# FEDERAL TRADE COMMISSION > Briefs, arguments, decisions, and more. URL: https://www.frixlaw.com/law-library/documents/agency%3Aftc%3A67e6526af395f187 ## Record - **Collection:** Agency decision - **Document type:** Agency decision ## Text 1 1 FEDERAL TRADE COMMISSION 2 3 In the Matter of: ) 4 WORKSHOP ON SECOND REQUEST AND ) 5 MERGER INVESTIGATION BEST PRACTICES ) 6 ------------------------------------) 7 8 JULY 10, 2002 9 10 Room 332 11 Federal Trade Commission 12 600 Pennsylvania Avenue, N.W. 13 Washington, D.C. 20580 14 15 16 The above-entitled workshop came on for comments, pursuant to notice, at 2:03 p.m. 17 18 19 20 21 22 23 24 25 For The Record, Inc. Waldorf, Maryland (301)870-8025 2 1 2 APPEARANCES: 3 4 5 6 7 8 9 10 11 12 ON BEHALF OF THE FEDERAL TRADE COMMISSION: MARY T. COLEMAN, Deputy Director, Bureau of Economics MICHAEL G. COWIE, Assistant Director, Bureau of Competition GABRIEL H. DAGEN, Acting Assistant Director for Accounting and Financial Analysis ELIZABETH SCHNEIROV, Special Assistant to Director, Bureau of Economics 13 14 PANELISTS: 15 MIKE DENGER, ESQUIRE, Gibson, Dunn & Crutcher 16 MEG GEURIN-CALVERT, Economists, Inc. 17 RICHARD HIGGINS, LECG 18 TIM DANIEL, NERA 19 JOHN WOODBURY, Charles River 20 DAVID PAINTER, LECG 21 22 23 24 25 For The Record, Inc. Waldorf, Maryland (301)870-8025 3 1 FEDERAL TRADE COMMISSION 2 3 4 I N D E X PAGE: WORKSHOP: 5 6 Mike Denger 5 7 Meg Geurin-Calvert 11 8 Richard Higgins 20 9 John Woodbury 27 10 Tim Daniel 42 11 David Painter 49 12 13 14 15 16 17 18 19 20 21 22 23 24 25 For The Record, Inc. Waldorf, Maryland (301)870-8025 4 1 P R O C E E D I N G S 2 - 3 MS. COLEMAN: - - - - Good afternoon. We're here for the 4 final of the Second Request and Merger Investigation Best 5 Practices Workshops and this one is involving data issues: 6 accounting and financial economic data. 7 8 9 I'm Mary Coleman, and I'm the Deputy Director of Bureau of Economics. With me is Elizabeth Schneirov who is Special 10 Assistant to the Director, Bureau of Economics, Gabe Dagen, 11 who is Acting Assistant Director for Accounting and Financial 12 Analysis, Bureau of Economics, and Mike Cowie, who is 13 Assistant Director, Bureau of Competition. 14 The purpose of these workshops is to try to get 15 feedback from the outside bar and consulting firms and 16 practitioners on issues of the merger investigation process 17 and how we can improve that process. 18 And our role here is to listen and answer 19 questions there are, to give any feedback to points raised by 20 individuals. 21 We have spoken to some people who were interested 22 in speaking at this workshop, and I'll introduce them and let 23 them give their discussion. 24 25 This is being transcribed. And we don't have a microphone up front just to try to make this a little more For The Record, Inc. Waldorf, Maryland (301)870-8025 5 1 informal, so we just ask that as people speak, that you 2 identify yourself and who you work with and speak up so that 3 the court reporter can hear what you're saying. Okay, I'm just going to start by introducing 4 5 folks. I'll start out with Mike Denger. He is at Gibson, 6 Dunn and Crutcher and he's worked in the antitrust field for 7 many years, has done work on merger investigations over that 8 time frame and he has prepared a proposed short PowerPoint 9 presentation which, I think, is out in the -- as you walk in 10 in the hallway. 11 MR. DENGER: Do you want me to come up there or -- 12 MS. COLEMAN: No; you can just speak from your MR. DENGER: I always find it hard to speak with 13 14 15 16 17 seat. my back to the audience. MS. COLEMAN: You are welcome to -- whatever makes you comfortable. 18 MR. DENGER: I haven't tried that. 19 I always feel nervous when they ask a lawyer to 20 come and talk about financial data and syndicated data and 21 cost data. 22 if they have me here on the data issues. 23 I guess I probably don't do too well as a lawyer, And what I want to talk about a little bit is what 24 I think is the advantages in cooperation and data-sharing 25 with the antitrust enforcement agencies during the merger For The Record, Inc. Waldorf, Maryland (301)870-8025 6 1 review process. 2 And I think from the standpoint of a party to a 3 prospective merger, there are some advantages to working 4 closely with the agency in this area. 5 First of all, I think you can get a common frame 6 of reference if you understand the data that they want and 7 need and the format that they're looking at. 8 I think it helps both the respondent understand 9 the focus and concerns of financial analysts and economists 10 and the staff at the agency and it helps for a more 11 constructive dialogue where you do not have two trains 12 passing in the night, so to speak, on data issues. 13 It also allows the FTC, if they have the same data 14 that you're working from, to replicate the type of work that 15 your economists or financial analysts may do, and I think 16 that is always helpful in establishing the credibility of 17 counsel. 18 I think if you work with the agency, you will also 19 reduce the cost of compliance oftentimes limiting the 20 production of duplicative or unnecessary information. 21 For example, in a retail merger, if one party has 22 IRI data and one has Nielsen data, if you can work together 23 and sort of agree, get everybody that's concerned with we'll 24 either use Nielsen or IRI, you're establishing a common 25 framework from which you can have the dialogue. For The Record, Inc. Waldorf, Maryland (301)870-8025 7 I also think it reduces the merger review time. 1 2 It takes away a lot of time when it is focused on compliance 3 issues and things that are not really the substantive issues 4 and focuses of the staff. 5 And I think ultimately what a party's counsel and 6 probably his economist and financial analyst have to offer 7 imparts their credibility and if you cut out the Mickey Mouse 8 work and try to get to a common set of data that you can both 9 work from, I think that helps. 10 For example, let me focus first on producing 11 syndicated data. 12 data in the form used internally by the company, explaining 13 what you can provide and what you can't provide, provide them 14 some samples explaining the limitations. It oftentimes results in the narrowing by the FTC 15 16 I would advocate providing early samples of staff of the information that they seek. You know, I can probably say I've never met an 17 18 economist or a financial analyst who didn't have data that he 19 wouldn't like or she wouldn't like if they had their choice, 20 but I think that the staff will be realistic if you can 21 explain what they're going to waste their time on, as well as 22 you wasting your time on in trying to find a meaningful data 23 set. 24 We had a case in a recent merger where we were 25 allowed to directly communicate with the FTC analyst and For The Record, Inc. Waldorf, Maryland (301)870-8025 8 1 economist to try to resolve the data problems sometimes 2 without FTC lawyers on the phone, and that came as a result 3 of, I think, the trust and credibility that both sides 4 developed in each other. I think if you work together, you can prioritize 5 6 data requests. 7 basis. 8 I would advise providing it on a rolling Sometimes if you give it on a rolling basis, give 9 the early years first, or the most current year's first, 10 oftentimes I think the FTC staff will say that's all they 11 need. 12 lot of information that you otherwise would have to do. They don't really need you to go back and to upload a 13 Similarly, in terms of financial data, we would -- 14 I think have found it very useful to provide exemplars right 15 early on, oftentimes in the first meeting that you have with 16 the staff of your sales data, how your cost data is 17 formatted, how your data is organized, if you have customer P 18 & L's that sort of thing, what their limitations are. 19 And again, when you look at what data is 20 reasonably available and obtainable, I think oftentimes you 21 can narrow it to what is meaningful to everybody concerned. 22 And also, it's not a one-way street. I think by 23 respondents learning what data the FTC is really focusing on 24 and wants, you understand how they are approaching their 25 analysis. For The Record, Inc. Waldorf, Maryland (301)870-8025 9 So again, provide an early sample, sequence your 1 2 production, arrange meetings between people who are 3 knowledgeable in the company with the FTC staff that's 4 actually seeking the data, the economists and the financial 5 analysts, and by having people who actually know what they're 6 talking about explain it, rather than always putting a lawyer 7 in as an intermediary, I think we found that makes it a lot 8 easier for everybody else concerned and oftentimes it leads 9 to a narrowing of the search requests by the FTC staff. 10 When you've agreed on what you're going to produce 11 in some form, we often provide sample reports to make sure 12 that everybody is agreed and understood on what each other 13 wants before you go to all of the expense of running it for 14 all of the years or all of the markets or all of the products 15 in question. And again, oftentimes that early cooperation can 16 17 eliminate costs and eliminate burdens on the company, as 18 well. 19 I guess our philosophy is sort of this: The facts 20 are the facts. 21 are all speaking about the same kinds of data and the same 22 facts. 23 And it's easy to have a dialogue when people It's important to have trust and credibility and I 24 think that that can best be obtained by -- frankly on the 25 substance, but not on all of these procedural issues. For The Record, Inc. Waldorf, Maryland (301)870-8025 10 1 And don't always view the FTC staff as the enemy 2 3 who is out to get you, which lawyers sometimes have a 4 tendency to do. The FTC employees just have a job and we expect 5 6 them to do it thoroughly. You also, I think if you cooperate 7 with them, can avoid a lot of waste, unnecessary duplication, 8 looking at data that is really not going to advance the 9 analysis on either side. 10 So I guess that's what I would say. And I guess 11 our philosophy in all of this has sort of been like the 12 philosophy we have in litigation. 13 And I will say I've tried a fair number of 14 preliminary injunction merger cases over the years, but when 15 you try to be hard-nosed and fight over every point, all you 16 do is waste your client's money and annoy the agency. 17 As every firefighter knows, blaring the siren can 18 19 be fun, but the point of the job is to douse the flames. 20 this case, the point of the job is to try to get the merger 21 through in an economically viable form. 22 fight over the substance all we want, but I think it really 23 pays to try to work with the agency to get the data and the 24 basic facts set out so then you can have a substantive 25 dialogue. Thank you. For The Record, Inc. Waldorf, Maryland (301)870-8025 In And we can 11 MS. COLEMAN: 1 2 Geurin-Calvert. Thank you, Mike. Next up is Meg She's currently at Economists, Incorporated. 