Barbara A. Caulfield and Zuzana Svihra
Orrick, Herrington & Sutcliffe LLP
San Francisco, California

As is the case in many areas of the law, the technological advances surrounding digital or electronic-based discovery have far surpassed the law’s current ability to regulate them. This is true both with regard to the Federal Rules of Civil Procedure and case law. In reaction, there has been a rash of legal writings attempting to alert the community to the lack of guidance available, offer perspective and interpretation, and provide solutions. Despite the rising costs, exposed abuses, and economic waste incurred due to the ever-widening scope of digital discovery requests, few alternatives seem to have been raised or applied.

The debate to change the Rules has become stagnant due to two entrenched and opposing perspectives. On the one hand are those who do not want to change the Federal Rules, either regarding the scope of discovery or in terms of shifting costs. This camp believes shifting costs would effectively silence poorer litigants from bringing many valid lawsuits and/or defending against non-meritorious claims. This side also argues the Federal Rules provide adequate guidance for the courts. [1] The other side consists of defense lawyers and corporations who often foot the bill for digital discovery, scope and breadth of which appears to be continually increasing. This side finds the exponential increase in discovery costs due to the expense of retrieving computer-based data alarming. This cost has been described as a new negotiating tool to "blackmail" corporate defendants and force many to settle rather than bear the financial burden.[2]

Due to the absence of a coherent body of law,[3] courts seem to decide digital discovery disputes based on an amalgamation of their own armchair knowledge of technology and precedent from traditional forms of discovery disputes. While this precedent does provide some degree of guidance, the issues that arise concerning electronic discovery result in unique problems that never surfaced in traditional discovery settings. Moreover, since discovery decisions are interlocutory, it is rare for a party to appeal a discovery decision and therefore there is little appellate case law. Greater guidance from the Federal Rules, reflecting these recently developed practical, economic and legal issues, would be an invaluable resource for future cases.

This article is divided into four discrete sections. The first section will provide a brief description of computer-based data to help explain why the costs are so high. The next section will discuss the proportionality amendment to the Federal Rules and compare the federal standards with the changes made in the Texas court rules. The third section will outline the status of case law. It will describe the general trends and outlying cases and the issues raised therein. The fourth section will advocate that the federal rules be modified to require the party who loses at trial to pay for all or a portion of the discovery expenses. Without greater regulation of discovery requests, litigants are effectively at sea when attempting to predict discovery costs for litigation.

While record-keeping systems for conventional paper-based discovery are far more organized than their electronic counterparts, the sheer mass of documents continually archived by companies makes routine expunction a necessity.[4] Except upon notice of litigation, expunction of paper documents, in the course of ordinary business, is legally acceptable and widely practiced. Electronic-record collection is not nearly so spatially limited and each year it becomes even less so. As a result, companies are able to store a much larger amount and assortment of materials in a much smaller area for a longer period of time. For instance, drafts of documents can be stored and retrieved, where normally they would have been destroyed.[5] Unfortunately, unlike their paper-based counterparts, electronic records management systems seldom exist.[6] Thus, the rising costs of electronic retrieval has raised justified concerns.

Electronic discovery encompasses four distinct types of data: active, replicant, archival and residual data.[7] Active or native data is what "actively" resides on the user’s hard drive and/or the network server, which is immediately accessible. Replicant data is what is termed "file clones," e.g., the redo and undo button, which are not immediately accessible and can be expensive to retrieve. Archival or legacy data is compiled in back-up tapes. Back-up tapes generally consist of everything on the server at a given time. This data is not immediately accessible because it is not saved in a user-friendly format. Collecting the data from back-up tapes usually requires hiring a technician to write a program to retrieve the data, which can be very expensive.[8]

And finally, and most importantly for the purposes of this article, information can be stored as residual data, which consists of deleted files and e-mails. Deleted files, also known as ghost files, are not actually deleted until the medium on which they reside has been completely overwritten by the system with another file. The possibility of these files actually being overwritten decreases as the typical storage capacity of computers grows. There have been a number of practical and theoretical issues raised surrounding these files.

First, retrieving these files is the most costly of all four of the data types. Second, there have been questions raised as to whether these deleted files are really "in the possession" of the party, as required by the Federal Rules.[9] Obviously, it would be absurd to force a responding party to retrieve shredded or discarded documents from the county dump. These paper documents are clearly not within the possession of the responding party. However, because the deleted files are still on the computer or perhaps on the backup files, albeit unintentionally, courts and parties appear to have presumed these files to be "documents" within the party’s possession.[10]

Third, the exponential increase in the volume of e-mail appears to be demanding ever-increasing storage capacity.[11] E-mails have begun to play a central role in litigation, often coined the new "smoking gun" for litigation.[12] This role seems to be predominantly due to the uniqueness of the medium itself. Users seem to view e-mail as a form of communication sharing attributes of phone and in-person conversations as well as company inter-office memos. This perspective encourages far more informal writing: e-mails are often carelessly worded and poorly written. The result can be devastating for a company.[13] As a result, propounding parties have sought to widen the e-mail requests as much as possible. In addition, the informality of e-mail often makes retrieval by computer-based word searching and screening for relevance and privilege costly and more time-consuming.[14]

There is a growing position that technological advances are already and will continue to make discovery cheaper and easier for both sides.[15] However, this has not occurred yet. Moreover, there is some question as to whether retrieval of archival or legacy data will ever be less expensive.[16]

This section will describe how the current federal rules have responded to deterring digital "fishing expeditions." It will describe the proportionality or undue burden test the Federal Rules have delineated as a means to prevent abusive discovery requests. It will then describe the Texas Rules of Civil Procedure and compare Texas’ blanket rule to the discretionary test federal courts utilize.

In 1983, Federal Rule of Civil Procedure 26(b)(1) was amended to address unreasonable burdens on the producing party in discovery requests involving computer databases. The rules were further changed in 1993, due to the "information explosion of recent decades [which] has greatly increased both the potential cost of wide-ranging discovery and the potential for discovery to be used as an instrument for delay or oppression." See Fed. R. Civ. P. 26, Advisory Committee Notes, 1993 Amendment. The Committee noted the changes made were intended to emphasize the courts’ "broader discretion to impose additional restrictions on the scope and extent of discovery." Id.

