rule 37(e) 

The unique challenges posed by electronic discovery (“e-discovery”) obligations are well-documented.[1] Parties must maintain records of electronic information in order to produce it during discovery in case of a law suit. Discoverable electronic data includes “any writings, drawings, graphs, charts, photographs, sound recordings, images, and other data or data compilations stored in any electronic medium.”[2] As companies and governments have digitized more and more of their operations, IT departments have struggled to manage the rapid increase in the volume of electronic data that must be stored indefinitely.[3]

One approach to the problem is to delete certain data on a regular basis so as “to prevent a build-up of data that can overwhelm the most robust electronic information systems.”[4] Unfortunately for large data producers,[5] courts have not taken kindly to routine deletions of electronic data. In the early 2000s, some circuit judges issued spoliation sanctions for merely violating the duty to preserve, that is, negligently deleting electronic data prior to litigation when the litigant did, or should have, “reasonably anticipated” litigation.[6] Some courts have even read breaching this duty to preserve as bad faith conduct, which allows for the harshest sanctions (including dismissal or an adverse inference instruction).[7] As a result, large data producers began to over-preserve electronic information.[8] The Judicial Conference, through the Federal Rules of Civil Procedure (“the Federal Rules”), has attempted to address this problem,[9] but it has largely failed. The current rule has been widely criticized[10] and the threat of costly spoliation sanctions has continued, largely unabated.[11] It is manifestly inefficient for large data producers, with a plethora of legal and institutional obligations, to be on the hook for incredibly costly spoliation sanctions solely due to their negligent conduct—that is, solely for breaching a duty to preserve. The e-discovery regime as it now stands requires these actors to divert human capital and other resources to the preservation of often utterly-irrelevant electronic data.[12]

But the largest efficiency concern is the current regime’s negative impact on the legal discovery process. Because the mere negligence standard applied by the courts under Rule 37(e) has led to over-preservation, litigants must process far more data when litigation actually arises. Parties must sift through thousands of pages of irrelevant emails just to find one that is possibly relevant. The costs of this should not be underestimated: it can be 1,500 times more expensive to review data than to preserve it.[13] Moreover, the sheer quantity of data that must be processed can obscure the truth, and render if more difficult to locate relevant information. The system as it stands now may enrich litigation lawyers under a billable-hours system,[14] but too much value is being diverted to rote review tasks.[15]

II.     The Solution Suggested by Proposed Rule 37(e)

A newly proposed amendment to Rule 37(e),[16] which completely replaces the current rule, is the Judicial Conference’s second attempt at offering a safe harbor for electronic data deletions.[17] On June 6, 2013, the Standing Committee approved the rule without much discussion or dissent.[18] The rule opened for a sixth-month public comment period on August 15, 2013, and may be promulgated as early as December 2015.[19]

The proposed rule is divided into two parts. The first, 37(e)(1), addresses the judge’s choice between curative measures or sanctions for a party’s spoliation of potentially relevant data.[20] A curative measure is not a sanction[21] and the proposed rule distinguishes between the two by, ostensibly, requiring more egregious conduct for sanctions. Under e(1) a party need only “fail[] to preserve discoverable information that should have been preserved” in order to face curative measures,[22] but under e(1)(B)(i) and (ii), a party must also “cause[] substantial prejudice” and act in “bad faith” or “irreparably deprive[] a party” of the opportunity to litigate its claim in order to face sanctions.[23] So, the rule in its first part creates two different inquiries—one into whether information should have been preserved, and, if so, another into whether there was bad faith or irreparably deprivation of an opportunity to present a meaningful claim.

The second part, 37(e)(2), provides a non-exclusive list of five factors that courts may consider when imposing either curative measures under e(1)(A) or sanctions under e(1)(B). These factors apply both to the inquiry into whether discoverable information “should have been preserved” and the “willful or in bad faith” inquiry.[24] These factors include “the extent to which the party was on notice that litigation was likely and that the information would be discoverable” and “whether the party received a [clear, reasonable, and good faith] request to preserve information.”[25] The judge, in considering these factors, may make either a “failed to preserve discoverable information finding” or a “bad faith” finding. Only the former finding need be made in order to impose curative measures, but absent a finding of irreparable deprivation, both are required in order to impose sanctions.

Observers have overwhelmingly praised the proposed rule, believing that the list of factors will prevent the judges’ imposition of harsh sanctions for negligent data deletions.[26] Many of these observers argue that the rule, consistent with the drafters’ apparent intent, will shield parties from liability for accidental or careless deletions done neither in bad faith nor with a desire to suppress the truth.[27] With all due respect to these commentators, I believe their analyses are quite misguided.

