It Takes Two:
Modernizing the Merger Review Process
Good morning, and thank you for inviting me to speak here today. It is a pleasure to be at Georgetown University for this year’s Global Antitrust Enforcement Symposium.
Events like these, which bring competition enforcement officials together to speak with members of the private bar, the business community, and the academic community, serve an important role in the continued development of antitrust law and its enforcement. They facilitate the rule of law by increasing transparency and predictability in enforcement. This, in turn, promotes competition and conserves both public and private resources by allowing the private bar to better counsel their clients so they can avoid anticompetitive behavior. Such events also allow competition agencies yet another opportunity to understand concerns that the business community may have about antitrust enforcement.
Today, I am going to address one of those concerns. My remarks will focus on modernizing the merger review process.
First, I will address the fact that mergers increasingly take longer to review and clear. Many think that is a problem. I do too, and I will explain why.
Second, I will discuss a series of changes that we are making at the Antitrust Division to address this problem and expedite the merger review process.
The Division is committed to achieving this result whenever possible without compromising the quality of merger reviews, but we cannot do it on our own. The merging parties and the Division both exert significant control over the pace and timing of compliance. It is important to recognize that shortening any investigation requires the cooperation of the merging parties. We need you to submit the necessary information promptly and work with us in good faith to resolve open issues expeditiously. It also means that parties shouldn’t hide the eight ball!
Why do I believe that the merger review process needs to be modernized?
There is widespread agreement that significant merger reviews are taking longer to complete. According to one source, in calendar year 2017, significant merger reviews conducted by U.S. antitrust enforcers took an average of 10.8 months to resolve. That’s up from an average of 7.1 months in 2013, which is a 65% increase, according to this source.
Several factors likely have contributed to this increase, some of which are beyond our control.
First, in this electronic age, merging parties frequently maintain enormous quantities of data and documents. It takes longer for the parties to produce them, and it takes longer for enforcement agencies to analyze them.
Second, more and more transactions are international in scope. Necessary coordination with our foreign counterparts — while beneficial— could add time.
Third, when divestitures are required to protect competition and remedy anticompetitive elements to transactions, we increasingly require upfront buyers that are pre-approved before consent decrees can be filed. That also adds time.
So the question is, what can we do, despite these factors, to increase the efficiency of merger reviews?
Nobody wins with unduly lengthy reviews — other than lawyers and expert economists, perhaps. The Antitrust Division’s mission is to protect competition for the benefit of American consumers. That means ensuring that pro-competitive mergers are not unduly delayed by our review process. Indeed, most mergers are pro-competitive, or at least competitively neutral.
As far back as 1982, the Horizontal Merger Guidelines reflected this idea, stating that, “While challenging competitively harmful mergers, the Department [should seek] to avoid unnecessary interference with the larger universe of mergers that are either competitively beneficial or neutral.”
Shortly after the guidelines were issued, then-Assistant Attorney General William Baxter elaborated on this idea. He faced criticism that the 1982 Merger Guidelines were too clear and provided too much guidance. This, the argument ran, would “encourage mergers right up to the line” drawn in the Guidelines. Baxter noted that this criticism “exhibit[ed] an inherent hostility to the phenomenon of mergers themselves.”
Baxter rejected this hostility, emphasizing that mergers are “an important and extremely valuable capital market phenomenon, that they are to be in general facilitated, and that it is socially desirable that uncertainty and risk be removed wherever possible to do so, subject, of course, to the very important limitation that where a merger threatens significantly to lessen competition, it should be halted.”
I agree with Bill Baxter.
Delay is a form of uncertainty and risk, and we should seek to remove it from the merger-review process whenever possible.
Changes made to the HSR Act nearly twenty years ago reflect the same goal. In my time on the staff of the Senate Judiciary Committee, I worked on legislation that raised the HSR Act filing thresholds to reflect inflation since 1976, and, perhaps more importantly, pegged them to inflation going forward. This dramatically reduced the number of reportable transactions, allowing the agencies to focus their resources on identifying transactions that are more likely to raise competitive concerns.
Let me pause here to make an important broader point.
Too often, effective leadership in antitrust enforcement is equated with the number of enforcement actions brought. The more enforcement actions brought, the thinking goes, the better the enforcement regime. In 1940, Justice Robert H. Jackson, then the Attorney General, rejected this thinking. As he said, “Any prosecutor who risks his day-to-day professional name for fair dealing to build up statistics of success has a perverted sense of practical values, as well as defects of character.” I stand with Attorney General Jackson.
Our job is not to bring the most enforcement actions. Our job is to get the right result. More than that, our job is to get the right result in an efficient manner, which often means reaching consent agreements with parties to remedy fully the sources of anticompetitive harm. Only when that is not possible, should we litigate.
That brings me to the second reason that unduly long merger reviews are a problem: they can waste public and private resources.
Certainly everyone in the business community who has received a second request, and everyone in the private bar who has worked on one, understands how burdensome compliance can be. We always want to minimize that burden consistent with upholding the antitrust laws.
