New FTC Second Request Procedures

The Hart-Scott-Rodino Antitrust Improvements Act of 1976 probably ranks as the most important piece of antitrust legislation in the last half-century.  It certainly has had the most practical impact, and the most significant part of that impact has come from its premerger notification provisions.  No antitrust issue has generated more debate, or created more tension between the private antitrust bar and the agencies, than the HSR merger notification procedures—particularly the right of the antitrust agencies to effectively enjoin a transaction without ever going to court, simply by issuing a “second request,” the odd moniker for the information request issued by the agencies when they seek to investigate a particular transaction.

This is, on its face, a spectacular power to hand over to a federal agency that does not have the regulatory mandate to approve a transaction.   It is one thing to have a federal regulatory scheme, such as in communications, where the FCC’s approval is required before certain actions (such as some mergers that require license transfers) can take place.  But neither the FTC nor the Antitrust Division of the DOJ has approval power over mergers, and both would be quick to deny that they are regulatory agencies.  Nevertheless, the HSR Act gives both agencies the ability to stop a transaction indefinitely by simply issuing a second request.  Given this power, it is not hard to imagine that at least some in the agencies might seek to use it not only to gather information, but also to simply slow down transactions so that the agencies can have plenty of time to investigate them at their own pace.  We can argue all day about whether this has happened or not, but certainly a lot of people believe it has, and there is absolutely no doubt that the potential for this kind of abuse exists in a system that gives essentially unlimited discretion to the agencies. 

On the other hand, merger analysis done right is hard stuff.  In the old days, pre-Chicago School, merger analysis was commonly based on simple market share and concentration analysis, without a lot of detailed factual analysis of the likely effects of the transaction.  Presumptions were often determinative.  But at least since the 1982 Merger Guidelines, and increasingly since that time, merger analysis is highly fact-specific and, in recent years, particularly focused on economic and statistical analyses.  Of course, econometrics requires data—preferably lots of data, since more data points usually result in more robust analyses.  Companies have also become more data-intensive, using econometrics and other forms of statistical analysis to inform business decisions.  The result is a lot more data in company files.  Over this same period, the use of e-mail and other electronic documents has greatly expanded, and since the storage of electronic bits is much cheaper and easier than the storage of paper, potentially responsive files have mushroomed in volume.  Put all these factors together, and it is inevitable that, to do their job right, the agencies will need a significant amount of information, and as data becomes more voluminous, so too will the desire to see more data grow.

The net of all this is that a second request is frequently enormously burdensome, and the costs associated with this process can be staggering (at least to the parties involved).  It is not uncommon for the response to a second request to require several months (four to six is very common and longer periods are not unheard of) and several million dollars (even double-digit millions in some cases).  Unless you are a staff lawyer at one of the agencies who has never worked in private practice and, in addition, believes that whatever costs you think necessary to do your job are justified, everyone familiar with this process agrees that it is a mess, and has been for decades.  As a result, every time there is a new head of one of the antitrust agencies, one of the items on their “To Do” list is always “Reform the second request process.”   Unfortunately, this is a lot easier to state as a goal than it is to execute, for several reasons.  First, the career staff at the agencies, quite understandably, want to protect their ability to investigate as thoroughly as they think necessary.  Unfortunately, there is quite a wide range of views among the career staff as to what is really necessary, so finding consensus on reforms from that perspective is difficult.  Second, the views of the private bar also vary dramatically; some think HSR (or maybe the antitrust laws in general) should be repealed altogether, some would rather complain about the process than constructively engage in finding a solution to the problems, and some may think that a long and cumbersome process is just fine for their bottom line.  Getting to common ground here is obviously a challenge.

Deborah Majoras, the current Chairman of the FTC, understands all these points about as well as anyone could.   She has spent about 10 years in private practice, a lot of that time doing merger work from the defense side, and about five years in senior positions at the Antitrust Division and the FTC, doing merger work (among other things) from the agency side.  When she was in private practice, she coauthored (along with the author of this Commentary) one of the most detailed critiques of the HSR second request process, which estimated the potential costs of what the authors viewed as overreaching by the agencies in the hundreds of millions or potentially billions of dollars.  And while at the Antitrust Division, she was one of the persons responsible for a number of HSR reforms adopted by that agency under the leadership of Charles James (in the interests of full disclosure, another former partner at Jones Day).  So it is hard to argue with the proposition that she is well placed to try to find the reasonable middle ground, and thus anything that comes out with her name on it deserves to be treated with respect as a serious effort from a serious (and extremely well-informed) authority. 