3 She also worked at the antitrust division for a 4 5 number of years and she is going to talk about issues related 6 to transactional-level data. 7 please. 8 MS. GEURIN-CALVERT: Meg, if you could come up, Thanks, Mary. I appreciate 9 the opportunity to be here. 10 and just cover some brief topics so there's lots of 11 opportunity for discussion. 12 And let me follow Mike's lead The perspective I bring is, as Mary mentioned, I 13 was at the antitrust division as an assistant chief and had 14 been a total of ten years at the antitrust division and at 15 the Federal Reserve Board with some time in between in the 16 private sector as a consultant. 17 And I think what I have learned best about the 18 particular topic we're addressing today, which is data and 19 both financial and accounting and transactional in assessing 20 mergers, was when I went back to the division as an assistant 21 chief working on my first second request where I understood 22 better, having worked on a very large number of second 23 requests on the receiving side, just what all the costs were 24 that were associated with compliance and what all the effort 25 was that it took to get the data together, but at the same For The Record, Inc. Waldorf, Maryland (301)870-8025 12 1 time understanding the kinds of data that were available in 2 firms somewhat better than I had when I had been in the 3 division the first time. 4 So I had more of a sense of what was out there, 5 but at the same time came to a little bit more of a 6 perspective of that don't ask for what you can't use. 7 And so I guess I have six words to present to think on 8 the topic of transactional data: testable hypotheses, 9 critical loss and don't assume. 10 And what I would like to put the perspective in is 11 I think that there's a common goal between the private 12 parties and the government agencies that are reviewing a 13 given merger. 14 The common goal is essentially to have a review 15 process that hopefully ends in a significant review of the 16 appropriate issues and from the party's perspective, a 17 resolution of those issues and the mergers proceeding or from 18 the government's perspective, to the extent there is a 19 significant anti-competitive concern, having a strong 20 litigation base built up such that you can go to court and 21 demonstrate persuasively to a court that the merger actually 22 does present a problem. 23 To get there, I think it's the same process. 24 need full integration from the beginning and up to the 25 decision-making point of the data with the analysis, you need For The Record, Inc. Waldorf, Maryland (301)870-8025 You 13 1 consistency of the data analysis and the data work with the 2 documentary review, the depositions, and so on, and you 3 obviously need time limits. And I think in any industry, essentially what that 4 5 means is you need clear theories to be spelled out on both 6 sides as early on very specifically as to what are the 7 concerns, the product, geographic entry and competitive 8 effects. 9 I think everyone knows the merger guidelines and 10 what needs to be laid out in-house very clearly in the 11 government and the agency, particularly by the economists and 12 lawyers working together as what is our theory of the case, 13 what are the two or three things on which this particular 14 case turns. 15 16 That leads them to the next thing, again in-house, which is what are the testable hypotheses. 17 And that is why I agree with Mike completely that 18 that's where the dialogue starts between the FTC staff and 19 the outside parties as to, "Here are our hypotheses. 20 are how we would like to test them. 21 we're looking for," and then I think you can focus lastly on 22 what are the results of the analyses. Here Here's the information 23 You may differ on theories and have a dialogue; 24 you may differ on the tests and have a dialogue; you may 25 differ on the reliability or the meaning of the results and For The Record, Inc. Waldorf, Maryland (301)870-8025 14 1 you can have a dialogue. And let me briefly talk about three types of areas 2 3 in which it plays out in different ways. There are three 4 kinds of transactional data that I think occur -- a context 5 in which they occur. 6 The first is what I would call industry-wide, 7 where both parties, private parties and the FTC or the 8 government agency have access to industry-wide databases. 9 This occurs largely in healthcare, in airlines, in banking, 10 in electricity. 11 regulated has some pretty good data. 12 13 Essentially anything that was or is Everybody has transactional-level data down to the zip code and the DRG, or whatever, with which to work. And as a result, you can focus. 14 The outside 15 economists can start much earlier, can do a full-scale 16 analysis. 17 can proceed much more quickly to very specific kinds of 18 analyses and discussions. 19 They'll know what the issues are, and the dialogue And so I think that is a case where everyone, more 20 or less, has a level playing field, but where it's very, very 21 critical that when the agency staff have particular concerns 22 about a narrow product market or are particularly concerned 23 about the standard by which one defines the geographic 24 market, inclusion or exclusion of certain kinds of zip codes, 25 that they articulate it so it can be discussed and tested. For The Record, Inc. Waldorf, Maryland (301)870-8025 15 1 The second area Mike dealt with in terms of 2 standard data, so I won't go into more other than to say that 3 there I think the issue is much more so, not the availability 4 of the data, but the methodologies that are applied or what 5 the testable hypotheses are and the debate over the nature of 6 the results. 7 productive dialogues helps that. 8 But I agree with Mike completely that having I think the hard world is the third one which is, 9 unfortunately, most of the transactions. 10 kind deals; it's almost every manufacturing industry merger. 11 It is the one of a When you think about it, the private parties have 12 extraordinary access to their own data. 13 detailed set of every transaction, most likely, that they 14 ever did, although it's surprising to some extent in the 15 computerized world how little data certain companies have 16 electronically and how much they have in hard copy, but where 17 the agencies have the advantage of having third party 18 discovery and third party data sets. 19 They have a good and And I think I'd give just some basic pointers to 20 talk about. I think that this is an area particularly where 21 the advanced work that the consulting economists or the 22 consulting legal team can do is to sit down and work with 23 their clients and identify what are the likely dispositive 24 issues, what are the testable hypotheses, 25 issues, is it private or is it geographic market, to set up For The Record, Inc. Waldorf, Maryland (301)870-8025 what are the key 16 1 evaluation of critical loss as to how many consumers would 2 have to leave to discipline pricing and then to work very 3 well with the people inside of the company to identify what 4 kind of data is available, who is knowledgeable about it and 5 to get as far up the learning curve as possible about the 6 data. 7 I think that will help them ultimately when the 8 staff at an agency wants data, that there is better 9 communication in terms of what is available. 10 I do think what is useful, though, having been an 11 assistant chief and worked on a very large number of mergers, 12 the approach I found I took after a couple circumstances of 13 unfortunately having asked for data and received it and then 14 found that we did not even have time to process it before the 15 decision had to be reached, before a recommendation had to go 16 out is I found what worked -- and this works on the outside, 17 too -- is to sit everyone down before the second request goes 18 out and ask two questions. 19 One is what are our testable hypotheses; what is 20 the key issue in this case. 21 unlimited computer resources, unlimited time, what would you 22 want to get to test those hypotheses; 23 And if we had unlimited data, And then to pose the second question, which is 24 then, we have a budget and a time constraint. 25 minimum sufficient data that you need to address that. For The Record, Inc. Waldorf, Maryland (301)870-8025 What is the 17 1 What I found consistently is that when everyone 2 focused together, oftentimes it was the case that people 3 could look at just the top 50 customers, what they had done, 4 who they had bought from, whether they had switched it or 5 not. 6 In a very large number of cases, one did not need 7 huge transactional databases of every single transaction that 8 anybody had ever done in the industry. 9 The other thing I found is that it served also to 10 meld the staff economists and the staff managers better with 11 the legal team. 12 Oftentimes there is a tendency in very complicated 13 industries where there is huge transactional data for the 14 economists to be doing a lot of empirical analysis that is 15 running on a separate track from the data and the documentary 16 -- the documentary and the evidentiary kinds of evidence. 17 And I think if you focused earlier on testable 18 hypotheses and minimum sufficient data, answers would be 19 available. 20 The last thing I say is one advantage that private 21 consulting firms have that the government typically does not 22 have is a lot of research assistants who are very skilled at 23 manipulating huge amounts of data. 24 25 And so to the extent, early on the FTC had to identify testable hypotheses data and asked that it be done, For The Record, Inc. Waldorf, Maryland (301)870-8025 18 1 to the extent it is doable, because more likely a lot of 2 processing could get done with results being given. In conclusion, I would just agree completely with 3 4 Mike. I think the process works best if there is a clear 5 dialogue, and that's my last set of words of don't assume. 6 I think there is a sense, because there is a very 7 good respect on both sides of the table for people, that 8 outside economists will understand or divine what the theory 9 of concern is. 10 I think the more that the staff early on can say 11 here specifically is -- you know, we are worried about -- if 12 I could fill in the blank -- price discrimination of this 13 sort, geographic market of this scope, customer concerns of 14 this type, the more it is that the outside economists can do 15 their best effort to put information forward. 16 If they are not able to address the concerns, I 17 think that gives you a better sense of having a strong case 18 for litigation. 