The rule’s amendment created the undue burden or proportionality test, which states "[t]he frequency or extent of use of the discovery methods … shall be limited by the court if it determines that … (iii) the burden or expense of the proposed discovery outweighs its likely benefit, taking into account the needs of the case, the amount in controversy, the parties’ resources, the importance of the issues at stake in the litigation, and the importance of the proposed discovery in resolving the issues."[17] Essentially, the rule provides the court with the ability to shift costs when it deems the economic burden on the party is too great, relative to the possible benefit derived from the production of the documents.

The Advisory Committee Notes explain how a court should interpret the facts relative to the factors. "The elements … address the problem of discovery that is disproportionate to the individual lawsuit as measured by such matters as its nature and complexity, the importance of the issues at stake in a case seeking damages, the limitations on a financially weak litigant to withstand extensive opposition to a discovery program or to respond to discovery requests, and the significance of the substantive issues, as measured in philosophic, social or institutional terms. Thus, the rule recognizes that many cases in public policy spheres, such as employment practices, free speech, and other matters, may have importance far beyond the monetary amount involved. The court must apply the standards in an even-handed manner that will prevent use of discovery to wage a war of attrition or as a device to coerce a party, whether financially weak or affluent."[18]

Despite the apparent intent of the Committee Notes and the "war of attrition" forewarning, few cases have actually decided to shift costs.[19] While there are both factually-specific and principled reasons for these courts’ decisions, the current atmosphere is extremely troubling for the producing party. One could argue the discretionary aspects of the rule make the court’s predilections towards the parties, the case, or its view towards corporate parties versus individuals far more determinative than the proportionality factors. This degree of discretion makes discovery decisions far less predictable and expensive. While the Federal Rules should not discourage relevant discovery, there needs to be stronger language to discourage frivolous suits that create "fishing expeditions."

The Texas Rules seek to strike a balance between the propounding party’s rights to discovery and the defendant’s rights to be protected from undue burden. Rule 196.4 of the Texas Court Rules of Civil Procedure is currently the only rule that specifically mandates shifting the cost for retrieval of digital discovery. The rule states:

Tex. R. Civ. P.196.4 (emphasis added). The operative phrases are whether the documents have been held in digital form during the "ordinary course of business." Upon such a finding the responding party must make "reasonable efforts" to retrieve the documents. However, if "extraordinary steps" must be taken, the propounding party will bear the expense.

While the Texas rule offers some degree of discretion for the court, it is far more limited than its federal counterpart. By placing the costs up front it necessarily precludes less financially stable parties from certain types of discovery. In addition, opponents argue courts have already created adequate standards and precedent for dealing with discovery issues.[20] It seems highly unlikely that similar language will be adopted for the federal rules.

This section will seek to provide a comprehensive depiction of courts’ decisions and their reasoning in digital discovery motions and describe the economic burden this precedent places on future litigants. Electronic retrieval is costly on two fronts. First, the sheer cost to hire technicians, develop software, and retrieve the documents can be financially disabling. Despite these facts, the majority of courts have rejected shifting any portion of these costs. In the alternative, courts that have allowed a propounding party to have on-site access to its opponent’s computer system have exerted greater safeguards and have shifted costs. Second, the concerns of waiving the attorney-client privilege or the work product doctrine require the producing party to implement additional standards and amass expense from the additional time and person-power it takes to review the documents. Courts have applied the same privilege standards, without greater deference due to the size of the documents and the expense it takes to review them. The privilege review has become even more crucial (and therefore expensive) given the recent decisions granting access for third party intervenors to review all documents attached to dispositive motions, regardless of a protective order being in place.

This subsection will outline the wide spectrum of cases that have addressed cost shifting motions. These decisions seem to vary based on the court’s responses to the parties. On the one hand are courts that presume the responding party can bear the expense and place a heavy burden on the responding party to explain why it shouldn’t bear the costs. On the other end, there are a handful of courts that appear to be genuinely concerned with the "war of attrition." These courts are more scrupulous as to the relevancy and specificity of the requests, taking into consideration the expense to the producing party and the degree to which the requests are overly broad or abusive.

The following is a list of the cases that have either shifted costs or, due to the expense, have simply refused to compel the production. In Zonaras v. General Motors Corp., 1996 U.S. Dist. LEXIS 22535 (S.D. Oh. 1996), the plaintiffs sought to compel discovery of data compiled pertaining to a number of different crash dummy tests. Zonaras is one of the few cases to rely on the Federal Rule’s proportionality test in determining whether to grant the motion. The defendant argued it had already produced a significant amount of data pertaining to a number of other tests. After balancing the elements outlined in Rule 26(b)(2)(iii), the Zonares court concluded the benefits of discovery outweighed the expense. But because it was unclear as to whether the evidence would be admissible at trial, the court ordered the plaintiff to pay half the cost of production. See Zonaras, supra at *10.

Applying a slightly different analysis, the court in Anti-Monopoly, Inc. v. Hasbro, Inc., 1996 U.S. Dist. LEXIS 563 (S.D.N.Y. 1996), required the propounding party to pay for the costs of producing computerized data. The court adopted the axiom that the requesting party must pay for data that can "be extracted only by special programming," which must be newly written and implemented. Anti-Monopoly, Inc., 1996 U.S. Dist. LEXIS 563, *4-5. The court rejected the argument that it was the responding party’s fault that the retrieval required a special program. Id; see also In re Air Crash Disaster at Detroit Metro, 130 F.R.D. 634 (E.D. Mich. 1989) (ordering propounding party to pay for the expense of the responding party duplicating the data onto a computer-readable tape, because data had not previously existed in electronic form). Moreover, the court concluded the undue burden test should focus not on whether the propounding party could afford the costs, but rather, whether the cost was substantial. Id. at *6.