Chairman Majoras has just released a paper outlining a series of merger process reforms she has instituted at the FTC that are intended to create a better balance between the legitimate need of the agencies to be able to investigate properly and the burdens that are placed on parties—who are, after all, only proposing a business combination, not a plot to destroy the Union.  This is a serious effort to find the solid middle ground, and given the Chairman’s background, one would have expected nothing less.  But it is also obvious that the Chairman was trying to find a position that would be at least tolerable to the FTC staff, who may be somewhat less inclined toward balance than the Chairman undoubtedly is.  And so the result is complex, and the effects of the changes unclear, at least without some experience.  It is at least possible that these changes will significantly reduce burdens without impairing, in any meaningful way, proper investigations.  But it is also possible that this will not be the case.  As we go through the various elements of these reforms below, we can see precisely which critical variables will determine whether these reforms will be any more significant than prior efforts.  Our instinct is that these reforms will in fact be a positive step toward a more rational process, but only time will tell for sure.

Let’s deal with the reforms one at a time and evaluate both the potential benefits and how these benefits could be lost.

Custodian Searches.   One of the most significant complaints about the second request process is that staff members frequently want to look under every rock, without balancing the likelihood of finding relevant information against the burdens imposed.  In a litigation context, a district judge or magistrate would referee any disputes, but in the second request process, this option is not practically available.  And so one way this overreaching manifests itself is by insisting on the search of a very large number of individuals, including many at relatively low levels in the company.  Any experienced merger lawyer knows that, in the vast majority of cases, a very small number of individuals at each company is likely to have the great bulk of truly useful documents dealing with the transaction being scrutinized.  But it is also undeniable that you never can tell what you will find if you look in a lot of files; maybe, just maybe, there will be very useful documents several layers down in the organization that will have the most useful information, and agency staff can give you plenty of examples where, in their view, that was in fact the case.  Since the agency staff do not have the same incentives that the parties do—speed and as low a cost as possible—they will quite understandably want to go deeper in the organization than a reasonable person might think appropriate if one was just balancing potential benefits and burdens.  So searches of more than 100 persons (called “custodians” in HSR-speak) are common, and those of 200 or more are unfortunately not really rare.  This is, of course, in addition to the ever more voluminous databases and central files that are generally also demanded. 

From experience, we know that this is probably the single most aggravating aspect of second requests to the parties—the need to spend time and money to search the files of employees considered very unlikely to have anything useful to contribute to the analysis.  On top of those costs, moreover, is the disruption to business operations that comes from conducting those searches.  Chairman Majoras has the experience to understand this point, and thus her first—and perhaps most important—reform is a presumption (rebuttable, and we will come to that in a moment) that no more than 35 custodians would have to be searched in a second request investigation.  One could quibble about the number—it could have easily been 25 or even less, since in most transactions there are probably five to 10 custodians whose files (in addition to central files and databases) would provide the useful information necessary for the agency to make an informed decision.  But this is really a quibble; 35 is certainly within the zone of reason and most importantly will, if faithfully applied, eliminate the abuse of 100 or 200 custodian searches, which are inarguably unnecessary except in very rare transactions.  So this is a really important reform—if the presumption holds.

The presumption is just that, however, and a larger search can be authorized by the Director of the Bureau of Competition—the FTC’s senior competition staff position—if “it is reasonably likely that a larger search group is necessary for the FTC to analyze a transaction’s potential competitive effects.”  Obviously, this is an exception that could easily swallow the rule, depending on how the Bureau Director exercises this essentially unlimited discretion.  We won’t know the answer to this critical question for some time, and it will be watched very closely, since the same kind of presumptive approach runs through the entirety of these reforms.  If the Bureau Director regularly expands the search group beyond 35, it will be fair to conclude that these reforms were not effective.  Only time will tell.