19 Then you'll know what it is that you would be up 20 against in that litigation were it to occur or, more 21 fortunately, could hopefully resolve the concerns. 22 23 MS. COLEMAN: Thank you, Mary. Next up, I would like to introduce Richard Higgins who is with LECG. 24 He's going to talk about standard data -- and it's been 25 discussed to some extent already, but he's going to discuss For The Record, Inc. Waldorf, Maryland (301)870-8025 19 1 in more detail the issues that arise not only in the 2 collection of standard data, but how it is used throughout 3 the merger review process. MR. HIGGINS: 4 Well, actually, I was watching you 5 very closely, because that's the first I heard of what I was 6 to do. 7 John. I thought I was just going to wait and comment on 8 MR. WOODBURY: Oh, really. 9 MR. HIGGINS: So far I don't have anything to say 10 about what was said before, except that I disagree with Mr. 11 Denger about agreeing on IRI versus Nielsen. 12 of disagreement. I like the idea 13 And I mean, there are a lot of parties to mergers 14 and I just can't imagine getting them all in the same room 15 and telling them what data set they should be using. 16 And there are only two companies thanks to you all 17 -- that is to say, there would have been one had it not been 18 for you all. 19 And I wanted to also ask a question. Panel data 20 is something that the Nielsen people and the IRI people have 21 and I've never gotten to use it. 22 all ever asked for those data. 23 And I just wondered if you I've not seen the second request, which was asked 24 for, and I have found that the IRI and the Nielsen people are 25 becoming a lot more jealous of what they ownand, in fact, For The Record, Inc. Waldorf, Maryland (301)870-8025 20 1 the last couple of times I had to work with data, I had to 2 sign something. 3 time, but the agreement worked out. It really wasn't onerous, but it took a long 4 MR. COWIE: 5 MR. HIGGINS: Richard, what is panel data? Well, Neilsen, for example, has 6 surveyed or has a sample 40,000 households and it gives them 7 a little machine that they can scan their own products. 8 have their own scanner at home. So 9 They are asked -- and I guess they're paid 10 something -- they are asked to scan all of their purchases. 11 And the value of it is it's a different type of data; it's 12 not just the time series stuff that we all have worked with. 13 We usually have, you know, lots of time and 14 especially if it's weekly data, and not -- well, I guess 54 15 cities or whatever is enough, but this is -- I think these 16 data mainly address some questions that haven't been 17 addressed and, in particular, they -- I think that they 18 account for couponing. 19 You know, a typical problem of ours is we've got a 20 merger involving products that are couponed a lot and the 21 kind of data that we normally get doesn't have any couponing 22 and -- in the case of ready-to-eat cereals, 23 part of the price. 24 25 would be a large So it would be interesting to see what one could do with the panel data. It would be a big problem, but there For The Record, Inc. Waldorf, Maryland (301)870-8025 21 1 are a lot of good economists, so I'm sure they could work 2 something out. 3 MR. DENGER: Mary, could I ask one thing? 4 MS. COLEMAN: Sure. 5 MR. DENGER: On panel data, I know it's been 6 requested because I believe we've provided it. 7 MR. HIGGINS: Oh, you did. 8 MR. DENGER: In fact, we've even provided it in 9 Oh, okay. the case Richard was involved in. 10 But panel data can be useful because they'll 11 measure what consumer purchases are in a particular point of 12 time, say a six-month period, and what happens, say, a year 13 later and you can track brand-switching and elasticities and 14 look at the changes in price and look at not only couponing, 15 but other things, and see how particular consumers react to 16 changes in price or other developments with regard to the 17 brand. 18 It can tell you how they move from one type of 19 outlets to other types of outlets. 20 example, the degree in which brands are particularly close 21 competitors of other type -- of other specific brands. 22 23 It can tell you, for So I think it's a very robust type of database in a lot of industries. 24 MR. WOODBURY: 25 MR. DENGER: Can I disagree? Sure. For The Record, Inc. Waldorf, Maryland (301)870-8025 22 MR. WOODBURY: 1 It's not -- I agree that there are 2 some utilities with that kind of panel data, you know, the 3 switching analysis that Nielsen and IRI can provide -- and 4 I'm guessing. 5 - but I mean, it would be great if you could have access to 6 the underlying individual data that's contained, because then 7 you could do -- there are some problems, but you could do, 8 you know, consumer level demand analysis directly using 9 individuals, rather than cities, as observations. 10 I don't want to put words in Richard's mouth - And as Richard says, the problem with getting 11 access to the data, as far as I can tell, is -- the hurdles 12 are unsurmountable because they will never give it out. We thought occasionally about talking to IRI and 13 14 Nielsen about providing a demand analysis, of course, at our 15 specifications, but we never see the data. 16 to do it for us. 17 -- 18 But we get them So I mean, we've never gotten that far, but MR. HIGGINS: Yes. I think they've become quite 19 competitive providing their own analyses and that's one 20 treasure they have that they're holding close and then 21 request a particular study to be done. 22 MR. WOODBURY: One thing to be leery of, which I 23 discovered recently, is that if you look at the switching 24 data, you can actually derive market shares based upon the 25 data that Nielsen and IRI give you. For The Record, Inc. Waldorf, Maryland (301)870-8025 23 1 And if you compare those data to the national IRI 2 and Nielsen data, you will find that there's usually a 3 substantial difference. 4 representative, so you have to be really careful about how 5 you use those data, even in switching studies. 6 MR. DENGER: 7 all of that sort of thing in switching data. 8 that the data can give you some insights. 9 It turns out those panels are not And there's also issues as to sample size and And I agree with, Richard. But I think If you're talking 10 about the underlying IRI/Nielsen data, it wasn't provided. 11 But there are lots of switching studies, and so forth, based 12 on that data that can be very useful, I think, at least in 13 giving certain insights that you might not otherwise draw. 14 MS. COLEMAN: In those cases, have you had 15 documents of Nielsen and IRI to understand what it is that 16 they did to do the switching studies and information or did 17 you get the data -- even if it was aggregated, the data in 18 enough detail yourself to understand what was done? 19 MR. HIGGINS: I have had conversations with 20 business people, marketing people, but I don't think I ever 21 understood what was done. 22 MR. WOODBURY: I mean, we've had direct 23 conversations. I mean, I've had direct conversations with 24 IRI and Nielsen about some of the switching studies so I can 25 understand, particularly when I discovered that I was getting For The Record, Inc. Waldorf, Maryland (301)870-8025 24 1 market shares that looked a lot different from national 2 shares, what I thought was going on and how the sample was 3 collected, and all of that. 4 Yeah, but I think -- you know, they're certainly 5 happy to talk to you about it, about how those samples are 6 constructed. 7 MS. COLEMAN: I think that those -- I do think 8 those types of data, it's interesting, can be useful to get 9 at issues, but it encourages helping us to -- you know, as we 10 get that data, helping us to understand what the collected 11 data is and how it's collected and how these studies were 12 done. 13 14 MR. DENGER: But all data has, you know, limitations -- 15 MS. COLEMAN: Right. 16 MR. DENGER: -- and we all have to be aware of 17 18 them and have to understand to effectively use the data. MR. HIGGINS: Well, I just -- actually, I'm going 19 to sit down and let John talk about. 20 more to say than I do. 21 MR. WOODBURY: 22 23 24 25 I know he has a lot No, no, no, no. I'm sorry. I shouldn't have -MR. HIGGINS: No, no, no. I know you prepared something. But before I do, a question I want to ask about For The Record, Inc. Waldorf, Maryland (301)870-8025 25 1 standard data is are you ever going to ask for it anymore? 2 MS. COLEMAN: 3 MR. COWIE: Yes. I have one question. Why do you 4 disagree with Denger that you have to see if the parties will 5 use the same -MR. HIGGINS: 6 Oh, I just think if I were working 7 for a party that was opposing the merger, however incredible 8 that may be, that you wouldn't -- they should be able to use 9 what their client has. 10 MR. COWIE: 11 merger and the FTC -- 12 MR. HIGGINS: But you agree that the parties to the Yeah, but that means you have to do 13 it both ways because God knows what the third party will 14 show. 15 And the data are different. John will probably 16 talk about the geographic areas are different, Nielsen and 17 IRI, and how they do it. 18 different results from time to time using the two. 19 prefer to try to comment on whatever he says. 20 MS. COLEMAN: I don't really know, but we do get Okay. Well, with that, I guess 21 we'll move to John who is with CRA. 22 customer data. 23 24 25 MR. WOODBURY: He is going talk about Well, actually, I thought I was talking about the vulnerable consumers. MS. COLEMAN: I Right, right. For The Record, Inc. Waldorf, Maryland (301)870-8025 26 1 MR. WOODBURY: And how would you use the second 2 request data to better identify vulnerable consumers. 3 actually found myself sort of confused for a while, but I 4 think I worked my way out of the confusion. 5 6 7 And I What I am going to do is address -- try to address three points. First of all, what do we mean by vulnerable 8 consumers and what might we mean by vulnerable consumers; 9 second, how does -- the second request process, how can that 10 be honed to better identify who those vulnerable consumers 11 are; and third, and frankly and most importantly, from the 12 policy stand point -- and I know it's not exactly a part of 13 this -- but what the heck do you do with the vulnerable 14 consumers when you find them. 15 16 17 In terms of definition I sort of think there are three categories of vulnerable consumers. One is an elasticity vulnerability, a second is 18 an over-aggregation of vulnerability, and the third is, as we 19 already mentioned, the unmentioned vulnerabilities. 20 Elasticity vulnerability, I think, is the one that 21 economists naturally think of when they think about 22 vulnerable consumers. 23 undifferentiated product, you know, like you have a product 24 where there are two groups of consumers. 