A unique approach was outlined in Torrington Co., et al. v. U.S., 786 F. Supp. 1027, 16 C.I.T. 76 (Ct. Intl. Trade 1992). In Torrington the court declined to order the defendant to create computer tapes from scratch, when plaintiff already had received all the documents in paper form. The court held the plaintiff "has not adequately articulated the need for the computer tapes" whereas the defendant "has validly enumerated extreme hardship." Torrington, 786 F. Supp. 1029, citing Williams v. Owens-Illinois, Inc., 665 F.2d 918 (9th Cir. 1982), cert denied, 459 U.S. 971 (1982) (holding appellants were not entitled to computer tapes when all the information was produced in hard copy).

In reaching its decision, the Torrington court stated, "[w]here the burden, cost and time required to produce the tapes is virtually equal on both parties, then the burden of producing the tapes falls on the party requesting the information." Torrington, 786 F. Supp. at 1030; relying Oppenheimer Fund, Inc. v Sanders, 437 U.S. 340(1978) (where expense of creating computer programs would cost the same for both parties, court ordered the party seeking the information must bear the cost).[21] Although this principle has not been cited elsewhere in the context of discovery motions, it offers an interesting formula for cost shifting analysis.[22]

The Oppenheimer court rejected applying a greater burden of production simply because the party maintains records on computer tapes. The court described the procedural history of the case, explaining how the district court and Second Circuit concluded the defendants should pay for the discovery costs because the defendants might try to "irretrievably bury information to immunize business activity from later scrutiny" or use needlessly complex programs for retrieval to discourage discovery." Oppenheimer, 437 U.S. at 362. The court held that absent an indication of bad faith, costs should not be shifted simply because of these conjectured risks. The court noted, "we do not think a defendant should be penalized for not maintaining his records in the form most convenient to some potential future litigants whose identity and perceived needs could not have been anticipated." Id. at 362.

The Seventh Circuit upheld a district court’s decision to shift half the cost of copying e-mails into hard copy from computer tapes, because the tapes were not computer readable without defendant’s software and equipment. Sattar v. Motorola, Inc., 138 F.3d 1164 (7th Cir. 1998). Sattar noted the district court authorized partially shifting the cost only as the last option. The district court offered three options: (1) plaintiff could get the information by on-site access to the system, (2) defendant could loan plaintiff the necessary software, or if the former two were not options, (3) defendant would download the data from tapes to computer disks or a computer hard-drive. Sattar, 138 F.3d 1171.

Williams v. E.I. du Pont de Nemours & Co., 119 F.R.D. 648 (W.D. Ky. 1987), required that the propounding party pay for a copy of the computerized database and accompanying software, that compiled thirty years worth of the propounding party’s employment file. Williams, 119 F.R.D. at 651. Although the data was originally the propounding party’s, it had been categorized and scanned in by the responding party. The court allowed the propounding party access to the digital version to enable it to view the data with greater convenience and ease for the purposes of preparing for expert testimony. Williams, 119 F.R.D. at 651. The court not only required the propounding party to pay for the copying, it also required it to pay for a portion of the fees and expenses incurred from the expert who encoded the data in the database. Id.; see also Fautek v. Montogmery Ward & Co., 91 F.R.D. 596 (N.D. Ill. 1980) (plaintiff required to pay for half of a computerized database created for litigation because plaintiff may not "piggy back" on the work done by the opposing party at great expense).

There has been one case that refused to order document production based on the expense. The court in Alexander v. FBI, 188 F.R.D. 111 (D.D.C. 1998), involving the infamous "Filegate" lawsuit, was persuaded the plaintiffs were not entitled even to restore the deleted files. The plaintiffs sought recovery of all of the deleted files for the past four years from the White House. In coming to its decision, the court relied heavily on the experts’ affidavits as to the cost and time it would take to retrieve deleted files. Alexander, 188 F.R.D at 117. The court was also persuaded by the fact that the plaintiffs could still seek data based on "targeted and appropriately worded searches" that limited the number of individuals and it could depose individuals with knowledge of individuals who might know of alternative ways to restore the files. Id. Although the Alexander case is promising, the fact that it was a high profile case and that the documents were being produced by the White House might help explain the degree of scrutiny applied.[23]

The leading cases on the other end of the spectrum presume that a higher threshold of expense can be born by the requesting party and thereby apply a stricter degree of scrutiny when determining whether to shift costs. See, e.g., Bills v. Kennecott Corp., 108 F.R.D. 459, 463-64 (D.C. Ut. 1985); In re Brand Name Prescription Drugs Antitrust Litigation, 1995 WL 360526, *2 (N.D. Ill. June 15, 1995). The predominant approach to computer database discovery is that it should be as readily provided as are traditional forms of documents. See Bills, 108 F.R.D. at 463-64. The Bills court noted that although a decision to shift costs will be made on a case-by-case basis, "certain propositions will be applicable in virtually all cases, namely that information stored in computers should be as freely discoverable as information not stored in computers, so parties requesting discovery should not be prejudiced thereby; and the party responding is usually in the best and most economical position to call up its own computer stored data." Bills v. Kennecott Corp., 108 F.R.D. 459, 463-64 (D.C. Ut. 1985). Courts that have adopted these "propositions" will likely not shift costs. For instance in In re Brand Name, the court required the responding party to pay for and produce non-privileged, relevant documents from its 30 million pages of e-mail data stored on its back-up tapes. 1995 WL 360526, *1.

Fundamentally, many of these courts appear to be skeptical of defendants’ assertions of expense for creating electronic files. For instance, one court stated it found "in this age of high-technology where much of our information is transmitted by computer and computer disks, it is not unreasonable for the defendant to produce the information on computer disk for the plaintiff." Storch v. IPCO Safety Products Co. of Penn., Inc., 1997 U.S. Dist. LEXIS 10118, *6 (E.D. Penn. 1997; see also Itzenson v. Hartford Life and Accident Insurance Co., 2000 U.S. Dist. LEXIS 14680, *3 (E.D. Penn. October 10, 2000) ("it is difficult to believe in the computer era" that the defendant could not identify files based on specific categories); Daewoo Electronics Co., Ltd. v. U.S., 650 F. Supp. 1003, 1006, 10 C.I.T. 754 (Ct. Int’l. Trade 1986) (court warned the defense that the "complex and unique" nature of the technology is an invalid reason for shifting cost).