This 35-custodian limit also has some potentially significant strings attached, and how these are implemented will also tell us a lot about the effectiveness of these reforms.   The parties have to supply organization charts and make someone available to the staff who knows what various people in the organization actually do, so that the staff can intelligently identify the 35 (or fewer, since this is only a cap) custodians from which it wants to see documents.  This is unobjectionable in concept, with the only issue being how fast the staff can decide on the search list; the reforms contain no time limit, and obviously one of the purposes of the reforms—to speed up reviews—would be defeated if the staff were to delay on this and other points where they have no deadline to act. 

In addition, the 35-custodian limit is also conditioned on the parties’ agreeing, if the FTC challenges the transaction, to at least a 60-day discovery period before any determinative hearing.  This also seems unobjectionable in concept; today, one of the reasons the staff sometimes overreach is that they want to be prepared in case there is litigation and the parties push the court, as they frequently do, for a quick hearing.  This process—limiting the search group in return for additional discovery if there is litigation—should be perfectly acceptable in most situations, since the likelihood of litigation is relatively small.  Litigation can be avoided by either convincing the agency that there is no problem or negotiating an agreed resolution of any problems that cannot be resolved.  Only a very small number of transactions are ever the subject of litigation, so this is a tradeoff that should almost always be acceptable to the parties.

Finally, there is a string that is perhaps a bit more problematic.  In order to receive this 35-person limitation, the parties must agree either to a “rolling production” (producing the documents in increments as they are processed rather than waiting until the end and providing all the documents at once) or to a production that occurs 30 days prior to certification of compliance with the second request.  In the normal course today, the parties certify compliance when they have completed their production and by statute are free to consummate their transaction after waiting an additional 30 days.  In practice, it is very common—almost universal—for that statutory timetable to be replaced by an agreement between the parties and the agency on a specific schedule that is actually longer than 30 days.  This commonly includes an agreement on when the staff will make their recommendation, what opportunities the parties will have to talk to staff or others in the decision chain during the process, and the like.  These reforms seek to build in an automatic additional 30 days, without the need to negotiate for that extra time.

This is hard to justify conceptually.  Why should the FTC be able to extract an extra 30 days on its statutory stay simply for being reasonable in the number of custodians that it agrees must be searched?  On the other hand, as a practical matter, these extra 30 days—and sometimes longer—are frequently negotiated away today, so in many (probably most) cases, this will not mean any further delay.  But this conclusion comes with a big caveat—it is true only if it is not followed with the same kinds of demands for additional time that parties typically see today (usually accompanied with explicit or implicit threats that staff will recommend a challenge in the absence of the additional time).  Again, time will tell how this plays out in the real world, but unless it has the practical effect of adding 30 days to what is already a lengthy process, this is a tradeoff that many parties to transactions will be willing to make.

Of course, the 35-custodian limit does not apply to central files or databases, and no limitations are placed on demands for that kind of information.  Since this is increasingly a focus of what are ever more economist-driven second requests, it is certainly possible that the burdens eliminated by limiting the custodian searches could be overwhelmed by additional burdens imposed in seeking data.  Again, time will tell whether this comes to pass.

Historical Information.   Another common complaint about second requests is that they seek too much historical information and they don’t accept reasonable cutoffs that enable efficient searching.  Until relatively recently, it was not uncommon for second requests to seek information from the last five or even 10 years.  In recent years, the standard has been to go back three years, although there have been many exceptions to that limit.  Another useful presumption in these reforms (assuming again, of course, that it is not regularly overridden) is that the relevant time period will be the last two years, which seems perfectly reasonable and appropriate.  That takes care of the back end, but the front end of the relevant time period has also been a problem.  The agencies have commonly required all relevant documents to be produced that were created at any time up until 30 days before certifying compliance—and for some categories, up to 14 days.  Since it takes time to process and review documents after they have been pulled from the files, this requirement has produced so-called “second sweeps,” where the relevant files have to be searched a second time right before the final production, with all the attendant disruption and cost.  It has sometimes been possible to negotiate this away with the staff, but only in return for some other concession.  Going forward, the presumption will be that only documents created up to 45 days before compliance (or before the end of production if taking advantage of the 35-custodian limit) must be produced.  This would eliminate or significantly reduce second sweeps, another very positive change. 