25 lower elasticity and then the other group. And then the lower These are -- think of an For The Record, Inc. Waldorf, Maryland (301)870-8025 One of them has the 27 1 elasticity consumers are potentially exploitable if a 2 hypothetical monopolists can discriminate between those two. He can discriminate between those two with two 3 4 conditions. 5 consumers are which and, second, the monopolists can prevent 6 -- hypothetical monopolist can prevent an 7 the two. 8 One is the monopolists can easily identify which arbitrage between So you find yourself in a relatively 9 undifferentiated industry where there are some groups that 10 have a relatively low elasticity with product. 11 thing to do is just basically apply the SSNIP test, the 12 standard, you know, market definition test to the lower 13 elasticity group and ask whether they constitute the same -- 14 if they constitute their own antitrust market, so you conduct 15 a merger analysis on that basis. 16 possibilities are -- well, I call them over-aggregation 17 vulnerabilities, but they actually just may be market 18 definition mistakes. 19 The first The other two In the undifferentiated product context, you might 20 find that, gee, you know, in looking at a particular input 21 some materials, you know, some chemical compound, you know, 22 we might conclude that the demand -- the antitrust markets 23 all supply us with that compound and all of the customers of 24 that compound are in the same antitrust market, but it may 25 turn out that for reasons that may have to do with differing For The Record, Inc. Waldorf, Maryland (301)870-8025 28 1 technologies, there is some subset of customers that, in 2 fact, can't use all of the material available, that their 3 range of choice is much more limited in that case, we've 4 identified another potentially vulnerable consumer group. 5 Similarly, if you're talking about geographic 6 markets, you might think that for some products that most 7 consumers are able to, say, input purchases or able to 8 purchase from anywhere in the United States because transfer 9 costs are relatively low, but in fact, there may be some 10 subset of purchasers for whom transport costs are pretty 11 severe and, again, their of alternatives are pretty limited. 12 And again, I call those over-aggregation mistakes, 13 but you know, they expose another set of vulnerability. 14 And the differentiated product contents of the 15 over-aggregation could arise because of our effort to try to 16 make the analysis trackable. 17 retail products, such as, you know, candy or cereals or 18 shampoo, the number of SKUs, you know, at the market level 19 can literally be in the hundreds, if not thousands, because 20 they come in different flavors, they look different, they 21 smell different, there are different package sizes. And if you think about some 22 So in an effort to make the analysis trackable, 23 you start aggregating up as some sort of mega-brand that 24 encompasses the whole slew of different package sizes and 25 different types of products. For The Record, Inc. Waldorf, Maryland (301)870-8025 29 1 In the process of doing that, we may have hidden 2 some consumers that were exposed because those consumers have 3 a relatively low elasticity for some of those products 4 compared to the average elasticity for that group as a whole. 5 Another kind of vulnerability that could result 6 again, and I'm thinking more in the consumer goods market, 7 suppose that -- suppose, in fact, that there are differences 8 in elasticities with distribution channels. By aggregating 9 all of the distribution channels will have hidden that 10 potentially vulnerable consumer. 11 Finally, there are these vulnerabilities that we 12 don't talk about very much. 13 gee, we're going to let this deal go through because there's 14 virtually no chance that a price increase will result and, 15 nonetheless, maybe some consumers are going to be harmed. 16 Suppose the agency decides that, If we're talking about a merger between retail 17 stores, there are some consumers who are going to discover 18 that post-merger, their favorite retail store is closed and 19 they have to further or go to a less preferred retail store 20 to purchase whatever it is they want to purchase. 21 They may -- again, post-merger there may be some 22 brands that are withdrawn from the marketplace and 23 consumers again will have to go to their second preference. 24 25 some And in many cases, we don't talk about who those consumers are or the fact they're out there, and I'm going to For The Record, Inc. Waldorf, Maryland (301)870-8025 30 1 come back to this in a second. Now, the anti-climatic part of this topic, since I 2 3 ignored those last vulnerability concerns, the 4 unmentionables, it seems to me that the second request as it 5 is now structured is actually a pretty good vehicle for 6 identifying vulnerable consumers because it solicits the kind 7 of information that at least lets you initially get a hold of 8 an estimate of who those consumers might be. 9 You know, in the three pieces of information, I 10 think that's solicited by the second request for documents, 11 they're data that could be used for econometric analyses and 12 they're surveys. Let me deal with this briefly one by one. In the case of documents, this is, you know, the 13 14 usual suspects, there are planning documents, there's 15 strategic documents, they're marketing documents. 16 documents that may enable the agency, as well as the parties 17 working for -- the merger parties to identify, you know, 18 efforts by the firm to segment the 19 groups. 20 alternatives available to different consumers. 21 They're market into different And they may highlight, you know, different They may actually talk about price differences 22 across what seems to be essentially the same product, but 23 across different consumer groups. 24 25 And it may be that there's enough information in the information provided to determine whether or not the For The Record, Inc. Waldorf, Maryland (301)870-8025 31 1 price cost margin differs across those consumer groups. For undifferentiated goods, we would be looking 2 3 for things in the documents that either talk about the 4 ability to arbitrage between, you know, low elasticity groups 5 and high elasticity groups or frustration at the inability of 6 resale from the guys that get it at a low price to the guys 7 that are going to be victimized by higher prices. 8 In terms of data -- and the data, again, can come 9 in a variety of flavors. 10 talked about which allows for direct estimation of own price 11 elasticity, but it also enables us to look 12 concerned about an aggregation, that we've aggregated the 13 brands too broadly, then you can use econometrics to test 14 that proposition. if we're Again, I'm putting aside the ongoing concerns 15 16 One is the standard data that we about how to interpret results from the scanner analysis. There's some groups of services that aren't 17 18 scanned. You know, there are other databases with market 19 data, you know, things like the airline data, and telephone 20 industry. 21 you to try to identify the extent to which you have 22 potentially vulnerable consumers. There are sample billings all of which go by -- 23 The one thing that isn't covered typically in the 24 second request, but I think would be useful in identifying 25 potentially vulnerable consumers are natural experiments. For The Record, Inc. Waldorf, Maryland (301)870-8025 32 1 In virtually every industry there are 2 unanticipated events that will affect pricing. These can be, 3 you know, the sudden exit of a player, a firm. It could be 4 the disruption of a distribution pattern or a temporary plant 5 shutdown, all of which can generate price increases. 6 And those allow you to determine, you know, are 7 there any difference in price effects felt by different 8 consumer groups; again, a signal that there may be more, 9 rather than less, vulnerable consumer groups. 10 Third, and I think this is really the large thing 11 to differentiate product surveys, surveys of the companies, 12 the merging parties themselves implement to try to understand 13 differences between consumer segments and whether or not, in 14 fact, there are strategies for separating those segments in 15 pricing differently between them. 16 All of that, of course, should be followed by 17 post-second request interviews of, you know, the rival 18 sellers, of customers, as well as the merging parties walking 19 them through concrete sales efforts to determine, you know, 20 exactly what kind of considerations play into this. 21 Now, once you've identified, of course, the 22 vulnerable consumer, that is at the end of the analysis, so 23 you have to ask the question, "Gee, if they pass the SSNIP 24 test, are there other alternatives that can prevent the 25 prices from increasing." For The Record, Inc. Waldorf, Maryland (301)870-8025 33 1 2 For example, there could be buyer strategies that could counter any attempts to engage in price discrimination. 3 4 But there is still something important, I think, 5 finance and policy questions, and I don't have any answers, 6 about how we weight some of these differences. 7 Let me give you the clearest example. 8 suppose that there is a merger between an upstream firm 9 decides to acquire one of its customers and it's that 10 customer that has been in an undifferentiated market, has 11 been playing the role of arbitrager between the high 12 elasticity groups and the low elasticity groups. 13 I mean, that So to eliminate that arbitrager, the upstream firm 14 acquires the downstream firm. 15 we'll have higher prices for the low elasticity guys, but 16 we're going to have lower prices for high elasticity guys. 17 The question is, well, how do we balance those -- 18 the gains versus the losses and should that play a role in 19 policy formulation, as opposed to sort of when we go forward 20 legally in court. 21 The post-merger effect is that And to give another example, suppose that there 22 are two firms that produce, you know, products A and B and 23 they merge. 24 variable cost synergies with production of product A. 25 And as a result of the merger, there are And as a result, the price falls and the cost For The Record, Inc. Waldorf, Maryland (301)870-8025 34 1 falls, but they fall in such a way, that the price cost 2 margin for product A is now higher than it was pre-merger. Well, now, the merging firm will have an incentive 3 4 to raise the price of product B, so as to divert customers 5 from B to the relatively higher margin of product A. Now, customers of A, they benefit by the merger; 6 7 customers of B are going to lose. 8 the losses to one group versus the gains of another group. 9 How do we weigh, you know, I don't think it's a defense to say that -- you 10 know, that we never do that merger analysis. 11 of the time. 12 ones that occur all of the time. We do that all And the unmentionable vulnerabilities are the So implicitly, there is some balancing of those 13 14 losses against the gains to the consumers who benefit from 15 the merger. I don't have an answer to the question. 