In Bills, the responding party argued it should be reimbursed for any additional costs that exceeded the costs accrued from the traditional method of simply providing the propounding party access to the documents. Bills, 108 F.R.D. at 460. The court rejected shifting costs because the defendant did not dispute the relevancy of the documents, it benefited "to some degree" from the production, and the plaintiff was not as capable of bearing the expense. Bills, 108 F.R.D. at 460-63. "The mere fact that the production of computerized data will result in a substantial expense is not a sufficient justification for imposing the costs of production on the requesting party." In re Brand Name Prescription Drugs Antitrust Litigation, 1995 WL 360526, *2 (N.D. Ill. June 15, 1995).

Sharing similar sentiments as the In re Brand Name Prescription Drugs Antitrust Litigation case, a number of decisions appear to blame the responding party for the cost. Despite acknowledging that the retrieval cost of $50,000 to $70,000 was expensive, one court declined to shift the cost to the plaintiffs, "where, as here, the ‘costliness of the discovery procedure involved is … a product of the defendant’s record-keeping scheme over which the plaintiffs have no control.’" In re Brand Name, supra, at *2, quoting Kozlowski v. Sears Roebuck & Co., 73 F.R.D. 73 (D. Mass. 1976) (where responding party argued because its classification was by claim number it was an "impossible task" to go through and find other complaints similar to the tort raised in the litigation); see also Rhone-Poulenc Rorer, Inc., et al. v. The Home Indemnity Co., 1991 U.S. Dist. LEXIS 8304, at *5 -8 (E.D. Penn. June 17, 1991) (an "unwieldy" record-keeping computer and hard copy system does not demonstrate "the most extreme showing of burdensome" required to shift costs); Delozier v. First National Bank of Gatlinburg, 109 F.R.D. 161, 164 (E.D. Tenn. 1986) (ruling defendant should bear expense of mirofilm retrieval process, because the "costliness of the discovery procedure involved is entirely a product of the defendant’s record-keeping scheme"); Danis v. USN Communications, Inc., 2000 U.S. Dist. LEXIS 2000, *151-152 (N.D. Ill. October 23, 2000) (in a shareholder derivative suit, court refused to shift costs incurred by defendant to rebuild database and application software to review back-up tapes at a cost of $159,632 because defendants were to blame for failing to preserve the information in a form that was easily accessible). These decisions seem to carry with them the residue of the obdurate defendant in Kozlowski, and hold corporations’ allegations of discovery expense to a higher level of scrutiny. Moreover, they don’t seem to reflect the fact that the retrieval costs (labor and software) are usually the fault of technology and not the company.

Essentially these courts apply the perspective that technology eases the expense of investigating information, i.e., the end product allows one to search the text for key words. However, this perspective does not account for the software and time it takes to get to that level. Responding parties are essentially in a catch-22. On the one hand, they are hesitant to "open up their computer banks for inspection," and, on the other hand, they are wary of the expense of technicians creating and applying programs to inspect their databases and of attorneys and assistants reviewing the plethora of documents for privilege and relevancy concerns. Bills, 108 F.R.D. at 462. Even after the party has produced a hard copy production, courts have required the responding party to also provide a digital copy at its own expense. See, e.g., Adams v. Dan River Mills, Inc., 54 F.R.D. 220 (W.D. Va. 1972) (requiring defendant provide electronic version of the printouts plaintiff already received, without addressing, nor requiring, a shift in costs); see also Daewoo Electronics Co., Ltd. v. U.S., 650 F. Supp. 1003, 10 C.I.T. 754 (Ct. Int’l. Trade 1986) (requiring hard copy as well as magnetic tape be produced).

But if the defendant refuses to provide the digital material, it faces the possibility of sanctions. See GTFM, Inc., et al. v. WAL-MART, 2000 U.S. Dist. 2000, *6-7 (S.D.N.Y. November 9, 2000) (charging WAL-MART for the expense attorneys generated from being forced to review transactional documents that would have been "readily provided by computer discovery"). However, a court is wary of granting sanctions or making a spoliation ruling except where there is clear evidence of the party’s intent to delete responsive documents. See Procter & Gamble Co. v. Haugen, 179 F.R.D. 622 (D. Ut. 1998). In Haugen, the plaintiff conducted searches of e-mail communications and deleted all files after the searches. The court denied sanctions except to the extent that the plaintiff did not perform the searches or retain the e-mail data with respect to five identified individuals who plaintiff had previously identified as persons having relevant information for the litigation. See Procter & Gamble Co. v. Haugen, 179 F.R.D. at 632.

Absent a finding of intentional expunction, it is unlikely a court will sanction the producing party, even where the documents were destroyed during litigation, unless there is evidence that this was the only option for discovering the information. See Danis v. USN Communications, Inc., 2000 U.S. Dist. LEXIS 2000, *129 (N.D. Ill. October 23, 2000) (court rejected sanctions where it found the destroyed digital data did not "prejudice" the plaintiff because it found the same data could be retrieved from other sources, the destruction of data was not intentional, and any financial prejudice was "offset by unnecessary costs that plaintiffs have inflicted on defendants in discovery"); see also NOW v. Cuomo, 1998 WL 395320 (S.D.N.Y. July 14, 1998) (rejecting sanctions for destroyed computer databases where there was no evidence of bad faith nor that plaintiffs would have been prejudiced by the loss, and a deposition of knowledgeable officials could track down paper copies).