III.    The Incentives Created by Proposed Rule 37(e)

The proposed Rule 37(e), however, will not put an end to large data producers’ over-preservation of data because it can easily be interpreted to (1) conflate the duty to preserve with bad faith; and (2) permit the consideration of factors separate from bad faith in the decision to institute sanctions as opposed to curative measures. That is, the rule by its text allows a court to find bad faith or institute sanctions even absent a culpable state of mind or a desire to suppress the truth.[28] I am not arguing that the proposed rule necessarily leads to these two interpretations, only that drafting permits judges inclined towards spoliation sanctions to levy such sanctions absent bad faith. Because this decision is entirely up to the judge’s discretion, large data producers rationally seeking to avoid drawing sanctions for bad faith conduct will continue to preserve data indefinitely.[29]

The factors in e(2) may create the conflation of the duty to preserve and bad faith. Two of the five factors identified for the bad faith inquiry relate intimately to the duty to preserve. Under e(2)(A), the court may consider “the extent to which the party was on notice that litigation was likely,” and under e(2)(C) “whether the party received a [clear, reasonable, and good faith] request to preserve information.”[30] So, if a party deletes data that the litigant had notice to preserve (thus the duty to preserve attached), the rule permits the judge to take this action into account in the bad faith inquiry. In discovery law, “notice” is a term of art. It is not limited to notice of a claim, but includes anything that should have led a party to reasonably anticipate litigation.[31] This is a low threshold,[32] so under e(2)(A) and (C), a party may not perceive future litigation, but the rule permits a judge to use this to make a bad faith finding.[33] In fact, by instructing judges that they may consider the duty to preserve in the bad faith inquiry, the rule threatens to deepen the conflation of the duty to preserve and become an engine for sanctions against many honest, though negligently implemented, data retention programs.

The fact that the proposed rule in e(2)’s introductory section seems to separate whether information “should have been preserved” (much like a duty to preserve inquiry) from whether it was deleted in bad faith does not necessarily sever this conflation. First, it is not clear that these are separate inquiries, for the proposed rule uses an “and” as a conjunction rather than “or.” This may suggest that the “should have been preserved” inquiry and bad faith inquiry are one and the same.[34] Second, even if the introductory text in e(2) is taken to create a distinction between these inquires, e(2)(A) and (C)—which relate intimately to the duty to preserve under current case law—represent a sort of Trojan horse by which a judge may be tempted to conflate the duty to preserve with bad faith.

Second, the rule provides no mechanism for restricting judges to curative measures for failing to preserve information “that should have been preserved,” while reserving sanctions for acts of bad faith. Because the five listed factors apply both to the curative measures/”should have been preserved” inquiry and the sanctions/bad faith inquiry, any one factor (not just those that restate the duty to preserve)[35] may be used to either authorize curative measures or sanctions. So although the judge is limited in the factors that are used to assess a party’s conduct (these must be “relevant”), she is not at all limited in the use of factors to choose between curative measures or bad faith sanctions. Courts, therefore, may consider a host of other factors, including whether there was a duty to preserve, the party’s capacity to absorb harsh spoliation sanctions, or the spoliation’s prejudicial impact, in opting to sanction as opposed to impose curative measures. For instance, the judge may find a party’s conduct improper under e(2)(E) because of a failure to seek the court’s guidance.[36] Because this is “relevant” to both the “should have been preserved” and bad faith inquiries,[37] a judge must choose whether to impose curative measures or sanctions. Although intuition may suggest that certain factors are only relevant to one inquiry and not the other, any judge can easily justify choosing sanctions for failing to seek the court’s guidance over curative measures by arguing, for example, that the party as a large data producer has the capacity to absorb a harsh spoliation sanction. In short, because the text fails to steer judges on the choice between curative measures and bad faith sanctions, there remains abundant leeway for judges to impose severe sanctions for a variety of reasons other than actual bad faith, which may catch a given litigant by surprise.

To be sure, sanctions under e(1)(B)(i) still require that the deletion of data cause “substantial prejudice.” I do not believe that this represents a serious barrier to spoliation sanctions, however, because some courts presume “substantial prejudice” if bad faith or gross negligence is shown.[38] As I illustrated above, the proposed rule makes such a finding quite easy for a judge. Also, as a practical matter, the absence of counterfactuals if the data in question is permanently deleted makes it is easy for the litigant to argue that the particular deletion in question “substantial prejudiced” their claims. Again, a judge is likely to defer to the arguments of such a litigant in the absence of counterfactuals or compelling evidence to the contrary.

