New York City.

Editor’s Note: This is the first in a two part series.

In Winfield v. New York, 15-CV-05236 (S.D.N.Y. Nov. 27, 2017), Magistrate Judge Katharine H. Parker ruled on the plaintiffs’ claims that the defendant did not properly produce e-discovery. The opinion is, paradoxically, interesting because it is boringEAs for the latter reason, readers will recall that I have written numerous columns about opinions the government has been obtaining, without making or showing probable cause, records of cellphone tower interactions with a user’s phone, most importantly, the tracking of a person’s movements by obtaining records for cell carriers as to when a user’s phone “pinged” various cellphone towers (simply “checked in” with the carrier at the closest tower so that if the carrier needed to route a call to the phone it would know where to send it), which records will provide a map of the user’s movements. The legal consequences of obtaining such records absent a showing of probable cause were unknown, of course, prior to cell phone towers.

The initial legal reaction was to deny defense attempts to suppress such evidence, reasoning that such tracking simply tracked the public movements of a user, i.e., something in which the user had no reasonable expectation of privacy. However, with the “Big Brother” aspects of such tracking having set in, some circuits having sided with defendants, and the Supreme Court recently having heard oral argument in such a matter, we see how new technology can create a new, interesting issue.

It is, then, heartening when an e-discovery opinion is not interesting, since that may indicate that more and more readers are familiar with both the technology and the legal issues in which e-ddiscovery is involved. We see that here, where the technical issues are identified but not discussed at length and where the legal issues are typical of discovery matters.


The matter involves the city of New York’s goal “to build or preserve 200,000 below-market-rate apartments by 2024 as part of its affordable housing programs.” Part of the city’s affordable housing programs, the “community preference policy” (CPP) allows city residents who want affordable housing and are otherwise qualified (including having a sufficient income) to “apply for certain affordable housing units through a lottery system.” Far more people, of course, apply than there are units available. The CPP at issue applies only to some of the housing lotteries. The policy sets aside half of the affordable housing units to be distributed through the lottery for people who live in the “community district” (59 in the city) in which the housing is located.

The plaintiffs are three African-American city residents who applied for affordable housing through lotteries, but were not selected to be interviewed for affordable housing developments. Their claim is that the CPP has a disparate impact on African-American and Latino applicants in “neighborhoods of opportunity” because their current residents are, predominantly, white. Application of the CPP perpetuates racial segregation and constitutes intentional discrimination in violation of the federal Fair Housing Act.

Background Discovery Issues

The parties’ initial disagreement was over what search terms should have been used in connection with the city’s review of electronic documents collected from five custodians from the Department of City Planning (DCP) and Steven Banks, the commissioner of the City Human Resources Administration/Department of Social Services (HRA) and head of the Department of Homeless Services (DHS). The plaintiffs’ complaints grew to include alleged deficiencies in the city’s electronic document review process, including “the scope of the city’s review of documents gathered from DCP and Banks, and the city’s alleged over-designation of documents in the prior review populations as non-responsive, which the plaintiffs claim has affected the reliability of the city’s predictive coding processes as a whole.” In order to judge the fairness of the defendant’s procedures in producing e-discovery, the plaintiffs sought, among other things, an order directing the city to provide the plaintiffs with samples of nonprivileged documents collected from the Department of Housing Preservation & Development (HPD), the mayor’s office, DCP, and Banks that the city has designated as “nonresponsive” in its review. Having reviewed the city’s in camera submissions, as well as sample documents provided by the plaintiffs, the court granted and denied in part the plaintiffs’ application.

HPD promotes both development of new affordable housing and preservation and renovation of existing housing that contains affordable housing units. One means of promoting affordable housing units is through a lottery. The CPP applies to only some of the lotteries and affordable housing units that become available in the city. In addition to HPD’s efforts, certain city housing developments “have been incentivized by tax exemptions provided under Section 421-a” of New York’s Real Property Tax Law (421-a developments). Under city policy, “50 percent of the community preference units in certain 421-a developments have been set aside for homeless residents of the city whose shelter or last-known address is in the community district where the 421-a development is built.”

The plaintiffs’ complaint alleges that the city has been and continues to be characterized by extensive residential segregation on the basis of race, ethnicity and national origin in all 59 community districts. While such segregation “can be traced to historical restrictions and intentional discrimination by all categories of actors in the housing market, including government entities, developers, landlords and others,” the CCP has had a disparate impact on African Americans and Latinos because those historical factors make some neighborhoods “higher opportunity” and African Americans and Latinos “do not have an equal opportunity to compete for affordable housing” in all neighborhoods, particularly the “higher opportunity” ones, and the CPP has ignored this reality.

The city denied the allegations of discrimination and claimed it was “deeply committed to fair housing and the creation of affordable housing.” The city moved to dismiss the complaint for lack of standing and failure to state a claim. In denying that motion, the court (not the instant magistrate) noted that discovery in the matter was highly important, as it could show that removal of the CPP “would have no impact on the racial and ethnic makeup of eligible affordable housing applicants (and ultimate recipients of housing) whether local residents within a community district are given a preference or not.”