A few cases have addressed discovery requests of computer databases created by attorneys. At first blush, one would presume these databases were necessarily protected by the work product doctrine. See Indiana Coal Council v. Nat’l Trust for Historic Preservation in the U.S., 118 F.R.D. 264 (D.D.C. 1988) (holding plaintiff could not gain access to defendant’s legal research resources and findings via a computer assisted legal research system because it was clearly work product doctrine); Santiago v. Miles, 121 F.R.D. 636, 640 (W.D.N.Y. 1988) (precluding discovery of computer printouts containing raw data but that was created for litigation and whose category design was created by attorney, because it would reveal mental impressions); Burroughs v. Barr Laboratories, Inc., 143 F.R.D. 611, 624 (E.D.N.C. 1992) (holding database compilation of documents reflects legal strategy and is not discoverable).

However, at least two courts have been persuaded that the ease of reviewing the files outweighed the mere fact the documents were categorized by an attorney for litigation. In addition, neither court shifted the costs. In Hines v. Windnall, 183 F.R.D. 596 (N.D. Ill. 1998), the court distinguished Fautek and Williams (see section III.A.1.) on the grounds that both of these requests were made in preparation of cross-examination of expert testimony. Hines, 183 F.R.D. at 600-601. Hines involved a dispute over access to scanned data, which had been collected for purposes of litigation. It was a compilation of documents from defendant’s own records and from that which had been produced by the plaintiff. The Hines court found the scanned data was not made for a testifying expert, rather it was made "voluntarily" for the litigation, and the defendant had "unlimited assets" (being the U.S. government), whereas the plaintiff was of "modest means." Id. at 601.

The Hines precedent represents the most serious example of allowing digital discovery without shifting costs. This is true for two reasons: First, because it granted access to the plaintiff for documents that were prepared by an attorney in anticipation of litigation, and second, because it did not require the propounding party pay for any of the expense despite the fact that the files were not made by or for the corporation in the ordinary course of  business.

An additional twist to digital discovery is where parties seek to have on-site access to defendants computer system.[24] In the few cases that have begun to allow on-site inspection, courts have shifted costs. As recognized in Bills, responding parties are reticent of "open[ing] up their computer banks for inspection." Bills, 108 F.R.D. at 462. Aware of these concerns, courts have been more aggressive about protecting confidential and privileged information and have instituted greater control over the procedure than typical exchanges. See Strasser v. Bose Yalamanchi, 669 So. 2d 1142 (Fla. Ct. App. 1996) (finding a likelihood of irreparable harm in allowing plaintiff on-site access to computer database where it was "unrestricted," without proper safeguards or restrictions to minimize harm to computer system or violate privileged or confidential information).

In Simon Property, the court allowed the plaintiff to attempt to recover deleted computer files from the company’s computers. In addition, it also allowed the plaintiff to review the files of the main executives’ home computers. The court ordered only a court-appointed expert could retrieve the data. The court held that although the record was "sparse" as to the degree to which on-site inspection would interfere with defendant’s business or the possible retrieval expense, the plaintiff would be required to pay for the retrieval process. 194 F.R.D. 639, 640.

In another internet site case, Playboy Enterprises, Inc. v. Terri Welles, 60 F. Supp. 2d 1050 (S.D. Cal. 1999), the court held the plaintiff could have access to the defendant’s individual and company’s hard drive after it was revealed that defendant had been deleting her e-mails in the course of ordinary business even after the litigation began. The court in Playboy weighed the benefit and burden of the discovery pursuant to the proportionality elements of Rule 26(c). Playboy Enterprises, 60 F. Supp. 2d at 1053-54. The court allowed the plaintiff to retrieve a "mirror image" of the hard drive at the propounding party’s (plaintiff) expense, pursuant to a protective order.[25] The court held the likelihood of relevant information being recovered outweighed the financial loss from shutting down the business for four to eight hours and any privacy rights. Playboy Enterprises, 60 F. Supp. 2d at 1054.

After giving defendant’s counsel an opportunity to review all documents for privilege, the court in Playboy instituted additional protections: (1) plaintiff’s expert must describe the probability of recovering deleted e-mails before the court would authorize retrieval; (2) only a court-appointed expert could make the mirror image copy; (3) the plaintiff was barred from raising waiver issues to the extent the expert read any privileged e-mails; and (4) the copied disk would remain with the defendant at all times. Id. at 1055.

In conjunction with the cost of retrieving documents is the additional cost of reviewing the documents for attorney-client privilege or work product doctrine concerns. Currently, the majority of courts review whether inadvertent disclosure constitutes waiver of either the work product doctrine or the attorney-client privilege on a case-by-case basis.[26] There are a number of district court cases that hold inadvertent disclosure can never constitute waiver of either the work product doctrine or the attorney-client privilege. See, e.g., Mendenhall v. Barber-Greene Co., 531 F. Supp. 951, 955 (N.D. Ill 1982); Georgetown Manor, Inc. v. Ethan-Allen, Inc., 753 F. Supp. 936, 938 (S.D. Fla. 1991); Corey v. Norman, Hanson & DeTroy, 1999 ME 196, 1999 WL 1261524, at *7 (Me. 1999); Helman v. Murry's Steaks, Inc., 728 F. Supp. 1099, 1104 (D. Del. 1990); Kansas-Nebraska Natural Gas Co., Inc. v. Marathon Oil Co., 109 F.R.D. 12, 21 (D. Neb. 1983). On the other end are courts that align with the "strict accountability" philosophy, automatically finding waiver, without delving into the intention or neglect of the party. See, e.g., Ares-Serono, Inc. v. Organon Int'l B.V., 160 F.R.D. 1, 4 (D. Mass. 1994); Carter v. Gibbs, 909 F.2d 1450, 1451 (Fed. Cir. 1990); In re Sealed Case, 278 U.S. App. D.C. 188, 877 F.2d 976, 980 (D.C. Cir. 1989).

Obviously, the possibility of waiver increases respectively to the amount of documents reviewed. As discussed above, digital discovery has exponentially increased the amount of documents discoverable. Therefore, in the majority of jurisdictions (which includes the case-by-case and strict accountability approach), the chances that inadvertent disclosure will occur through computer based discovery and that a court will find waiver have increased tremendously.[27] Necessarily, a producing party is under an even greater burden to ensure that privileged and confidential documents are not disclosed.