16 I just 17 raise it. 18 be dealing with when they deal with the agencies: how do we 19 evaluate instances in which mergers will provide clear gains 20 to some groups of consumers, but the potential of loss is to 21 other groups. 22 23 24 25 But I think it's an important one for parties to Thanks. MS. COLEMAN: Thanks, John. I have one comment on the second request about natural experiments. What do you think could be done to improve that to make sure -- I agree that it can be very useful for a number For The Record, Inc. Waldorf, Maryland (301)870-8025 35 1 of things. MR. WOODBURY: 2 I think it would start to do things 3 like ask a party to identify -- take a list of potentially 4 disrupted events, say, plant shutdowns, recent plant 5 shutdowns, firm exits, sudden shortages of input and supply 6 and ask questions about have any of these occurred in the 7 past X-years and that will give you a basis to start. 8 9 MS. COLEMAN: And what kind of data would you MR. WOODBURY: Well, if you're doing econometrics, think -- 10 11 I mean, you think about doing before and after studies where 12 we look at the price before and after the event. 13 But it doesn't have to be econometrics. 14 simply look at averages because you can learn a lot of other 15 things. 16 the outcomes of these natural experiments even in somewhat 17 statistically more crude terms. 18 MR. DENGER: You can But it would be insightful effort to look at some of The only thing I would say, Mary, is 19 you may not ask for it specifically, but I think you pretty 20 adequately if you understand the second request bring within 21 the scope these natural experiments and other events in the 22 marketplace. 23 You may not know what they are unless you review 24 the documents, but I think that to the extent that there are 25 these events in the industry, you will pick them up. For The Record, Inc. Waldorf, Maryland (301)870-8025 36 1 MR. WOODBURY: Oh, yeah, that may well be true. 2 You do it for the second -- if I'm sitting with this guys, I 3 do a second request. 4 5 And you know that we on the other side would be looking for those things. 6 MS. COLEMAN: Does anybody have any -- 7 MS. GEURIN-CALVERT: I think one of the things, 8 too, is that this is an area where we're assuming that what 9 is going on consecutive with the second request is a lot of 10 meetings between -- or phone calls between the economists and 11 lawyers within the agency and the economists and lawyers on 12 behalf of the merging parties. 13 again, what is useful is to identify events that people think 14 really are representative natural marketing experiments. It would seem, 15 And I think where the -- what I would say as well, 16 is it is oftentimes very difficult at that point for anybody 17 to have sufficient data to be doing really good econometrics 18 in the sense that the merging parties in most industries do 19 not have a complete data set. 20 21 22 They've got their own data which can give some insights, but oftentimes is not sufficient. The agencies probably are not at a point to have 23 had the full data set either, and so the more people can come 24 up with what it is they expect the natural market experiment 25 to be showing, what it is they're looking for and, as a For The Record, Inc. Waldorf, Maryland (301)870-8025 37 1 result, where information can be brought to bear that is 2 empirical in terms of actual data analysis, but also 3 documentary, and so on, would be, I think, be very helpful to 4 try and potentially reach some closure as to whether or not 5 the next stage needs to be occurring. 6 MR. WOODBURY: 7 think that's absolutely right. 8 9 10 Actually, let me step back. I I focus on experimental -- you know, natural experimental context for identifying vulnerable consumers. But obviously, a natural experiments, you need 11 some referrals including data entry, looking at substitution, 12 and the like. 13 And I'm glad you pointed that out. MS. GEURIN-CALVERT: One example would be a lot of 14 manufacturing cases involve hundreds, if not thousands, of 15 consumers. 16 It is -- there are some industries where there are 17 three or four very, very substantial consumers. 18 typically, it is the case that there are at least a hundred 19 fairly good-sized consumers. 20 More And if the concern at issue is that certain 21 customers are more vulnerable to a price increase than 22 others, I think it's a useful point to lay out, you know, in 23 terms of -- particularly if the staff has at this point 24 significant complaints from customers is to identify we 25 believe from each segment of your customer base, we have at For The Record, Inc. Waldorf, Maryland (301)870-8025 38 1 least one complaining customer; you know, show us for each 2 customer grouping in the context of a natural experiment for 3 others why it is they have a plausible alternative other than 4 the merging parties. 5 I think again the more you can move from what 6 inherently is a customer database to the key issues, it helps 7 both sides. 8 MR. DANIEL: I would -- in a similar framework or 9 frame of mind, I would hope there would be open dialogue and 10 frequent dialogue between the agencies and the parties if the 11 vulnerable consumer issue is at work because it's a little 12 bit at odds with the Meg's point, which is maybe you can 13 learn all you need to learn from maybe the top 50 customers. 14 You can learn a lot from the top 50 customers, but 15 if there are 200 or 500 customers, you might be able to find 16 some down there that would not have as many alternatives. 17 But allowing for those discussions as early and as 18 frankly and as openly as possible animated with whatever 19 support you can share with the parties, given the 20 confidentiality of information, would really be helpful 21 because oftentimes those vulnerable customers pop up 22 relatively late in the process and then you try to figure out 23 how can we analytically track where they are and what their 24 options are. 25 MR. DENGER: That's why the interviews with the For The Record, Inc. Waldorf, Maryland (301)870-8025 39 1 customers is so important to understand if they feel 2 potentially victimized or why they feel vulnerable and the 3 source of that vulnerability. 4 MS. COLEMAN: Okay, thank you. 5 MR. WOODBURY: Richard. 6 MR. HIGGINS: I just want to know what an example 7 is of a vulnerable customer. 8 numbers. 9 10 MR. WOODBURY: You must have worked those I've never found a vulnerable customer -- no. 11 You know, it's the usual. It's the residential 12 versus business consumer, and for example, a long distance 13 company. 14 I mean, at least there's the conventional wisdom 15 that a residential consumer is more at the mercy of long 16 distance companies than the business consumer. 17 18 19 It's the leisure versus business travel with the airlines, you know, those kinds of distinctions. You know, it's the generic versus brand-named 20 drugs, those preferring the generic when the generics come 21 out, then the prices rise. 22 MR. HIGGINS: I thought it was something new. 23 MR. WOODBURY: No. 24 MR. HIGGINS: All right, thank you. 25 MS. COLEMAN: Next up is Tim Daniel from NERA He's For The Record, Inc. Waldorf, Maryland (301)870-8025 40 1 2 3 4 going to talk about a different type of data. MR. DANIEL: Thanks, Mary. Thanks for the opportunity to be here. Switching gears a little bit, I guess, and what 5 Mary has asked me to talk about involves mergers where the 6 competition arises in bidding- or auction-kind of 7 environments. 8 In looking at some of the second requests that 9 we've worked on where auction or bidding issues have arisen, 10 the bid specs, if there are such things, they can be pretty 11 sparse and limited and I don't think that's necessarily a 12 failure of the second request so much as a recognition, if 13 you will, that when bidding is involved -- what I'm thinking 14 of here is not so much bidding in a formal process, as the 15 government tenders, where you're going to know exactly what 16 the bid is for, but rather, competition that occurs for a 17 durable product in the medical business for a CAT scan or a 18 purchase of electric generation equipment by a utility where 19 it puts out a request for bid in a less formal sense, but the 20 competition is very much in that kind of framework where the 21 sellers are negotiating with the customer, and vice-versa, 22 and that's how the competition is described. 23 The data on that kind of competition are often 24 pretty idiosyncratic to the firm at issue and sometimes even 25 within the firm are very different qualities of data For The Record, Inc. Waldorf, Maryland (301)870-8025 41 1 collected across different product lines and divisions and 2 geographic areas. And I think that what I would take from that at 3 4 least in the process of doing a merger investigation is that 5 there's even a bigger benefit early on after the parties and 6 the government sit down and try and understand what data the 7 companies do, in fact, have when the competition occurs in 8 that kind of a bidding environment because it can be very 9 spotty in certain areas, it can be quite incomplete and it 10 can be quite frustrating down the road. 11 If the expectations are higher, I mean, I think 12 that the parties would keep good records on their bidding 13 competitions, good records on what their rivals are bidding, 14 good records on why they won or why they lost certain bids, 15 when, in fact, that's not the case. One option I think that I would entertain early 16 17 on, therefore, is if you do have those dialogues and discuss 18 what is available in the companies is if you do hone in on a 19 particularly rich set of data, be it on a limited product 20 line or a limited time period or a limited geographic area, I 21 think the discussion could ensue at that point as to whether 22 or not a careful analyses of the good data might provide 23 enough of a basis to generalize to the markets what are at 24 issue. 25 That's a potentially difficult discussion to have, For The Record, Inc. Waldorf, Maryland (301)870-8025 42 1 but I think one worth thinking about. It's essentially the 2 same, I guess, as to whether you can take a sample of the 3 data. If you happen to have data records that might be 4 5 complete, but they're on all paper and they're, again, 6 difficult to deal with, they're not Nielsen data, they're not 7 IRI data, they are just messy bid records, and you want to go 8 back five or ten years, or whatever time period you're 9 interested in, I mean, it might be possible early on to agree 10 to take a sample of those records. 