Therefore, contrary to the Rules drafters’ intent, the proposed rule, by providing two avenues for courts to impose sanctions absent actual bad faith, does not provide a meaningful safe harbor for negligent data deletions. If promulgated in its current form, the over-preservation of data will persist.

IV.    A Solution to Proposed Rule 37(e)’s Deficiencies

The defects to the proposed rule could be remedied if the drafters created a brightline standard that limited judicial discretion by creating a clear nexus between conduct and consequences for spoliation. This can be achieved through three alterations to the proposed rule.

First, the committee should eliminate any references to the duty to preserve as factors in the imposition of spoliation sanctions. As things stand, courts that have imposed sanctions for mere negligence may invoke both e(2)(A) and e(2)(C) to continue imposing spoliation sanctions.[39] By eliminating the duty to preserve as a factor, judges cannot rely on the Federal Rules to impose spoliation sanctions simply for negligence.[40]

Second, the rule should more clearly state that a party who behaves negligently—that is, breaches a duty, but the breach was not in bad faith—is insulated from sanctions. To do this, the rule should have a two-tiered consequences structure in e(2). There should be a list of factors to determine negligent conduct, which, if found, should be remedied only by curative measures under e(1)(A) and a separate list of factors to evaluate bad faith conduct, which, if found, may be remedied by spoliation sanctions under e(1)(B). The separation of consequences, based on an list of factors, will limit judicial discretion and dampen the perennial duty to preserve by communicating to potential litigants that data retention programs implemented in good faith cannot be sanctioned, absent the very strong “irreparably deprived” standard in e(1)(B)(ii).

Third, the rule should lucidly embrace the principle of proportionality. The principle of proportionality holds that the court must, in its decision to issue sanctions or allow discovery requests, weigh the costs of document preservation or complying with a discovery request against the remedial value of that request.[41] In the pre-litigation context, a court may consider the costs that a party would bear in order to maintain electronic data; if a judge finds that to maintain every email would have cost a large amount, then spoliation sanctions may not be imposed even if litigation could have been reasonably anticipated. Proportionality would benefit large data producers because of their high data preservation costs.[42] I believe that if a party acts negligently, then a judge can insulate that party even from curative measures if the costs of preservation would have been greater than the benefits. So, whereas the proposed rule makes proportionality a factor under e(2)(D),[43] I believe that to protect large data producers it should be the factor in the case of negligent deletions wherein the probable costs of preservation would have exceeded its benefits.


Proposed Rule 37(e) claims to reduce incentives for large data producers to over-preserve data as a protection against sanctions; the proposed rule’s failure to properly differentiate a breach of the duty to preserve from bad faith, however, threatens to negate its otherwise promising reforms. As such, the Judicial Conference should consider an amendment to Rule 37(e) that properly distinguishes the duty to preserve from bad faith through a two-tiered analysis, and further creates a corresponding system of proportional remedies.

[1] See generally John M. Barkett, E-Discovery: Twenty Questions and Answers (2009); Rachel K. Alexander, E-Discovery Practice, Theory, and Precedent: Finding the Right Pond, Lure, and Lines Without Going on a Fishing Expedition, 56 S.D. L. Rev. 25, 27 (2011).

[2] 16 C.F.R. § 2.7.

[3] Oracle, Lowering E-Discovery Costs Through Enterprise Records and Retention Management (March 2007), at 6, http://www.oracle.com/technetwork/middleware/webcenter/content/records-retention-whitepaper-130956.pdf.

[4] Roland C. Goss, Hot Issues in Electronic Discovery: Information Retention Programs and Preservation, 42 Tort Trial & Ins. Prac. L.J. 797, 806 (2002). This is especially true because data can be 1,500 times more expensive to review than to store it. Symantec, Survey Finds Infinite Data Retention Leading to Costly Information Management Mistakes (Aug. 4, 2010), http://www.symantec.com/about/n ews/release/article.jsp?prid=20100804_01.

[5] I focus on large data producers, such as large companies and government agencies, because they face large e-discovery compliance costs as repeat players in civil and criminal litigation. My observations about the problem and failures of the attempted solution, see infra Part II, may also apply to, say, small businesses with video surveillance cameras. Although such small data producers have to preserve comparatively little data, they also tend to lack the sophisticated infrastructure and capital needed to store it most efficiently, which results in high marginal costs of storage.