Initially, discovery focused on “production of data from the affordable housing lotteries needed for disparate impact analysis, depositions of key city witnesses, and production of documents concerning the community preference policy and the reasons for its promulgation and amendments, as well as documents concerning the city’s defenses, including community resistance to affordable housing projects.” The plaintiffs sought, inter alia, “documents related to affordable housing policies more generally, city policies related to zoning or homelessness, interactions with the U.S. Department of Housing and Urban Development (HUD), and strategies for mitigating displacement of individuals from gentrifying communities.” The court, during a conference with counsel, “narrowed a number of the plaintiffs’ discovery requests, including by topical scope and the number of custodians that the city needed to search in response to 23 of the plaintiffs’ document requests.” Finally, the parties also disagreed as to the custodians from whom documents would be collected. After numerous meet and confers and rulings, the court ordered the city to collect documents, including electronic documents, from 50 custodians. These included 44 individuals who worked within HPD and the mayor’s office, five individuals who worked in DCP, and Banks.

The city began its document review process with HPD and the mayor’s office. As it pertained to ESI, the parties disagreed as to the search terms that would be applied to the electronic documents collected from these custodians. Ultimately, after considerable negotiations between the parties and judicial guidance, the city applied a set of search terms. The pace of the city’s first searching for files which “hit” on search terms and then using review experts to examine each “hit” to determine whether it was relevant and not privileged, however, led to the plaintiffs lodging “numerous complaints” which, in turn, led the court to direct the city complete the review by “using Technology Assisted Review (TAR) software (also commonly referred to as ‘predictive coding’) to hasten the identification, review, and production of documents responsive to the plaintiffs’ document requests.” The court noted that Rio Tinto v. Vale, No. 14-cv-3042 (RMB) (S.D.N.Y. Mar. 2, 2015) and Da Silva Moore v. Publicis Groupe, 287 F.R.D. 182 (S.D.N.Y. 2012) supported TAR as having “been shown to produce more accurate results than manual review.” The court explained that for “TAR to work properly, the producing party must prepare a training, or seed set, of responsive and nonresponsive documents to train the computer system how to distinguish between them,” and that the city had done so.

The city next reviewed ESI gathered from the DCP custodians and Banks. Because at a discovery conference the city “explained that it had applied search terms to the DCP/Banks review population that were targeted towards only a subset of the plaintiffs’ requests for production that the court had ordered the city to search,” and that “it was unaware of whether the plaintiffs had been provided with the list of the search terms,” the court “directed the city to provide the search terms it used to the plaintiffs” and further ordered that, if the plaintiffs believed additional search terms should be applied to these custodians, they should “propose the additional terms to the city within a week.” The city’s list of search terms proved to be “far more limited than the terms used for HPD and the mayor’s office,” leading the plaintiffs to object as well as to propose “over 800 additional search terms.” The city, in turn, opposed the plaintiffs’ request, “contending that it would require the city to review approximately 90,000 additional documents” and emphasizing that the requested searches would cost $248,000 in addition to the $350,000 which the city had already spent to review the 100,000 document already examined.

Further negotiations led to the city to propose acceptance of all of the plaintiffs’ “modified proposed terms,” which the city would, in turn, use along with predictive coding. The plaintiffs objected to that proposal, asserting that the city’s application of “an impermissibly narrow view of responsiveness during its review process” led to the “over-designation” of documents as privileged and non-responsive, and thus to “improperly trained” TAR software used by the city that was “unable to recognize documents” that were truly responsive.

To address the plaintiffs’ concerns, the court issued several orders, requiring the city “to submit a letter for in camera review describing its predictive coding process and training for document reviewers,” ordering briefing on the “plaintiffs’ challenges to the city’s privilege designations on documents and in depositions,” requiring the city to provide a privilege log “for a sample set of 80 documents that the city designated as privileged in its initial review,” and requiring “submission of deposition transcripts and a privilege log concerning privilege objections made during the depositions.” The city withdrew its privileged designation of 51 of the 80 documents and, but the court nevertheless ordered the city to submit all 80 documents and detailed privilege log.

The court further ordered the city to respond to a complaint of the plaintiffs’ that the city had improperly designated documents as nonresponsive because they were “slip-sheet” documents, even though the plaintiffs could still view the extracted text. After much back and forth, the city provided its explanation and conceded that these slip-sheeted documents were “arguably responsive.”

Next week, I’ll discuss the court’s lengthy and considered response and other current e-discovery issues.

Leonard Deutchman is vice president, legal for KrolLDiscovery, which he helped build into the largest e-discovery provider in the United States. Before joining KrolLDiscovery, he was a chief assistant district attorney at the Philadelphia District Attorney’s Office, where he founded the Cyber Crime Unit and conducted and oversaw hundreds of long-term investigations involving cybercrime, fraud, drug trafficking and other offenses.