In USA v. Keystone Sanitation Co., Inc., et al., 885 F. Supp. 672 (M.D. Penn. 1994), the court held the attorney-client privilege had been waived due to two e-mails that had been inadvertently disclosed in what was described as a "massive production." The court applied the five-prong test (widely followed) to determine whether the inadvertent disclosure waived the privilege. The test asks the following: (1) the reasonableness of the precautions taken to prevent inadvertent disclosure in view of the extent of the document production; (2) the number of inadvertent disclosures; (3) the extent of the disclosure; (4) any delay and measure taken to rectify the disclosure; and (5) whether the overriding interests of justice would or would not be served by relieving a party of its error." Keystone, 885 F. Supp. at 676; Amgen Inc. v. Hoechst Marion Roussel, Inc., 190 F.R.D. 287, 291 (D. Mass. 2000); In re Sause Brothers Ocean Towing, 144 F.R.D. 111, 115 (Ore. D.C. 1991).

The court in Keystone concluded almost every prong weighed in favor of finding waiver. The following were the reasons, respective to each prong: (1) the party did not "contact the court" after producing the documentation; (2) there was no set, immediate deadline for completion; (3) the defendants did not request extra time to devise a statement of privilege; the disclosure as to the subject matter was "complete"; (4) the fact that there was no substantial delay between date of disclosure and the date defendants sought to rectify the issue was inconsequential; and (5) the e-mails were directly related to the defendants’ liability. Keystone, 885 F. Supp. at 676. One could argue the Keystone decision was based on the importance of the issues in the litigation (defendants were trying to evade paying their share for environmental clean-up). The court went on to order the defendants produce all attorney billing statements and the identities of the attorneys that were related to the subject matter of the e-mails. Id. Regardless of the reasons, Keystone represents the most troubling precedent for the producing party.

In order to avoid a ruling of waiver, the responding party must prove, at a minimum, that it implemented protective measures. If a court finds an absence of "reasonable precautions to preserve the confidentiality" of the documents, then the court will likely find waiver. Ciba-Geigy Corp. v. Sandoz, Ltd., 916 F. Supp. 404 (D.N.J. 1995). In Ciba-Geigy, the defendants produced all documents from a database without conducting a privilege review. The Ciba-Geigy court explained the privilege is waived where the disclosure is a result of "gross negligence." Ciba-Geigy, 916 F. Supp. at 411. In applying the four-prong test (described above), the court held the defendant had waived the privilege.

Defendant argued it did not conduct a privilege review because of former associates’ representations that the documents had been reviewed. The court found this explanation unpersuasive. The court was also unimpressed with defendant’s argument that the protective order immunized a party from such a review. Id. at 412. Finding no time constraints and the relatively small size of the documents produced also pointed in favor of waiver. Id. at 413-14. Ciba-Geigy makes clear, although a protective order and an assertion of the privilege are prerequisites, a party cannot rely solely upon these measures. The Ciba-Geigy case is descriptive of the most blatant type of waiver.

Logically, courts seem more lenient when the parties are required to produce larger amounts and where they actually performed some degree of review for privileged or confidential documents. The leading case against waiver is Transamerica Computer Co. v. IBM, 573 F.2d 646 (9th Cir. 1978). In Transamerica Computer, the Ninth Circuit held the defendant was effectively "compelled" to produce the documents, due to the size (17 million pages) and the demanding, accelerated schedule. Transamerica Computer, 573 F.2d at 650-53. The producing party also did attempt to review the documents for privilege. Transamerica Computer, 573 F.2d at 650-53. However, the court noted this was a "truly exceptional and unique situation." Transamerica Computer, 573 F.2d at 651; see also IBM v. Comdisco, Inc., 1992 Del. Super. LEXIS 255, at *4-5 (June 22, 1992) (disclosure of one privileged document did not constitute waiver, where plaintiff had an "elaborate review process" and it had reviewed and produced millions of pages). The disparity between the Transamerica Computers and Keystone decisions is disconcerting for a responding party who seeks some degree of predictability as to how to protect against waiver while limiting costs. As more information is requested and produced, the producing party will bear an even greater economic burden to prove it implemented proper protections.

The costs of retrieving and reviewing digital documents are amplified by recent circuit court decisions providing third party intervenors access to documents attached to dispositive pretrial motions. The Ninth Circuit is the most recent circuit court that has weighed in on the issue. In San Jose Mercury News, Inc. v. U.S. District Court - Northern District (San Jose), 187 F. 3d 1096 (9th Cir. 1999), the court held the intervenors had a federal common law right to review any prejudgment documents that were attached as exhibits to dispositive motions. San Jose Mercury News, 187 F. 3d at 1102, citing Republic of the Philippines v. Westinghouse Electric Corp., 949 F.2d 653 (3d Cir. 1991); Joy v. North, 692 F.2d 880, 893 (2d Cir. 1982). Declining to decide on First Amendment grounds, the court noted there was a common law presumption that the public had a right to the documents.[28]

Therefore, the party resisting the disclosure had the burden to make a "particularized showing of good cause with respect to any individual document." San Jose Mercury News, 187 F. 3d at 1103. This burden was especially important where a blanket protective order was entered into. Id. Absent a showing that the documents are trade secrets or that disclosure might harm the party’s competitive standing, a court will likely grant access. See Westinghouse, 949 F.2d at 662-63.

The Third Circuit has gone so far as to allow a third party intervenor a common law right of access to all pretrial motions and attached documents of a non-discovery nature. Leucadia, Inc. v. Applied Extrusion Tech., Inc., 998 F.2d 157, 164. Although the Ninth Circuit’s decision decided an intervenor’s right of access only with regard to "dispositive motions," the distinction is minimal. This precedent adds an additional concern, and therefore expense, for practitioners when reviewing documents. It is possible that opponents could attach documents of slight relevance, for purposes of the motion or the case, to publicize the document, e.g., a particularly embarrassing e-mail, solely for harassment purposes. Litigants must face the increasing possibility that non-relevant, embarrassing documents will be disclosed due to the sheer number of digital files available. The repercussions of this disclosure being made public add an additional caution to protect confidential data produced in electronic discovery.