[6] See, e.g., Residential Funding v. DeGeorge Fin., 306 F.3d 99 (2d Cir. 2002); Danis v. USN Commc’ns, Inc., 53 Fed. R. Serv. 3d 828, 828 (N.D. Ill. 2000); Philip M. Adams & Assoc., L.L.C. v. Dell, Inc., 621 F. Supp. 2d 1173, 1994 (D. Utah 2009). See Zubulake v. UBS Warburg LLC, 220 F.R.D. 212 (S.D.N.Y. 2003), for a thorough and widely applied standard and discussion of electronic data preservation requirements.

[7] See, e.g., Thomas Y. Allman, Conducting E-Discovery After the Amendments: The Second Wave, 10 Sedona Conf. J. 215, 220 (2009) (“Some courts seem to assume that once a duty to preserve attaches something akin to a strict liability follows for any [data] losses.”) (citation omitted).

[8] Throughout I will use the term “over-preserve” to refer to data preservation policies made out of fear of sanctions and what is preserved will actually prove irrelevant later. I do not define a precise referent point for what is the “right” level of document preservation to empirically justify this use of the term, but, as others have and I will argue, it is clear that the system as it stands now requires more preservation than is efficient. See infra notes 27 – 30 and accompanying text.

[9] See Fed. R. Civ. P. 37 (e). This rule remains operative.

[10] See, e.g., Thomas Y. Allman, Defining Culpability: The Search for a Limited Safe Harbor in Electronic Discovery, 2006 Fed. Cts. L. Rev. 7, at *16 (2006).

[11] Dan H. Willoughby, Jr. et al., Sanctions for E-Discovery Violations: By the Numbers, 60 Duke L.J. 789, 828 (2010) (“Parties … seeking refuge from the increasing sanction-motion practice will be able to reach Rule 37(e)’s refuge only in very limited situations. Since the rule’s adoption, approximately two cases per year have met its requirements.”). Readers should peruse the tables in this article to see numerous specific cases of judges imposing sanctions despite the rule. Id at 829 – 64.

[12] For example, federal agencies hire data custodians and spend millions of dollars to manage the thirty billion emails they collectively receive per year. George L. Paul & Jason R. Baron, Information Inflation: Can the Legal System Adapt?, 13 Rich. J.L. & Tech. 10, at *1.

[13] Symantec, Survey Finds Infinite Data Retention Leading to Costly Information Management Mistakes (Aug. 4, 2010),  http://www.symantec.com/about/news/rel ease/article.jsp?prid=20100804_01.

[14].         This may be a major policy concern by itself, depending on who you ask (e.g. a junior associate without partner ambitions who has received one too many doc review assignments).

[15] It is worth noting that predictive coding, a bundle of programs that target relevant electronic information, may substantially mitigate these costs in the long-term, but these technologies are in their infancy. See Moore v. Publicis Groupe, 287 F.R.D. 182, 186 – 87 (S.D.N.Y. 2012), for an example of a judge requiring predictive coding for a case with vast electronically stored information.

[16] I will refer to this throughout as the “proposed rule,” for it has yet to be promulgated. The proposed rule is not nearly as pithy as the current one. Therefore, I will only quote certain parts of it. For full text of the rule that the Standing Committee accepted without modification as well as the Advisory Committee’s Notes, see Advisory Committee on Civil Rules, Report to the Standing Committee 43 – 59  (May 8, 2013) (hereinafter Report of Rule 37(e) Draft), http://www.uscourts.gov/ uscourts/RulesAndPolicies/rules/Reports/CV05-2013.pdf.

[17] Interestingly, the proposed rule applies to both electronic data and “traditional” paper document discovery. In this Remark, however, I will focus exclusively on electronic data. Discussing the reason for why the proposed rule was expanded to include all types of data is beyond the scope of this Remark and does not change my thesis about the proposed rule’s likely effectiveness.

[18] Henry Keltson, Proposed Discovery Amendments Move to Public Comment (June 6, 2013), http://www.law.com/jsp/lawtechnologynews/PubArticleLTN.jsp?id =1202603039841.

[19] Lauren Aguiar & Jeffrey Geier, Skadden, Proposed Federal Rule 37(e): Savior From Sanctions? 3 (Apr. 10, 2013), http://www.skadden.com/sites/default/files/publications/070041319%20Skadden.pdf.

[20] Report of Rule 37(e) Draft at 43.

[21] Curative measures are remedial, meaning they seek to make a litigant whole if potentially relevant data is deleted. Sanctions may have a remedial purpose, but they are also punitive and tend to be more costly.

[22] Report of Rule 37(e) Draft at 43.

[23] Id. Whether a party failed to preserve discoverable information that should have been preserved” under (e)(1), then, is a threshold requirement for both curative measures and sanctions.