Both of these time-limit changes can be overridden, but in this case by staff, not just by the Bureau Director.  This obviously leaves a lot of discretion in the staff’s hands; exactly how that is handled will determine the utility of these changes.

Finally, there is what could be a very important exception to these cutoff dates that, improperly applied, could swallow this good rule—it does not apply to data.   With respect to data, the reforms simply require the staff to communicate with the parties about the data they need and why—i.e., the theories of competitive harm to which the data relates—and then provide a meeting with a senior official from both the Bureau of Competition and the Bureau of Economics if the parties think the staff are overreaching.  Since the data requests are already the most burdensome part of many second requests today, and that will be even more true in the future if the 35-custodian limit is actually applied, this leaves a mighty big hole in the reforms.  Still, there is merit to the argument that the data needs for transactions will vary with the matter and the parties, and that arbitrary limitations are hard to set, given that variety.  So this too is an area where we will have to wait for experience to determine the impact, if any, of these reforms. 

Backup Tapes.   Another point of controversy in the past has been backup tapes.  Almost every organization has some form of backup for its electronic materials, but in the majority of cases, it is intended as a last resort in case of catastrophe.  As a result, backup materials are generally not indexed or easily searchable, but nevertheless, at least literally, they contain potentially responsive documents—somewhere in there.  In many situations, the staff will agree to exempt backup tapes from the second request, but not always, and not always without some other concession.   These reforms seek to eliminate the possibility of strategic leverage here by creating a presumption that the FTC will require production from backup materials only when really necessary, and in any event the parties can elect to preserve only the materials from two calendar days that are designated by the staff.  Again, in principal this is a positive change, but since these are only presumptions, we will need to see how this plays out in practice. 

Privilege Log.   The privilege log has long been a source of controversy.  As a practical matter, FTC staff rarely seek to impeach a claim of privilege, but the detailed requirements of producing a privilege log are among the most time-consuming and burdensome parts of a second request production.  First of all, privileged materials have to be segregated.  Then, under past practice, they had to be logged with a large amount of descriptive material—to, from, date, subject, etc., and, most difficult, the full name, title, and other information of every addressee (many of whom will have left the company, changed positions, or done something else that makes collecting this particular data extremely time-consuming).  While it is difficult to argue that this information would not be relevant if there ever was a fight over privilege, the actual fights are rare, but the burdens apply to all parties in every transaction.  This does not make sense in any rational analysis.  Chairman Majoras has tried to split the baby here, but it is not clear that the changes implemented will actually have the desired effect.  The reforms provide that a party may elect to submit a partial privilege log, with only very limited information, and the staff can then identify either five individuals or 10 percent of the total number of custodians searched (whichever is greater) for which the party will have to produce a complete log in order to certify compliance with the second request.  In the case of any litigation, the party will have to agree to submit a complete privilege log for all custodians no later than 15 days after it is requested.

This sounds reasonable at first blush, but on closer scrutiny, it may not really help much.   It is likely to be fairly common that picking the right five people will encompass the vast bulk of all privileged documents.  The same purpose—to run a check on legitimacy—could have been accomplished by allowing the staff to pick one or two custodians.  In addition, the fact that parties cannot certify compliance until after the completion of the full privilege log for the five persons selected by the staff may well mean that it will be more effective to just do the full log for all 35 custodians rather than run the risk of losing additional time at the end of the process.  And this will be especially true where parties have not been willing to make the commitments necessary to get the 35-custodian limitation.  So here again, experience will disclose the importance of this reform. 

Staff Lawyers.   There are a few other smaller reforms—limiting the definition of “agents” to those involved with the transaction (seems obvious but was not the case in the past); acknowledging that the requirement to produce information about every facility that has been involved in the manufacture or sale of any relevant product was frequently overreaching, and thus encouraging staff to limit this demand as appropriate; and removing from the definition of “documents” tax and other materials that are not relevant to a competitive analysis.  But there is one additional little reform hidden in the text that might be really important—that a staff lawyer with “substantial merger experience” must participate in all negotiations over the second request.  Too often in the past, these negotiations have been left to junior members of the staff, who understandably are scared to death of being criticized for giving away too much, and not at all concerned about ever being criticized (internally at least) for being too tough.  This results in negotiating with someone with a hand on his or her wallet, which all experienced lawyers know is a lot harder than dealing with someone with real experience in the subject matter.  This rule alone might have as much practical significance as any in this package.