Some perspective is important. Last year, the Division opened an investigation into only 2.3% of proposed transactions, and issued second requests in less than 1% of proposed transactions. That 1%, however, is expensive. It is also resource intensive.
The government also spends enormous amounts of time and money reviewing mergers that go to a second request. Now, let me be clear, this is time and money well spent when it is necessary to protect competition. Some transactions present knotty problems and it takes a while to untangle them. Once we are in a position to understand the competitive significance of the transaction, however, every additional minute and dollar spent reviewing the merger is deadweight loss. We have limited resources. As a related side note, over the past ten years the Antitrust Division’s budget has stayed roughly constant in nominal terms, which means it has declined in real terms, as salaries and other expenses have risen. A significant part of my job, then, is ensuring that we use those limited resources wisely.
So, how can we improve the process?
In a moment, I will announce a series of reforms we are taking to modernize merger review process, but let me lay down an important benchmark to help orient our thinking on this: Provided that the parties expeditiously cooperate and comply throughout the entire process, we will aim to resolve most investigations within six months of filing.
We are not unilaterally disarming, to be sure.
We will never compromise our ability to enforce the law effectively.
With the merging parties’ cooperation, we can expedite our review without compromising quality, and provide expeditious certainty to the business community, and the markets.
Of course, every case is different. Not every investigation can be resolved in six months. Indeed, in some cases, the parties may prefer a longer investigation rather than a quick decision by the Division. Or the parties may wish to extend the timing for various reasons. If the goal of the business community is a shorter review, however, we share that goal and are committed to working towards it.
With that said, let me discuss what we are doing going forward, and will continue to do, to modernize the merger review process. Every merger investigation has distinct stages, and I will discuss improvements that we are making at each stage in the merger review process.
1) Meet with the Parties Early. Often the first stage in an investigation is for the parties to meet with staff. One improvement that we are making is that the Antitrust Division Front Office will be open to an initial, introductory meeting. We expect that these meetings will be most productive if the parties include key executives from relevant businesses. We want to understand their deal rationale and any other facts they believe will be important to our analysis.
2) Model Voluntary Request Letter. Another improvement is that we soon will publish a model voluntary request letter on the Antitrust Division’s website. As many of you know, the Division often asks for the voluntary production of key information during the initial waiting period. This information is crucial to resolving mergers during the initial waiting period for the simple reason that there are enormous information asymmetries between the parties and the enforcer, especially early in a merger investigation. Parties should be prepared to provide key information within the first few days of their HSR filing, if not before filing, to allow the staff time to confirm critical facts through the parties’ document productions and our own independent investigation. One of my deputies, Barry Nigro, gave a speech on this topic last February, and I would encourage you to read it.
Publishing the model voluntary request letter is intended to give the parties a head start in identifying the kind of information they should look for, so that they can be proactive and submit it as early as possible in the process. A tailored version of this letter is sent to the parties in almost every investigation we open. It identifies information we believe will allow us to assess quickly whether there is any potential anticompetitive harm that would require a longer, more in-depth investigation. The sooner we get this information, the sooner we can close investigations that do not raise competitive issues. When a more in-depth investigation is required, we can use that information to narrow the scope of the investigation as much as possible.
3) Pull-and-Refile Accountability. We also have implemented a system to track what happens when parties pull-and-refile their HSRs. The Division cannot, of course, resolve all concerns within the initial 30-day waiting period. The parties, rather than face a second request they believe unnecessary, may choose to pull-and-refile their HSRs. Our new system is designed to ensure that we have an investigative plan in place to maximize our use of the additional time. We will determine whether we can close the investigation without issuing a second request, or, if one is necessary, narrow it as much as possible.
4) Model Timing Agreement. We are also publishing a model timing agreement on our website. As many of you know, timing agreements are a mechanism to encourage an orderly process by which the parties comply with the second request and the Division analyzes the transaction and decides whether to clear it, seek remedies, or seek to block it. Done right, timing agreements are mutually beneficial. The Division gets certainty on timing — which is in the parties’ control — and the parties get certainty, among other things, on the number of custodians, the number of depositions, and the availability of meetings with the Front Office.
Too often, though, negotiations over timing agreements are taking on a life of their own. Both the parties and the Division are spending too much time on process, slowing our ability to get to the heart of the matter. To facilitate reaching an agreement, we will be publishing a model timing agreement on our website.
5) Reforming Timing Agreements. We are also cognizant that timing agreements are a deviation from the process that Congress outlined in the HSR Act, which sets a deadline of 30 days for the Division to decide once the parties certify compliance with the second request. Thus, we also intend in the coming months to make changes to the model timing agreement in order to narrow potential areas of disagreement, facilitate more efficient reviews, and bring the process closer in line with the HSR Act. While we are still fine tuning the specifics, here are a few examples.
