Stipulated motions can be more troubling than they seem. The cooperation that leads to stipulations is a beguiling reprieve from the combativeness of litigation. Parties constantly file contested motions raising vexing questions that require judges to pick winners and losers. In contrast, stipulated motions appear simple and enable win-win outcomes. Busy judges might be tempted to grant these stipulated motions without much thought. Minimal scrutiny is tolerable if an order truly is win-win for everyone who the order affects. But if the order has collateral consequences, then a win for the parties can impose a troubling loss on the public. Rulemakers recognize that private and public interests do not always align and craft procedures to manage the incongruity. For example, Federal Rule of Civil Procedure 26(c)(1) requires judges to police self-interested parties seeking protective orders. Even if the parties agree that information obtained though discovery should remain private, judges must find “good cause” before shielding information from the public. The rule gives judges discretion to protect trade secrets and personal data while enabling disclosure of threats to public health and safety.
Secrecy by Stipulation reveals that the “good cause” requirement often fails to prevent dubious protective orders. When parties stipulate that secrecy will facilitate discovery, courts are reluctant to disrupt the agreement. The authors—Nora Freeman Engstrom, David Freeman Engstrom, Jonah B. Gelbach, Austin Peters, and Aaron Schaffer-Neitz—acknowledge that prior scholarship has considered the potential harms of protective orders and the need for careful judicial scrutiny. Many commentators believe that judges are side-stepping the “good cause” standard. However, a lack of data has hindered efforts to assess the frequency of stipulated protective orders and the intensity of judicial review. Critics of secrecy have relied on anecdotal accounts and empirical studies with limited scope. This evidence has not convinced rulemakers that judges are departing from the requirements of Rule 26(c)(1). Proposed reforms therefore have failed to gain traction.
The new article responds to the impasse by providing better evidence of how Rule 26(c)(1) operates in practice. The authors employ cutting-edge empirical methods to show that stipulated protective orders are more common and less scrutinized than rulemakers assume.
Measuring the prevalence of stipulated protective orders is difficult. Researchers must parse docket sheets to find orders with a range of designations. Moreover, the frequency of orders may vary by time, place, and subject matter. Scholars hoping to paint a complete picture must therefore analyze a massive sample of cases. Unsurprisingly, no one has completed this labor-intensive inquiry until now.
Modern technology makes the task of identifying stipulated protective orders less labor-intensive than manual coding techniques. The authors cleverly combine text processing and machine learning to extract information about protective orders from more than two million docket sheets in civil cases filed between 2005 and 2012. These techniques reveal the frequency of stipulated protective orders with far more precision than prior studies. The authors explain that the passage of time since 2012 should not undermine the current relevance of their findings.
The study reveals that at least one party files a motion for a protective order in roughly 8.5% of the cases that survive long enough for defendants to file an answer. Roughly 45% of these motions are stipulated. Despite the “good cause” requirement, judges grant approximately 96% of stipulated motions. More than half the judges who reviewed at least twenty-five motions for stipulated protective orders granted every motion in its entirety. The authors contend that this 100% grant rate is evidence that “searching case-by-case scrutiny is not the norm.” They reinforce this finding of “rubber-stamping” by qualitatively reviewing four hundred cases from the sample. Most of the orders granting stipulated motions did not mention “good cause” and did not provide any case-specific analysis. In contrast, judges who denied stipulated motions often found glaring errors. The authors “identified many of these same errors” in their review of granted motions. The prevalence of errors in granted motions suggests that courts are not rigorously applying the “good cause” standard.
Of course, data does not have self-evident implications. A high grant rate for stipulated protective orders is troubling only if judges should have denied some of the granted motions. Determining whether motions should have been denied requires a normative framework for analyzing judicial secrecy. The article therefore does more than present the results of an innovative empirical study. The authors situate their study within a broader context of debates about the purpose of civil adjudication. If adjudication should promote public interests in addition to private interests, then secrecy may be undesirable. Secrecy is especially troubling under an emerging theory positing that the tort system should aspire to transfer privately held information about wrongdoing into the public domain. This transfer would enable government officials to implement regulatory objectives more effectively. The authors hope their empirical findings will invigorate conversations about the information-disclosing function of adjudication.
Secrecy by Stipulation should generate a wide-ranging discussion. The authors have provided a sobering reminder that the FRCP’s general commands can be ineffective when judges lack the benefit of adversarial briefing. The article also highlights how secrecy minimizes the ripple effects of private discovery. Cabining these effects can undermine both private and public enforcement of health and safety regulations. In addition, the authors’ creative integration of machine learning and qualitative analysis can shape future empirical studies of judicial behavior. Millions of docket sheets are ready to be mined by creative researchers. The article is therefore notable for its descriptive findings, normative implications, and empirical methods.