11 When I was here at the FTC and doing consumer 12 protection matters and we talked about discrimination and 13 lending practices, that was a common occurrence when you had 14 a large number -- thousands and thousands of applications for 15 credit to just take a sample of those, work with the parties 16 to draw up a representative sample and then agree that those 17 are the data that are going to form the basis of the analysis 18 or for the decision at least up through the investigative 19 period. 20 One issue that I think is particular to an auction 21 or bidding environment that has to be taken into account is 22 if you look back at the economic models and ask what is the 23 likely price effects from a merger when two firms get 24 together in a bidding situation; is the expected price 25 increase from a unilateral context. For The Record, Inc. Waldorf, Maryland (301)870-8025 It depends critically on 43 1 the costs and the offerings of the non-merging parties. 2 So to get the two low cost firms in a bidding 3 competition merged, your typical concern is they're going to 4 raise the price up to the cost of another third-level lowest 5 cost provider. 6 What that means is that very often there's an even 7 bigger asymmetry, I think, between the information available 8 to the government who can obviously subpoena and gather 9 information from all of the major players in the market and 10 the information immediately available to the parties. 11 That's not a gap that's going to be necessarily 12 closable, but I think it needs to be recognized and, again, 13 to me, at least counsels for as much open dialogue and as 14 much disclosure on the part of the agencies as to what's 15 animating their concerns, what's driving their analytical 16 framework and why they came and those competitions in those 17 situations of merger might be likely to raise prices. 18 Virtually, everything else I have on my short set 19 of bullets has been raised already, and so I think I'm going 20 to leave it there just with an endorsement of what Mike had 21 said about having open dialogue and discussions with the 22 agency as early as possible, agreement, if possible, between 23 the agencies -- agency and the parties regarding what data 24 will be sufficient to test the hypotheses and, if possible, 25 early on to agree on the economic model at issue because the For The Record, Inc. Waldorf, Maryland (301)870-8025 44 1 economic models that underline an auction or bidding 2 environment are very different from the models underlying 3 standard differentiated product or homogenous product merger 4 concerns. 5 And if those agreements can be reached up front, 6 yes, indeed, this is the kind of analysis that we need to do, 7 then that can help to hone in on what information the parties 8 will have as a matter of course and what part of the 9 information the parties will not have, particularly on the 10 cost of their rivals as a matter of course, and then trying 11 to deal with some of the information and asymmetries that I 12 think are even more pronounced in these kinds of mergers and 13 these kinds of investigations that are made. 14 MS. COLEMAN: Okay, thank you. 15 MS. GEURIN-CALVERT: Does anyone -- I just want to echo what Tim 16 said about -- I think it is particularly the case where the 17 evidence to resolve the specific concern does not reside 18 within the merging parties and is not likely to be readily 19 available in public form as data on cost, but could only 20 maybe be inferred from particular things. 21 It is very important, I would echo his words, to 22 identify what the specific theory is because one could think 23 in a bid model a concern that it really is the case that two 24 merging parties, as Tim expressed, the lowest cost and that 25 there is no real sense that the non-merging parties may have For The Record, Inc. Waldorf, Maryland (301)870-8025 45 1 in any way changed their cost position from the last time the 2 bidding occurred or could not do so somehow post-merger, 3 there really is a circumstance where there is a reason for 4 concern. 5 I think the more it is articulated that the 6 concern is a cost difference as opposed to, let's say, a 7 likelihood of somebody even submitting a bid. 8 One of the things I think that I found both in the 9 antitrust division and outside is people spend a lot of time 10 focusing on the wrong question because they thought it was 11 the correct question, that they thought, you know, the 12 agency's concern was we want to know about likelihood, as 13 opposed to we want to know about magnitude or we want to know 14 about likelihood and magnitude. 15 And I know that sounds relatively simplistic, but 16 I think that the more the questions can be articulated very 17 clearly and early on, the more there is the opportunity to 18 come back with information. 19 The other thing is I think it highlights the 20 circumstance where transactional level data, particularly if 21 it's old, may or may not be informative of what the post- 22 merger competitive process will be like. 23 If it's an industry that -- where the occurrence 24 of who has won in the past is different from, say, who hasn't 25 won in the last year, you can end up with a very different For The Record, Inc. Waldorf, Maryland (301)870-8025 46 1 merger result than if you -- or merger prediction than if 2 what happened in the last year is almost identical to what 3 happened in previous years. 4 So I would respond to Tim's statement a little bit 5 that in terms of viewing any sampling, it would be worthwhile 6 to give the parties an opportunity to address whether or not 7 the past is truly a prologue or not. 8 9 10 MR. DANIEL: In the natural experiment, again, I would raise the same point. A lot of the markets with large durable medical 11 equipment, electrical equipment have seen a lot of mergers in 12 the last half dozen years and someone might be able to use 13 that information and a change in market structure to see 14 whether or not past mergers, of which this current merger in 15 front of you may be somewhat similar in appearance, one might 16 be able to do some kind of prediction with these models. 17 MS. COLEMAN: Thank you, Tim. I'm going to switch 18 gears a little bit here and move more to the accounting 19 financial side. 20 David Painter with LECG, also formerly with the 21 FTC, is going to talk about some of those key issues that 22 arise looking at a different type of accounting and financial 23 data that he worked with. 24 MR. PAINTER: And I appreciate this opportunity to 25 speak to you on this, as well. For The Record, Inc. Waldorf, Maryland (301)870-8025 47 1 I guess let me reiterate at the outset something 2 Mike said and I think everybody else has certainly agrees 3 with and that is that having been here for 30 years, the FTC 4 is not the enemy. The facts are the facts. And I think that while I'm going to speak to 5 6 financial and accounting data, it can probably apply to 7 anything. 8 that, for merging parties to take is to provide the factual 9 information they think the Commission is seeking. And that is the best strategy, if I can call it 10 I think it behooves the Commission and certainly I 11 think the merging parties, the lawyers representing merging 12 parties should make every effort to have a sharing of 13 theories because, clearly, a sharing of theories, once the 14 theories are honed in on, allow them to come up with the 15 accounting, with the factual information of what is financial 16 and accounting information, the Nielsen information or 17 anything else. 18 theories as early as possible is something that would be 19 essential. 20 theories are, that it behooves the merging parties through 21 the lawyers, through retained consultants to provide a 22 crystal clear road map and analysis that takes that factual 23 information and melds it into something that presumably comes 24 up with some conclusion and explain it to the Commission. 25 So I think that the sharing of And I think last recognizing what those With that said, let me just sort of tell you that For The Record, Inc. Waldorf, Maryland (301)870-8025 48 1 I'm going to limit my focus because so much else has been 2 said, as well, the use of accounting and financial 3 information for purposes of preparing an expert analysis, 4 such as a failing company and efficiencies or for that 5 matter, econometrics, or anything else. 6 And I guess I've been on the outside for five 7 years, but I was with the Commission for 30 and perhaps that 8 30 years is what's going to drive me to sort of suggest the 9 conclusion. 10 It was a long time in trying to get there. My experience at the Federal Trade Commission, and 11 it's confirmed I think in large part on the basis of my 12 experience on the outside, is that the underlying support 13 that is provided by the merging parties with respect to any 14 expert analysis is too frequently -- not always, but 15 frequently incomplete or inadequate for the Commission staff 16 to try, if they might, to be able to get a complete 17 understanding of the analysis in a way that says -- allows 18 them to say it's reliable. 19 it reliable. 20 I mean, this is the key thing: is And that isn't because there is some inadequacy in 21 the second request. The second request is as thorough as is 22 imaginable and, in fact, I just recently right before I came 23 here looked at the second request, the one that was received 24 by the merging parties in a case that I worked on, and 25 essentially it asked for the identification of all experts For The Record, Inc. Waldorf, Maryland (301)870-8025 49 1 and consultants retained by merging firms and all documents 2 and data that were provided to such experts and consultants 3 and also asked for all instructions, programs and other 4 documents that are necessary to use and interpret the 5 finished analyses prepared by these consultants and the data 6 that was used to prepare the finished analyses. 7 covers it all. 8 And that And yet, despite that -- and I'm assuming this 9 dates back certain years. 10 was similar when I was here -- expert analyses often are 11 fruitless. 12 It wasn't that exact language. Efforts are made to submit them in a way that's 13 understandable. 14 them in a way that allows them to factor it into the 15 decision. 16 and my feeling is that what's lacking is the road map. 17 Efforts are made by the staff to understand But somewhere along the way, something is lacking, And I think we all might be able to better 18 appreciate this if you look at the Efficiencies Amendment, 19 the Efficiencies Amendment before it came out, before 20 submitting anything and everything that we could to try to 21 demonstrate efficiencies. 22 It And we all had a sense that they had 23 practitioners on the outside, most of us that worked on the 24 inside what was needed with the Efficiencies Amendment and 25 sort of laid out a road map as to what is going to count with For The Record, Inc. Waldorf, Maryland (301)870-8025 50 1 respect to at least the efficiency claim and this could have 2 been applied. 