The fundamental issue surrounding the debate to shift costs is the silencing effect any modification might have on certain litigants. However, the cases discussed above demonstrate the astonishing expense the producing party can incur. The most obvious way to appease these concerns while at the same time trying to reduce the transactional costs of digital discovery could be by shifting costs at the time of judgment. Similar to deposition costs that are shifted at the end of trial, the losing party would be required to reimburse the other side for all or part of discovery costs.[29] However, the court would have discretion to determine the amount that would be shifted based on the economic resources of the party, the expense incurred, and the degree to which the requests were frivolous or irrelevant. At the same time, the proportionality test would still be applied during discovery. This solution would provide for an additional balancing of hardship and resources after trial and a determination on the merits.

The benefits are multi-fold. Such a rule would prevent abusive and overbroad requests. It would force the propounding party to particularize the requests and seek only truly relevant information. This requirement would curb all parties, regardless of their resources, from waging a war of attrition. Such a solution provides direct incentive, for all, to self-regulate the scope and content of the requests, likely making the process more efficient and perhaps less contentious. Self-regulation would also prevent "fishing expeditions" without the need for the court to make conjectures as to the parties’ true intentions. And thus, it would likely decrease the need for court intervention overall. This solution would also provide an incentive for the weaker side to settle rather than risk additional costs besides simply losing at trial. At the same time, it would not preclude litigants with fewer resources from discovery, as a court could still require the producing party pay for the production up front.

The transactional costs may increase with this rule. Corporate litigants may be more willing to go to trial in hopes of recouping some of their expenses, and trial costs would extinguish any savings from discovery expenses. While the solution provides for a greater incentive to settle, it would only be after discovery, which would marginalize any economic benefit, absent that already provided by the proportionality rule. Regardless, the benefit of increasing settlement would not deter the cost of defending frivolous lawsuits.

While this option is not perfect, it offers the most responsive compromise between the current disparate positions. By providing a bright-line rule, all the parties will be required to substantively evaluate the documents they need. Necessarily, wasteful expenses will be reduced. At the same time, no party will be prevented from continuing with a non-frivolous claim. While no solution provides for a perfect result, the degree to which discovery has gotten out-of-hand demands rules that do not presently exist.

1. Arenson, Gregory K., Fleming, Thomas F. and McGanney, Thomas, Does Discovery of Electronic Information Require Amendments to the Federal Rules of Civil Procedure?, New York State Bar Association, Commercial and Federal Litigation Section, Committee on Federal Procedure Publication (February 22, 2001). Posted at ‹http://www.kenwithers.com/articles/rules_debate/›. [back]

2. See Giacobbe, 57 Wash. & Lee. L. Rev. 257, 268 & 268 n. 73. ("Lawrence Aragon, E-Mail Is Not Beyond the Law, PC Wk., Oct. 6, 1997, 111 (discussing instance in which defendant chose to settle rather than to incur enormous electronic data discovery costs); Karen L. Hagberg & A. Max Olson, Shadow Data, E-Mail Play a Key Role in Discovery, Trial, N.Y. L.J., June 16, 1997, at S3 (discussing problem involving plaintiffs using discovery rules to harass defendants); James J. Marcellino & Anthony A. Bongiorno, E-Mail Is the Hottest Topic in Discovery Disputes: One Litigant Seeks Facts Buried in a Data Base; the Other Seeks to Avoid Burdens of Production, Nat'l L.J., Nov. 3, 1997, at B10 (discussing potential for abuse of discovery rules involving electronically stored data requests); Janet Novack, Control/Alt/ Discover, Forbes, Jan. 13, 1997, at 60 (referring to use of cost of electronic data discovery to force settlement as "blackmail"); Geanne Rosenberg, Electronic Discovery Proves Effective Legal Weapon, J. Rec., Apr. 27, 1997, available in 1997 WL 14390671 (discussing use of electronically stored data discovery requests as negotiation tool). Aragon cites one instance in which a corporate defendant faced an electronic data discovery request that cost between $500,000 and $750,000. Aragon, supra, note 41, at 111. The defendant was not as concerned with whether or not the requested backup tapes contained detrimental data as it was with the enormous burden of complying with such a discovery request. Id. Ultimately, "[t]he huge tab weighed on the company's decision to settle the case"). [back]

3. See, e.g., Scheindlin, supra, 41 B.C.L.R. at 351. [back]

4. See Withers, Kenneth J., Computer-Based Discovery in Federal Civil Litigation, 2000 Fed. Cts. L. Rev. 2, *II.B.2 (2000), posted at ‹http://www.fclr.org/2000fedctslrev2.htm›, noting paper-based record keeping systems manage documents in "business-record order." [back]

5. See Withers, 2000 Fed. Cts. L. Rev. 2, *I.2 (2000), citing Johnson, Gregory S., A Practitioner's Overview of Digital Discovery, 33 Gonz. L. R. 347, 360 (1998). [back]

6. See Withers, 2000 Fed. Cts. L. Rev. 2, *II.B.2, summarizing a recent ABA and PricewaterhouseCoopers survey that found only 11.5% of the corporate clients interviewed had an electronic data classification mechanism that allows one to quickly locate data relevant to a particular area of litigation. It also quoted Prof. David Wallace at the University of Michigan School of Information who stated, "developing an electronic record-keeping system is 'the single most important priority' for the record-keeping professions." [back]

7. For a more detailed description of the types of electronic documents, see Ruanne, William J. and Lehman, James K., Running the Gauntlet: Responding to Discovery of Electronic Documents, 10-2 (February 22, 2001). [back]

8. See Giacobbe, Corinne, Note: Allocating Discovery Costs in the Computer Age, 57 Wash. & Lee. L. Rev. 257 (2000), and articles cited therein. The Note offers a fairly comprehensive explanation for how much and why retrieval costs are so high. For example, on page 265, the Note states one electronic document retrieval company's estimate that the retrieval of twelve monthly backup tapes would cost on average $100,000. [back]