Use of Search Software.   Finally, there is one additional piece of these reforms that sort of comes in from left field and is definitely not a “reform” in the common meaning of the word.  Parties today are commonly using electronic search software to sort out responsive and nonresponsive documents, to eliminate duplicate documents, and to help identify privileged and significant documents.  Given the tremendous increase in the volume of what are increasingly electronic rather than paper documents, the use of this kind of advanced technology is essential if transactions are to be completed in a timely manner.  Properly used, electronic review software can increase the speed of review by five to 10 times and, as a result, reduce the need for live bodies to look at materials, thus cutting costs dramatically—and all with generally better-quality results.  In other words, the software typically makes fewer mistakes than the human reviewers do!  The FTC (and the Antitrust Division, for that matter) is behind the curve with this technology, and as a result, reviews and negotiations dealing with this subject have been more complicated than necessary.  They are trying hard to catch up, and will eventually do so, but in the meantime, they seem to view these tools more as problems than solutions.

And so we see, tucked in at the end of these reforms, not a reform but an additional requirement that seems unnecessary and, at least potentially, both inappropriate and unenforceable.   The FTC says it will add an instruction to its version of the second request that requires a party intending to use such electronic review software to get permission to do so.  There are any number of objections to this.  Frankly, how the parties comply with their obligations to produce responsive documents is none of the agency’s business.  The notion that it is perfectly OK to decide how to do this using humans, and not OK to carry out the same legal obligations using software, seems atavistic at best.  The statement justifying this intrusion into the lawyer’s work product identifies the concern that the use of this software might eliminate responsive information by the use of de-duplication or “near de-duplication” techniques.  Well, the short answer to that is that the parties and the lawyers do so at their peril, just as is the case with all the other determinations that are made about whether documents are responsive or not, or privileged or not.   And it is hard to see how this discussion could take place without inquiring into the lawyer’s work product—the tools and criteria by which lawyers select and sort the materials that they review. 

This very jarring exception to the very positive overall tone and approach of these reforms is hard to understand.  It reeks of the agencies’ old approach that the other reforms seem to be trying to put to rest—“our way or the highway.”  It seems to be based on a profound fear that somehow, someway, the parties or their lawyers are going to pull a fast one over the agencies, which is the universal excuse for being unreasonable, and almost never a proper justification.  And even if the odd lawyer misbehaves, why should all parties to all transactions be penalized for that misbehavior?  This issue will—and should—get more attention in the near future.

Notwithstanding this last exception, these are very desirable changes in principle, and now we will have to see how they work in practice.   I suspect they will work fairly well, since both Chairman Majoras and Jeff Schmidt, the Bureau Director, understand clearly that this will have been a great waste of time if these presumptions turn out to be worthless.  It will also be interesting to see if the Antitrust Division, which is working on its own reforms, and traditionally has been at least slightly less regulatory in its approach to such matters, follows this lead or takes a different direction.  In the meantime, Chairman Majoras and the FTC deserve credit for taking a serious look at this area, where opinions run the gamut and tempers frequently run high, and for implementing changes that at least have the potential to make this process work better for all parties. 

Lawyer Contact

For further information, please contact your principal Firm representative or the lawyer listed below. General e-mail messages may be sent using our “Contact Us” form, which can be found at

Joe Sims

Jones Day Commentaries are a publication of Jones Day and should not be construed as legal advice on any specific facts or circumstances. The contents are intended for general information purposes only and may not be quoted or referred to in any other publication or proceeding without the prior written consent of the Firm, to be given or withheld at its discretion. The mailing of this publication is not intended to create, and receipt of it does not constitute, an attorney-client relationship.

We use cookies to deliver our online services. Details of the cookies and other tracking technologies we use and instructions on how to disable them are set out in our Cookies Policy. By using this website you consent to our use of cookies.