3 You could have an econometrics amendment or you 4 could have failing company amendment in some respects, and so 5 forth, but the bottom line to it was that -- what it said was 6 that, look, we know we're requesting everything in the second 7 request, it gets in everything, but despite that, independent 8 of the second request, we are going to tell you now that you 9 have to substantiate the efficiencies in a way that allows 10 the staff to verify it. 11 12 13 You know, I think it could apply it to anything, I think, and that's the key. And so I think that from my standpoint, I think 14 that there are lots of shortcuts with respect to accounting 15 information data. 16 We're all aware of -- that doesn't mean that 17 they're not -- that they can't be meaningful when put into an 18 expert analysis. 19 shortcomings are and what the strengths are and coming to 20 some sort of a reasoned conclusion that allows the staff to 21 say yea or nay on it. 22 It's a matter of explaining what the I will tell you that if all of you are going to 23 count on the outside with the submission of boxes and boxes 24 of documents that underlies the efficiencies claim, it's not 25 going to do the trick. For The Record, Inc. Waldorf, Maryland (301)870-8025 51 And the staff is not going to be able to uncover 1 2 it. 3 to do that through depositions and investigational hearing. 4 I mean, I was here for 30 years and we tried desperately You know, as cooperative as the merging parties 5 attempt to be and the lawyers representing the merging 6 parties attempt to be, it's just extremely difficult to sort 7 of get at the subtleties, get at the sort of fine-tuning of 8 what these analyses are all about in a way that allows you to 9 say, hey, it's reliable. 10 So here is what I am suggesting. I am suggesting 11 that the Efficiencies Amendment, whether formally or whether 12 on the outside, effectively use it as a tool, as a backdrop 13 for what we submit to the staff on other kinds of analyses 14 involving accounting and financial data be the -- be an 15 additional standard, if I can call it that, that standard 16 being a substantiation that allows the Commission to verify. 17 And I can tell you having worked on the Staples 18 case and having worked on the Heinz baby food case that it 19 isn't a matter of whether you are submitting five boxes of 20 documents or five hundred boxes of documents to support your 21 position; it's the road map. 22 On Staples, I worked, I'm sure, a thousand hours 23 trying to go through hundreds of boxes of depositions, 24 hearings, everything else to get to the truth. 25 And even there, half of the efficiencies were For The Record, Inc. Waldorf, Maryland (301)870-8025 52 1 effectively dismissed because there was no way of verifying 2 whether they were valid or not. 3 On the other hand, I worked on the Heinz baby food 4 case and there were a fraction, a small fraction of the same 5 number of documents and -- but there was a road map, I 6 believe. 7 Now, that doesn't mean that the Commission 8 accepted it. 9 don't think there was a dispute as to the numbers themselves 10 and what the numbers were saying. 11 They didn't. We all know that. But at least I There were theories that effectively refuted what 12 was being presented. 13 then I think we can make some reasonable arguments for and 14 against. 15 And if we can get down to that, I mean, So I think that if we recognize that what's needed 16 is a road map, what's needed is really some really crystal 17 clear explanation, we realize that most of these analyses 18 that we sent to the Commission were started long before the 19 initial filing even took place. 20 We recognized that we have the benefit on the 21 outside of talking to business people at length, ad nauseam, 22 you know, time and time again. 23 We were able to pick and choose based on those 24 discussions all of the documents we wanted. 25 these constraints are, time constraints are with respect to For The Record, Inc. Waldorf, Maryland (301)870-8025 And we see what 53 1 the HSR process. 2 I think we can better appreciate what the 3 Commission is faced with and we can better serve our clients 4 because we have a strong incentive to get this merger to pass 5 muster, so we have a very strong incentive to make it crystal 6 clear to the Commission. 7 We can better serve our clients by getting these 8 things done in a way that allows them to have -- to be given 9 some weight. 10 11 MR. HIGGINS: But they never give us their -- the benefit of their advantage. 12 MR. PAINTER: The Commission staff? 13 MR. HIGGINS: Yes. They're sitting there with a 14 document from Bicardi that talks about whether they're going 15 to make the pineapple rum or not. 16 us if they're trying to get a rum deal. 17 MR. PAINTER: They're not going to tell You know, again, I can't speak for 18 the staff here, but I can tell you that when I was here, 19 while I may not have named parties and I might not have named 20 names, I think that, conceptually, the notion was thrown out 21 that, you know, what would preclude somebody from entering, 22 you know, by some means -- I mean, it was obvious -- and 23 maybe those exact historical examples that so many of you 24 have talked about, you know, were actually identified in a 25 request and explanation was made. For The Record, Inc. Waldorf, Maryland (301)870-8025 54 1 Yeah, it may not be an absolute perfect two-way 2 street, but I don't think that -- since we want this merger 3 to get through, you know, it's just something we, perhaps, 4 have to face. 5 MR. COWIE: Richard, there are confidentiality 6 concerns that at times prevent us from telling you what kind 7 of rum Bicardi is planning to make. 8 The fact of the matter is it's very difficult to 9 share third-party information. 10 MR. HIGGINS: I've been bluffed by Mr. Newborn 11 enough times to know there's more to it than that. 12 suggesting that it is an adversarial situation. 13 contest and I don't see where we should -- 14 MR. PAINTER: But I am It is a You know, again, I will tell you 15 from experience, I mean, obviously, this is a short -- 16 everybody has their own view, but if it becomes a contest and 17 you have an economic analysis that can't be -- you know, that 18 the staff is unable to sort of discern whether it's reliable 19 or not, then your best is case is going to court. 20 That's your best case because I don't think 21 they're going to accept it and go through, you know, without 22 an understanding of whether it's really, you know, an expert 23 analysis that's presumably supporting the arguments that 24 we're making to the staff and they're not sure if it's 25 reliable or not. For The Record, Inc. Waldorf, Maryland (301)870-8025 55 1 So, you know, maybe we can win in court and maybe 2 -- you know, I may be saying tomorrow the direct opposite, 3 but if we want to take our chances in court, then obviously 4 that's something that's factored into the overall strategy 5 from day one when we make the initial filing. But again, I don't want to -- I think that's more 6 7 the exception than the rule. I think that the facts are 8 going to come out. 9 late, so that the theories -- you know, it's better that the 10 explanation theory comes out early, so that, in fact, there 11 are valid considerations on both sides, that they can be 12 explored. It's better that they come out early than 13 MS. COLEMAN: Okay. 14 MS. GEURIN-CALVERT: 15 you. 16 it is so difficult to do. Mary. I in many respects agree with I have one point of disagreement just because I think 17 I think your idea of having a clear road map as to 18 what is the task, how do we do it and to be able to present 19 it in a form that stays objective and as verifiable as 20 possible is the best for everyone because I think that's 21 ultimately, as well, what is going to be presented in court 22 if it has to get there. 23 That's the same standard. I think where it is useful to have this workshop I 24 think where it is particularly difficult now to apply that 25 road map, although I think the most effort needs to go -- is For The Record, Inc. Waldorf, Maryland (301)870-8025 56 1 in large transactions where there is not a Nielsen and an IRI 2 or an industry database where what you are working with is a 3 huge number of consumers, very large and messy transactional 4 or customer databases in the merging parties so that even as 5 the economists start six months ahead of time working with 6 the most accommodating client to say let's understand what is 7 going on here, to the extent -- and this just is picking up 8 on what John said -- to the extent for whatever reason, 9 whether it's third-party information, or otherwise, the staff 10 has a concern about a particular customer group that they 11 believe has inelastic demands and a particular other group 12 they view may have more elastic demands. 13 I think the sooner that that can get articulated 14 by the staff to say we are particularly concerned about this 15 kind of set of customer, then I think it is more feasible for 16 the economists on the outside to try to marshal the data and 17 put it together in as testable form as possible. 18 But I think it's frustrating for everyone if the 19 way a meeting or a set of meetings proceeds as we have an 20 overall concern about lack of availability of substitutes. 21 I think the more progressed in the direction is 22 something you indicate on the efficiency side where everybody 23 agrees on what are we looking for, the better 24 industries. 25 and it's more straightforward for the staff to ask for the in some It's very straightforward to do that early on For The Record, Inc. Waldorf, Maryland (301)870-8025 57 1 information in the second request. 2 And I would just say where you're having to ask 3 for huge transactional databases and it's a wide range of 4 industries it helps everybody to give more guidance from both 5 sides. 6 MR. PAINTER: Absolutely. 7 MR. DENGER: If I could add just one thing here 8 and it transcends data issues, but I think it's really 9 important. 10 I think the key thing is everyone wants to get a 11 sound merger policy and to do that, you have to have dialogue 12 on the issues that are of concern. And I agree totally with Dave to the extent that 13 14 we can learn from the parties to the mergers with respect to 15 what the concerns of the agency are early on. 16 that there are confidentiality concerns that you have to 17 protect. And I realize 18 On the other hand, you have to be careful that you 19 don't use confidentiality concerns as an excuse for blocking 20 a dialogue. 21 think, surface issues oftentimes. 