Antitrust agencies routinely require the disclosure of thousands (or more) of internal documents, including those produced or received by senior executives. Merging parties should therefore anticipate the need for substantial document productions when planning their transactions.

Whilst the many practical issues may seem entirely process-driven, document production can have a significant impact on the substantive prospects of the case. Agencies increasingly use ordinary course and deal-specific documents as a core piece of evidence when assessing a transaction, so they can often be pivotal to whether a transaction gains approval.

At the same time, document requirements are not uniform across jurisdictions, and handling multiple large-scale productions can affect the transaction timetable.

In this post, we have teamed up with colleagues Laura Kippin (Managing Director) and Andrew Johnston (Senior Managing Director) at FTI Consulting to highlight our top five issues to consider when navigating the complexities of document production in parallel merger processes. Planning this work stream ahead of engagement with the regulators may help to avoid procedural pitfalls and minimize unnecessary delays to a transaction timetable.

1. What types of data need to be disclosed?

Agencies are increasingly focused on the evolving types of data being created within companies. Requests are becoming more intrusive, going beyond routine document and email searches to include communications via chat platforms such as Teams, Slack, WhatsApp, Signal, SMS messages and other ephemeral communication applications.

Another important focus area is “modern attachments”, which are links to collaborative documents in emails or chat platforms. During collection, the attachment itself is not collected. In order to identify and collect this data, additional custom work is required to determine where the data resides and ultimately gather the data. This can result in additional back-end processing and iterative requests from agencies to track down unproduced documents from specific links they see in other documents, and will need to be factored into the production timeline.

2. How do you search for responsive documents?

Agencies’ willingness to accept the parties’ methodology for document collection and review differs considerably. Some agencies will negotiate search terms ahead of the document collection, while others do not agree upfront on a set of search terms and instead request incremental search terms to be run, generating additional and follow-up productions, as their review progresses. This iterative process is burdensome for the parties, who need to organize document collections and productions of the identified custodians on a rolling basis, usually under significant time pressure.

In addition, search terms frequently yield large volumes of irrelevant documents, burdening reviewers with extensive manual review, and leading to longer review and production timelines. Technology-assisted review models – known as TAR – enable a faster and more efficient review of documents, as a small set of documents coded by attorneys is extrapolated by the technology to a larger set as a one-off, or the ongoing coding by the reviewers is used to develop an evolving TAR model. In the US, FTC and DOJ practices are converging and parties typically must provide metrics on the underlying TAR model and sample null sets (documents the TAR model deems non-responsive) for agency staff review. There will be circumstances where other global agencies will accept TAR, especially where the process has already been completed in the US, but otherwise TAR-based production may need to be topped-off with manual productions based on search terms/key words.

3. Beware diverging attorney-client privilege rules

When producing documents to the agencies, the merging parties have the right to withhold or redact parts of documents that benefit from legal professional privilege. The definition and scope of privilege varies across jurisdictions, adding a layer of complexity to the document review and production process for cross-border transactions.

For instance, most – if not all – jurisdictions recognize that attorney-client privilege protects certain communications between a client and an attorney in private practice. However, EU law – unlike US/UK law – does not recognize privilege over communications with in-house attorneys. In practice, this means that they could be producible before the EC, but withheld from the CMA or the US agencies. Similarly, EU law – again unlike US/UK law – does not recognize that disclosure of privileged material to a third party sometimes does not constitute a waiver of privilege (so-called “common interest privilege”).

To explain whether and why any document has been withheld or redacted for privilege reasons, the parties generally have to submit “privilege logs”, including information such as document production ID (“bates numbers”), document authors and recipients, document types, and the basis for claiming privilege.

These logs are typically supplied to the regulators and, to avoid inadvertent waivers of privilege, parties should put in place privilege log protocols to ensure that documents that must be disclosed to one agency under a narrower interpretation of privilege are not disclosed to another agency with a broader interpretation of privilege. Further, if documents being prepared for production in the US are likely to be requested later in the process by the EC, for efficiency merging parties may want to consider adopting a review protocol that addresses coding for EU privilege on the first pass.

4. Where should the data be hosted?

Different countries have varying legal frameworks for data protection and privacy, which affects the consideration of where to host client data. For example, some EU countries require that data stored in folders labelled “Private” is handled and searched separately as this is deemed to be private data and will likely be added to a Privacy log. Alternatively, some regulations mandate that data cannot leave a specific jurisdiction, which affects options on where the data can be hosted. Even in some US-based cases, where custodian data must be hosted in the UK/EU, the entire data set across all custodians is hosted in the UK/EU to have the data efficiently centralized in one platform. This avoids complications that can emerge when using multiple hosting platforms for a single matter. For example, deduplication, the use of TAR (see above), and visual analytics are harder to execute across the entire dataset or across platforms.

5. The final production: Be aware of different production specifications

Once a final set of responsive documents has been identified, it will need to be prepared and transferred to the agencies. Here too, production specifications vary between agencies. Some will accept the use of FTPs to transfer a production, while others still require a hard drive to be delivered.

Agencies also require documents in different formats for their review. For instance, some agencies will only accept PDF files for all the producible documents, while others require the documents in native and/or image-format (where redactions are applied). Even if a document universe is properly coded and ready for production, merging parties should factor in the timing implications of applying different production specifications to prepare the submission to another agency.