Litigation Discovery Cannot Be Optimal But Could Be Better: The Economics of Improving Discovery Timing in a Digital Age


Litigation Discovery Cannot Be Optimal But Could Be Better: The Economics of Improving Discovery Timing in a Digital Age

  • 1 Tags
By Scott A. Moss, Professor, University of Colorado Law School

courtsCost-benefit “proportionality” limits on discovery have long been prescribed by a wide range of commentators. The judiciary codified a proportionality requirement in Federal Rule of Civil Procedure 26(b)(2)(C) and later in e-discovery rules based on proportionality principles, such as the Rule 26(b)(2)(B) proviso that only upon “good cause” can there be discovery of computerized data “not reasonably accessible because of undue burden or cost.” Drives to limit discovery typically gain strength when technology increases discovery cost. The original proportionality rule arose when the spread of photocopying technology changed discovery from in-person inspections to massive document productions. Similarly, the e-discovery rules arose once mass digitization of corporate and other records expanded discovery from paper exchange to costly examination of data existing in high quantities, in deleted media, and on obsolete hardware—endeavors that can cost tens or hundreds of thousands of dollars just to find old e-mails, digitally scanned paperwork, or database content.1

This Article dissents from the consensus in favor of proportionality rules, but not for the same reasons as those criticizing proportionality as too strong or too weak a limitation on discovery cost. Many who criticize proportionality rules as ineffective place the blame on bad rulemaking or judicial decisionmaking;2 even if those critiques are correct, this Article diagnoses the problem as a more fundamental one that better rules and decisions cannot fully fix. This Article sees proportionality limits as impossible to implement effectively: sometimes they fail to curb discovery excess or allow costly discovery on meritless claims; other times they disallow discovery that meritorious cases need.

The problem with proportionality rules is that they ask the impossible: judges must decide when discovery cost is proportional to some measure of “value” that includes both evidence value to jury deliberation and case value to the parties and society. This yields a fundamental information-timing problem: discovery disputes occur before parties marshal all the evidence, so how can courts measure the value of particular evidence, much less case merits? Further, discovery has more probative value in close-call cases than in the strongest and weakest cases (in which more evidence is less likely to affect case outcome)—an observation paralleling Professor Robert Cover’s classic 1975 argument that litigation-procedure rulings cannot truly be independent of case merits, contrary to the conventionally assumed “transsubstantive” nature of the federal rules. Case merits, though critical to discovery decisions, typically remain hidden in a cloud of uncertainty during discovery because the court is not yet able to sift fully through the evidence and arguments.


Due to the information costs (including time) of assessing case merit during discovery, courts often cannot tell which litigants’ braggadocio is cheap talk and which reflects real case merit. As a result, courts must ignore parties’ merits arguments and adjudicate discovery disputes as if all cases of a similar type in the pool (that is, cases arising under the same statute that are neither facially frivolous nor obvious winners) warrant similar discovery. Courts’ discovery rulings should be based on the merit of the claims, but merit cannot be communicated effectively; as a result, those rulings must be based on the average value of all cases in the pool. In this pooling equilibrium, the best available strategy for courts is to rule the same on all cases in a pool regardless of case merit—even though these rulings are suboptimal in the sense of yielding too much discovery in low-merit cases and too little discovery in high-merit cases. Other scholars and commentators have noted that discovery featuring pooling equilibria requires making rulings based on rough averages; this Article depicts the problem as pervasive and intractable because the rules require proportionality inquiries that demand of judges an unrealistic level of knowledge not only of the disputed evidence, but also of the case merits.

Under this analysis, the quest for better discovery limits has disappointed not because of bad decisionmaking or bad rulemaking, but because courts and parties are stuck in a pooling equilibrium. This is a fundamental information-timing problem inherent in the discovery stage of litigation: optimal evidence-gathering decisions require more merits analysis, but merits analyses require more evidence gathering.

Deferring close decisions on potentially useful but costly evidence until case merit is clearer—until meritorious cases distinguish themselves, turning a pooling equilibrium into a separating equilibrium—is one possible solution to the pooling equilibrium. Fortunately, litigation reaches just such a point—after summary judgment. In deciding summary judgment, courts allow to proceed to trial only claims a reasonable jury could decide in favor of either party, weeding out both claims with the lowest probability of merit (summary judgment grants to defendants) and claims with the highest probability (grants to plaintiffs). A case reaching trial, having survived summary judgment, has a reasonable probability of merit. Even without adopting the old theory that cases reaching trial likely have fifty-fifty odds, cases reaching trial are more likely than others to be close calls, and they certainly have higher average merit than the pool of all filed complaints. More evidence, like costly electronic data, has more value to the jury in close-call cases than in very weak or strong cases.

Accordingly, much of the scholarly debate on discovery misses the mark by focusing on how much to limit costly discovery, such as with proportionality rules and numerical caps. Rather, the debate should focus on when in litigation to allow costly discovery. Specifically, decisions regarding costly discovery should be postponed until after summary judgment to ensure that costly discovery is imposed only in cases with a greater probability of merit. Although this proposal might enable judges to deny or postpone more discovery, any discovery denied or postponed under this proposal would probably already be denied based on judges’ proportionality discretion. Thus, the main utility of this proposal would be to explain how courts could allow more discovery—only after summary judgment—of helpful but costly evidence that courts often disallow and declare nondiscoverable.