- A) Fewer Custodians: We will seek documents from fewer custodians than we generally have in the past. While every investigation is different, as a general matter we will assume that 20 custodians per party will be sufficient unless the Deputy AAG in charge of the investigation explicitly authorizes more.
- B) Fewer Depositions: We will also take fewer depositions than in the past. Again, while every investigation is different, we generally will not seek more than 12 depositions unless the deputy in charge of the investigation authorizes a greater number.
- Together with reducing the number of custodians, these changes will tangibly reduce the burden on the parties of complying with a second request.
- C) Shorter Time from Compliance to a Decision: We also will strive to make a decision as quickly as possible from the time the parties’ certify compliance. We will make a decision in no longer than 60 days — sooner, if possible —with the proviso that a Deputy AAG can authorize an extension if necessary.
In exchange for these benefits, we also are going to modify what we expect from the parties:
- D) Faster and Earlier Productions of Documents: We will expect to receive documents and other information earlier in the compliance period. If the parties employ traditional document reviewers, this will mean a more robust rolling production, with the parties producing several tranches of documents roughly evenly spaced over the compliance period. For parties employing technology assisted review, it will mean completing the bulk of the production a certain number of days in advance of certifying full compliance.
- E) Earlier Production of Data: Data is increasingly important in many investigations. Data is often the key to getting merger decisions right. Frequently, there is no reason that data cannot be produced substantially earlier than production of the main bulk of documents. We will expect to receive early cooperation on identifying relevant data for our economists to analyze. We will further expect production of useable data substantially before the second request compliance date.
- F) No More Privilege Log Gamesmanship: We also want to eliminate gamesmanship on privilege issues. The Division respects the attorney-client privilege and the work product doctrine, but too often we see parties game the process, withholding large numbers of documents as privileged, only to de-privilege and dump many of these documents on us much later in the process, often on the eve of a particular deposition. While some of the de-privileged documents might be close calls, most never should have been withheld in the first place.
- G) Longer Post-Complaint Discovery Period: Finally, we will require adequate time to conduct post-complaint discovery in the event that it results in contested litigation. Because most merger investigations do not result in contested litigation, and parties therefore can reduce the burdens of the second request review by agreeing to defer some discovery until after a complaint is filed.
6) CID Enforcement. In addition to those improvements to the timing agreement, which you will see on our website when the model timing agreement is posted, we are going to take steps to ensure that third parties comply with our civil investigative demands in a timely manner. A second request is the primary tool by which we get information from the merging parties. While the merging parties are generally the most important source of information about the competitive significance of a transaction, third parties often possess critical documents and data.
As you know, the Antitrust Civil Process Act empowers the Assistant Attorney General of the Antitrust Division to issue CIDs for relevant materials. In significant merger investigations, this can be an essential tool to collect third-party documents, data, and testimony. As you can imagine — in fact, as many of you know firsthand — third parties rarely greet CIDs with enthusiasm. Compliance is often slow and incomplete. This hampers the Division’s ability to analyze the transaction at issue expeditiously. Going forward, we are going to hold CID-recipients to the deadlines and specifications in the CIDs we issue. When necessary, we will not hesitate to bring CID enforcement actions in federal court to ensure timely and complete compliance.
7) Demonstrating Leadership in Parallel Investigations. We are also seeking ways to improve coordination with foreign enforcers in parallel investigations. To facilitate the process, when facing parallel investigations, parties should consider aligning time periods with other jurisdictions if possible, and where appropriate ask us to help with that. If there is cooperation on the parties’ side, we will gladly help reduce the burdensome red tape wherever possible.
8) Withdrawing the 2011 Remedies Guide. We are also taking a close look at our remedies policy. Negotiating remedies to anticompetitive mergers often adds significant time to the merger review, and our commitment to shortening the duration of merger reviews extends to the remedies phase. While our review is underway, I want to be transparent with the bar about what the Division’s practices will be. To that end, today, I announce the withdrawal of the 2011 Policy Guide to Merger Remedies. The 2004 Policy Guide to Merger Remedies will be in effect until we release an updated policy.
9) Release of Merger Review Statistics. I have described a long list of changes we will be making.
How will you know if they have had an impact?
We will increase transparency.
Going forward, we are going to release statistics that show, on average, how long the Division is taking to review mergers. Specifically, we will release the average duration of second request investigations and the average length of time from the opening of a preliminary investigation to the early termination or closing of the investigation, including investigations that did not result in the issuance of second requests. The Division previously released these statistics in connection with the 2006 amendments to the Merger Review Process Initiative. We believe that releasing them periodically going forward will increase our own accountability and give the private bar and the business community greater insight into our process.
To conclude, we are committed to accelerating the pace of merger review consistent with enforcing the law because we believe that doing so is good for American consumers and taxpayers. We have taken action already, and we plan to do more in the coming months. It does takes two, however. To that end, I welcome the private bar and any other interested parties to reach out to discuss any other ideas we should consider as we continue to make our processes more efficient.