9. See Schiendlin, 41 B.C.L.R. 327, 365, discussing how this definitional question has not been challenged or addressed in the courts. [back]

10. See Simon Property Group, L.P. v. mySimon, Inc., 194 F.R.D. 639, 640 (S.D. Ind. 2000) (holding deleted documents are discoverable under rule 34). [back]

11. See Dryer, Anthony J., When the Postman beeps twice: The Admissibility of Electronic Mail under the Business Records Exception of the Federal Rules of Evidence, 64 Fordham L. Rev. 2285, 2291 (1996), noting the projected 40 million e-mail users for the year 2000 were expected to produce 60 billion messages annually. And see Withers, 2000 Fed. Cts. L. Rev. 2, *II. C. 1. n.22, referring to one research company's estimate that a 100-employee company will "accumulate 211,200 messages annually, not counting copies or backups." [back]

12. See, e.g., Llewellyn, Virginia, Discovery the E-Way, Texas Lawyer, Vol. 16, No. 47 (January 29, 2001), Scheindlin, Hon. Shira A. and Rabkin, Jeffrey, Electronic Discovery in Federal Civil Litigation: Is Rule 34 Up to the Task?, 41 B.C.L.R. 327, 339 (2000). [back]

13. See Olmsted, Betty Ann, Electronic Media: Management and Litigation Issues When "Delete" Doesn't Mean Delete, 63 Def. Couns. J. 523, 525 (1996) (describing the type of cases that have arisen from e-mail discovery, including "defamatory remarks, obscene and indecent language, electronic stalking, racial discrimination or harassment, copyright infringement, conspiracy, retaliation for whistle-blowing"). [back]

14. See Withers, 2000 Fed. Cts. L. Rev. 2, II.C.1. [back]

15. See Llewllyn, supra, Texas Lawyer (stating electronic review not only allows one to discover information not possible through a paper trail (e.g. when the document was created and then modified) but it can do it cheaper by preventing duplicate documents being produced, lawyers and staff reviewing the same documents, and human error); see also Withers, 2000 Fed. Cts. L. Rev. 2. [back]

16. See Marcus, Richard L., Confronting the Future: Coping with Discovery of Electronic Material, 64 Law and Contemporary Problems 101 (2001), for a comprehensive discussion of the possibility of costs decreasing. [back]

17. FED. R. CIV. P. 26(b)(2) & (b)(2)(iii). [back]

18. Advisory Committee Notes, 1983 Amendment. [back]

19. See Section III discussing the cases that did authorize cost-shifting; see also Wright, Miller, and Marcus, Federal Practice and Procedure: Civil 2d, ch. 6 §.2008.1 [back]

20. Arenson, Gregory K., Fleming, Thomas F., and McGanney, Thomas, Does Discovery of Electronic Information Require Amendments to the Federal Rules of Civil Procedure?, New York State Bar Association, Commercial and Federal Litigation Section, Committee on Federal Procedure Publication (February 22, 2001), arguing the balancing test in the federal rules is fully adequate. Posted at ‹http://www.kenwithers.com/articles/rules_debate/›. [back]

21. It should be pointed out the Oppenheimer court was reviewing a trial court's decision ordering the defendant to pay for and provide a computerized list of potential class members for the plaintiff's suit. The Supreme Court noted establishing class certification was plaintiff's burden, pursuant to Rule 23. Thus, the court was not deciding whether to shift costs based on the undue burden test under Rule 26, although the court did make reference to its language. [back]

22. For another alternative see Danis v. USN Communications, Inc., 2000 US Dist. LEXIS 2000, *151-152 (N.D. Ill. October 23, 2000). In Danis, the court held it would have shifted costs after production if it were determined that the plaintiff failed to copy or use the documents produced. However, because defendant was equally abusive throughout the discovery process, the court decided against any shift in costs. [back]

23. As a caveat, one court granted plaintiff's request that the documents produced be on disk, rather than hard copy. Storch v. IPCO Safety Products Co. of Penn., Inc., 1997 U.S. Dist. LEXIS 10118, *6 (E.D. Penn. 1997). The court based its decision on the fact that plaintiff had established it would incur significant expense from the encoding fees that would be incurred if it had to reenter the data into the computer. Storch, 1997 U.S. Dist. LEXIS 10118, *6. This is an option for easing the expense for both sides. [back]

24. See Withers, 2000 Fed. Cts. L. Rev. 2, G, discussing the hurdles of on-site inspection as an alternative to producing tapes and/or hard copies. [back]

25. One could argue the Playboy decision was based on the fact that the plaintiff was a corporation that appeared willing to pay for the expense. [back]

26. The same standards are applied to determine waiver for the attorney-client privilege as are those for work product doctrine. See, e.g., In re Sause Brothers Ocean Towing, 144 F.R.D. 111, 114 (Ore. D.C. 1991). [back]

27. For a more extensive discussion of digital discovery and inadvertent waiver, see Comment: Making A Wrong Turn On The Information Superhighway: Electronic Mail, The Attorney-Client Privilege And Inadvertent Disclosure, 26 Cap. U.L. Rev. 347 (1997); Inadvertent Disclosure Of Privileged Information And The Law Of Mistake: Using Substantive Legal Principles To Guide Ethical Decision Making, 48 Emory L.J. 1255 (1999); Note: E-Mail: The Attorney-Client Privilege Applied, 66 Geo. Wash. L. Rev. 624 (1998); Comment: E-Mail and the Attorney-Client Privilege: Simple E-Mail in Confidence, 59 La. L. Rev. 935 (1999); Comment: The High-Tech Legal Practice: Attorney-Client Communications And The Internet, 69 U. Colo. L. Rev. 851 (1998). [back]

28. But see Sun Systems v. Microsoft Corp., Northern District Court of California (San Jose), Case No. C 97-20884 RMW (PVT), Order Amending Stipulated Protective Order, filed August 10, 1998, where the district court granted San Jose Mercury News and others access to documents under seal pursuant to a protective order based on First Amendment grounds. [back]