DOJ Approves Use of Predictive Coding in a Proposed Merger Document Review

The Department of Justice approved the use of predictive coding for a review involving over one million documents in the recent proposed merger between Anheuser-Busch InBev and Grupo Modelo.  The DOJ arrived at a settlement with Anheuser-Busch InBev and Grupo Modelo in April 2013.

The DOJ requested documents from Constellation Brands Inc. and Crown Imports LLC. Constellation was the potential buyer of assets to be sold in the transaction. Crown Imports LLC was the joint venture between Modelo and Constellation.  Upon the DOJ’s approval, the companies’ attorneys reviewed the seed set of documents on which the computer trained to identify responsive documents.  They continued their review until they and the DOJ were confident that predictive coding (also known as technology assisted review) would make the correct responsiveness calls on the remaining documents. As a result, Constellation and Crown Imports LLC produced a multitude of thousands of documents to the DOJ.

In a Wall Street Journal article, counsel for Constellation and Crown Imports asserted that his clients “spent 50% less than they would have using more traditional methods . . . .  ‘Something that would easily cost three, four, five million dollars, you can do in the range of one to two . . . .’”  Predictive coding also made the review more efficient.

A spokesperson for the DOJ commented that the “antitrust division ‘has worked with parties who choose to use this new technology in complying with the division’s civil investigative requests.’”  In a February 2013 DOJ antitrust presentation by Renata B. Hesse, Deputy Assistant Attorney General of the Antitrust Division of the DOJ; Ms. Hesse commented that the DOJ has allowed technology assisted review in certain instances “to help streamline [their] process.”  She stated, “When it works well, predictive coding reduces the document review and production burden on parties while still providing the division with the documents it needs to fairly and fully analyze transactions and conduct.”  She cautioned, though, that for “predictive coding to work . . . the division . . . require[s] a high degree of cooperation and transparency about the implementation and structure of the predictive coding process.”