RealtimeNobody previously has suggested solving the dilemma of costly discovery with post–summary judgment discovery, which might seem to be a counterintuitive idea; under Federal Rule 56(f) and the case law applying that rule, summary judgment typically comes only after all discovery is completed. But there is no requirement that all discovery must precede summary judgment, which is why courts sometimes allow summary judgment motions limited to threshold questions (like a defendant’s immunity from suit).

Unusually costly evidence should be another exception to the rule of summary judgment coming after all discovery. Surviving summary judgment means a case is likely the sort of close call warranting more fact gathering, so courts should allow truly costly discovery, like the heavy e-discovery that they commonly disallow, only once a case survives summary judgment. No rule change is required to implement this Article’s proposal that courts revisit denials of burdensome discovery if a case survives summary judgment: existing rules give courts broad case management authority, including authority over the timing of discovery (Rule 16(c)(2)(F)), the timing of summary judgment motions (Rule 16(c)(2)(E)), and the sequencing of discovery (as the Rule 16 Advisory Committee’s note elaborates). Thus, this proposal could not only improve litigation discovery, but it could also provide a welcome answer to courts’ riddle of how to rule on proportionality without circular, premature case-merit evaluations. A new rule would be advisable, though, to minimize the risk of courts misusing the proposal to deny discovery excessively.

This Article’s analysis and proposal illustrate a broader point about economic analysis. Fitting into a line of scholarship analyzing litigation as a series of points in time when information emerges,3 this Article indicates that for economic analysis of litigation to provide accurate diagnoses and useful recommendations, it must do more than just prescribe cost-benefit comparisons; it must consider the timing-and-stages nature of litigation, such as by delving into the details of discovery, prelitigation settlement, and other events short of trials and dispositive motions.


Copyright © 2009 Duke Law Journal.
Scott A. Moss is Associate Professor, University of Colorado Law School.
Scott A. Moss can be reached by e-mail at

This Editorial is based on the following full-length Article:  Scott A. Moss, Litigation Discovery Cannot be Optimal but Could be Better, 58 DUKE L.J. 889 (2009). Click here for the full article

  1. See, e.g., In re Brand Name Prescription Drugs Antitrust Litig., 94 C 897, 1995 U.S. Dist. LEXIS 8281, at *2-3 (N.D. Ill. June 13, 1995) (granting class action plaintiffs’ motion to compel the defendant to produce computer-stored e-mail at the defendant’s own expense, estimated at $50,000 to $70,000); PSEG Power NY, Inc. v. Alberici Constructors, Inc., No. 1:05-CV-657, 2007 WL 2687670, at *1, *9-*10 (N.D.N.Y. Sept. 7, 2007) (ordering the plaintiff, in a $4.4 million construction contract claim, at a cost of $40,000 to $200,000, “to produce all electronically stored emails, numbering approximately 3000, conjunctively with their corresponding attachments as ‘married’ documents”); W.E. Aubuchon Co. v. BeneFirst, LLC, 245 F.R.D. 38, 44 (D. Mass. 2007) (ordering the defendant, in a claim that an employee benefit administrator breached its fiduciary duty, to produce thousands of employee claim forms and medical bills stored electronically as unindexed images, at an estimated cost of $80,000 and 4000 hours); Wiginton v. CB Richard Ellis, Inc., 229 F.R.D. 568, 577 (N.D. Ill. 2004) (requiring class action harassment plaintiffs to pay 75 percent of a $249,000 e-mail search for known pornographic and other harassing e-mails); Best Buy Stores, L.P. v. Developers Diversified Realty Corp., 247 F.R.D. 567, 569-72 (D. Minn. 2007) (denying defendants’ request for plaintiff’s database on other landlords’ lease charges because the data was not in a searchable format, required restoration costing $124,000 plus $27,823 per month, and could be compiled from paper discovery).
  2. See, e.g., Henry S. Noyes, Good Cause Is Bad Medicine for the New E-Discovery Rules, 21 HARV. J.L. & TECH. 49, 71 (2007) (criticizing proportionality and e-discovery rules as too vague to rein in excess discovery that courts are too unwilling to limit); Martin H. Redish, Electronic Discovery and the Litigation Matrix, 51 DUKE L.J. 561, 563-64 (2001) (noting that “the rules’ drafters and revisers over the years . . . have failed to fashion a discovery process that satisfies most people,” and specifically criticizing discovery rules for lacking more cost shifting or spoliation provisions); Thomas D. Rowe, Jr., A Square Peg in a Round Hole? The 2000 Limitation on the Scope of Federal Civil Discovery, 69 TENN. L. REV. 13, 14 (2001) (criticizing the federal rules’ discovery limits as vague and therefore unable to change judicial decisionmaking).
  3. See, e.g., Joseph A. Grundfest & Peter H. Huang, The Unexpected Value of Litigation: A Real Options Perspective, 58 STAN. L. REV. 1267, 1270-71 (2006) (using “real options theory,” “{a} tool{} applied to the economic analysis of research and development projects,” to model litigation as a series of discrete stages with progressively greater information available to the parties); Scott A. Moss, Illuminating Secrecy: A New Economic Analysis of Confidential Settlements, 105 MICH. L. REV. 867, 877 (2007) (analyzing settlement confidentiality based on information distinctions between settlements reached before and after litigation commences).
Almanya sohbet anal yapan escort