[24] The advisory committee notes explicitly state that the factors apply to both inquiries. Id. at 48 (“These factors guide the court when asked to adopt measures under Rule 37(e)(1)(A) due to loss of information or to impose sanctions under Rule 37(e)(1)(B) (emphasis added)).

[25] Id. at 43 – 44.

[26] See, e.g., Aguiar & Geier, supra note 18, at 3 (“[Rule 37(e)’s] adoption would end the current circuit split.”); H. Christopher Boehning & Daniel J. Toal, Paul Weiss, Propose Rule 37(e): A Step in the Right Direction? 3 (June 4, 2013), http://www.paulweiss.com/media/1666131/nylj_4jun13.pdf (“[T]he revised Rule 37(e) is almost certainly a step in the right direction.”); Gibson Dunn, Electronic Discovery Year-End Update: Moving Beyond Sanctions and Toward Solution to Difficult Problems (2012) at 1 to 2 (“If adopted, the revised rule should help prevent the imposition of harsh sanctions for inadvertent preservation failures.”). The law firm also claims that the revised rule is “real and meaningful progress.” Id. at 10.

[27] Both the 2006 rule and the newly proposed rule contemplate that sanctions only be issued for those parties with a “culpable state of mind” above mere negligence. See Summary report of the Judicial Conference Committee on Rules of Practice & Proc. C-88 (Sept. 2005), http://www.uscourts.gov/uscourts/RulesAndPolicies/rule s/Reports/ST09-2005.pdf.

[28] It is clear that this is contrary to the Rules drafters’ intent, for the drafters claim explicitly that the amendment rejects court cases that have made Rule 37(e) a mere negligence standard. Report of 37(e) Draft at 35.

[29] This is because large data producers ought to “assume that [they] will be sued in the one circuit where negligence is grounds for sanctions.” Alexander Dahl, Spoliation, Uncertainty, and Sekisui (Nov. 7, 2013), http://www.law.com/jsp/lawtechno logynews/PubArticleLTN.jsp?id=1202626301277&slreturn=20131016142028.

[30] Id. 43 – 44 (emphasis supplied).

[31] Kronisch v. U.S., 150 F.3d 112, 126 (2d Cir. 1998) (defining notice as including “when a party should have known that the evidence may be relevant to future litigation”).

[32] For example, even informal complaints, like a written notice, see Broccoli v. Echostar Commc’n Corp., 229 F.R.D. 506, 511, 515 (D. Md. 2005), or threat of litigation, see In re Napster, 462 F. Supp. 2d 1060, 1069 – 70 (N.D. Cal. 2006), give rise to a duty to preserve.

[33] Elsewhere in e(2)(B) the rules assert that judges may take into account the reasonableness of preservation efforts, but this is merely optional. Id. at 44.

[34] Report of Rule 37(e) Draft at 44. Although this interpretation would not confer any meaning to the text in e(1), which seems to separate the two inquiries.

[35] See supra note 31 and accompanying text.

[36] Report of Rule 37(e) Draft at 44.

[37] Id. at 43.

[38] A. Benjamin Spencer, The Preservation Obligation: Regulating and Sanctioning Pre-Litigation Spoliation in Federal Court, 79 Fordham L. rev. 2005, 2017 & n.73.

[39] Not all the circuits have used Rule 37(e) to impose sanctions for mere negligence, but the Second Circuit has been especially quick to sanction parties if there was a duty to preserve. For a pithy breakdown of the circuit split, see Aguiar & Geier, Proposed Federal Rule 37(e): Savior From Sanctions? 3 & n.1 (Apr. 10, 2013), http://www.skadden.com/sites/default/files/publications/070041319%20Skadden.pdf.

[40] It is true, however, that the courts may rely upon their “inherent powers” to sanction parties and circumvent the Federal Rules of Civil Procedure. See Chambers v. NASCO, Inc., 501 U.S. 32, 45 – 46 (1991).

[41] See, e.g., Mancia v. Mayflower Textile Servs. Co., 253 F.R.D. 354, 358 (D. Md. 2008) (condemning “kneejerk discovery requests served without consideration of the cost or burden to the responding party”).

[42] See supra notes 3 to 4 and accompanying text.

[43] Report of Rule 37(e) Draft at 44.

J.D. Candidate, Yale Law School Class of 2015

Cite this piece: Stanley Richards,

The False Hope of Proposed Rule 37(e): Why It Will Not End Data Producers’ Over-Preservation Habits

, Yale L. & Pol'y Rev. Inter Alia (April 19, 2014), http://ylpr.yale.edu/inter_alia/false-hope-proposed-rule-37e-why-it-will-not-end-data-producers-over-preservation-habits.