22 You can protect confidentiality, but still, I And nothing is more frustrating to a party to a 23 merger as to make a presentation and not really understand 24 what the concerns are coming back. 25 And I would urge the agency -- and I realize For The Record, Inc. Waldorf, Maryland (301)870-8025 58 1 oftentimes you don't fully understand all of the concerns 2 early on in the process until you've conducted your 3 investigation and talked to customers and so forth -- but I 4 would think that the thing that will make for the best merger 5 policy for the dialogue is to let's take off our litigation 6 hats just for a little bit at the beginning and try to have a 7 dialogue where we can understand the concerns of each side 8 and try to address them and try to provide the substantiation 9 if we can and I think that will make for the most 10 constructive approach in the data areas, as well as others. 11 MR. WOODBURY: But if I could echo, you know, a 12 little bit of Richard's presentation and focus on 13 econometrics, for example, I know CRA, and working with 14 Richard on some projects, LCEG has produced you know, pretty 15 detailed road maps. But if I can make an analysis given the agency 16 17 given the underlying data that -- and frequently they come 18 back and say well, we ran something a little bit different. 19 Well, what was it? Well, it was just a little bit 20 different, and you get frustrated because you don't 21 understand what they do, you know, how sensitive. 22 no idea how different it is. 23 You have And this has been an ongoing -- as you probably 24 all know, an ongoing issue, particularly scanner data 25 analysis. In the mass econometric projects, it would be nice For The Record, Inc. Waldorf, Maryland (301)870-8025 59 1 to have a completely two-way street. And you're right, you can't win the contest. 2 3 We're bound to lose it. 4 can't understand sometimes what's going on on the other side 5 in terms of why they find our results, you know, not 6 credible. 7 MS. COLEMAN: But it's just frustrating that we It's very clear from the discussions 8 that have been happening that the dialogue process is the key 9 to all of this and this is something that (inaudible) data 10 myself, trying to encourage as much as possible and as early 11 as possible, you know early discussion among the economists 12 and certainly the lawyers, as well, about the issues in 13 trying to come to agreement or at least an understanding of 14 what the theories are and what data tasks and other types of 15 analyses would be useful. 16 And also, we'll encourage to the extent we can, 17 given the confidentiality concerns, potentially where we are 18 in the process if it's got a very late date, this dialogue, 19 you know, the results that we're finding in as much detail as 20 we can so that those types of issues don't arise. 21 It's clearly something that both Dave and I have - 22 - are encouraging as much as we can. 23 continue to encourage that. And we're going to 24 Allen, did you have a comment? 25 ATTENDEE: One comment from the perspective from For The Record, Inc. Waldorf, Maryland (301)870-8025 60 1 someone on the inside. So many times parties have come up with, let's 2 3 say, white papers with documentation where the white papers 4 documentation would have been very effective but they tend to 5 come after everything has gone to the Commission and at the 6 last minute they come forth with the arguments. 7 late. 8 That's too The one time I was working on a case where I 9 thought the expert was most effective was when they received 10 the materials very early while we were still in the midst of 11 the investigation and that expert's analysis went a long way 12 toward resolving the concerns and getting the merger through. 13 14 There have been times when I've seen parties come 15 up with some material that really devastated the case or sort 16 of devastated the case, but it came up after all of the 17 memos, including the Directors' memos were forwarded to the 18 Commission, and it was too late. 19 MS. COLEMAN: I would, again, comment that the 20 later that these types of analyses come in, the less weight 21 they're going to be given and the less chance that we'll have 22 to not only look at them and the dialogue back. 23 That leads me to the question I sort of wanted to 24 throw out. We have a few more minutes. 25 and some of the other workshops point of issues have been For The Record, Inc. Waldorf, Maryland (301)870-8025 The timing of data 61 1 raised, nature of the data request and how long it takes to 2 gather the data particularly when we ask for it in ways that 3 the company's keep it and we end up getting the data 4 sometimes when the compliance happens and the data ends up 5 being a waste of everyone's time because by the time we get 6 the data, when it's that late, we can't do anything with it. 7 It's just too late in the process. 8 throw out a question about what can be done to encourage 9 people to submit data earlier as they get it and to enable 10 the process to be able to actually gather it early enough. 11 MS. GEURIN-CALVERT: So I wanted to One thing that may not work 12 in every case, but let's assume we're at the second request 13 stage or a point where there's been, you know, a voluntary 14 agreement for data, I think that is exactly the time to sit 15 down first internally and ask what exactly do we want, what 16 are we trying to test, what are our highest priority types of 17 information. 18 Obviously, in industries you've worked in before, 19 you know what you're looking for. 20 worked in before, you're less likely to know. And then to sit 21 down and have that dialogue with not just the attorneys, but 22 also the economists on the other side, and to identify as 23 clearly as possible what kinds of levels of data you're 24 needing in the highest level of priority and have the 25 shortest turnaround that you can to get back -- either to For The Record, Inc. Waldorf, Maryland (301)870-8025 In industries you haven't 62 1 have a discussion just with that group again or to identify 2 the relevant business people. 3 effective, both on the outside and in the government, to go 4 into the equivalent of a plant tour, to go do a data tour and 5 have people go down and see what are the forms in which the 6 data exist and that I have found expedites, even if it's on a 7 rolling production, data as quickly as possible in as 8 formatted a form as possible to get to you. 9 I have found it's And I think that covers also the issue of the 10 white paper. 11 not gotten the data either in a form that's really usable 12 much in advance of the time frame in which the staff gets it; 13 and so, again, to the extent issues can be identified and the 14 data identified sooner. 15 Sometimes it is that the outside parties have But I think having conference calls and visits 16 with the data people is one way to expedite it and try as 17 quickly as possible, Tim's point, if sampling is going to 18 work, have somebody come up with an agreed-upon method as 19 quickly as possible if it has to be the universe. 20 I was in one negotiation a long time back when I 21 was an assistant chief and it turns out that the parties had 22 a very small data set that was exactly what we wanted, that 23 they were able to turn over on a diskette the next day. 24 25 But that we never would have found it if we hadn't been kind of sitting and having that kind of dialogue. For The Record, Inc. Waldorf, Maryland (301)870-8025 So 63 1 that's what I would recommend, particularly if you're looking 2 at large companies that have not done a lot of work in the 3 industry before. MR. DENGER: 4 5 Let me just add one other thing, that we have used lots of times, and I think it's very helpful. The first meeting we come to to negotiate; the 6 7 second meeting we use to bring a couple of people who are 8 very knowledgeable about the data, samples and everything 9 like that. 10 And oftentimes, you can say here's what we got 11 from this person and you say, well, we can't do it that way, 12 but we can do it this way, and it will cost us X-zillion 13 dollars to do it this way, but if you can take this, we'll 14 provide it to you very quickly. And if you take -- and a lot of people like to 15 16 bring very knowledgeable data people to the first meeting. 17 But I think it oftentimes helps to focus the 18 request to get it in quickly to weed out a lot of data that's 19 really not going to be meaningful because even if you've had 20 it all, there are some limitations to it that effectively 21 preclude you from using it for the purposes you might want to 22 use it. 23 And rather than have the lawyers go back and try 24 to find people in the company, and so forth, assume you've 25 got some knowledgeable people, bring them along and let's get For The Record, Inc. Waldorf, Maryland (301)870-8025 64 1 the dialogue going and here's what we got and here's what we 2 can do for you and then if we can't do this, what about this 3 as an alternative. And I think sometimes the lawyers get in the way 4 5 and slow things down by not having the business people there, 6 and you can bring your accountants -- to tell you we looked 7 at the data and this won't work. 8 Give me this type of economic analysis and this 9 will work and I think that speeds things along and allows a 10 more meaningful focus earlier in the process. 11 MS. COLEMAN: 12 MR. DAGEN: Does anyone else have any questions? Just there aren't as many attorneys 13 here as maybe there might be, but is there a reason why a law 14 firm wouldn't want to engage in the process you're 15 describing, Mike? MR. DENGER: 16 Yeah, there are a lot of reasons. 17 One they can think they'll be in litigation with you and they 18 will think that information flow is a one-way street as you 19 sit there and you talk and you talk and you talk and you 20 don't get any dialogue coming back from the concerns. 21 There are a whole host of reasons why you wouldn't 22 do it. You would be worried if you bring some businessman, 23 there is always a risk that he or she could say something 24 that you may not wish that they would say and you might 25 rather have that occur in a context where you can control it For The Record, Inc. Waldorf, Maryland (301)870-8025 65 1 a little more. But I think in balance, when you have to weigh 2 3 these, I sort of go back to the day that let's not start out 4 looking at each other as the enemy. You've got a job to do and we have a transaction 5 6 we want to clear and let's try to see if we can have an 7 informed merging decision. 8 And the way to do that with the least burden and 9 the least cost on everybody I think is to step up to the 10 plate as early on the process. 11 step-up. And it's got to be a two-way 12 The agency has to be -- the personnel have to be 13 reasonably candid about what their concerns are as soon as 14 they have them keeping in mind the confidentiality respect. 15 MS. COLEMAN: 16 questions or comments? 17 (No response.) 18 MS. COLEMAN: 19 coming. 20 your input. 21 Anyone else? Okay. Does anyone have any Well, thank you all for This wraps up the workshop sessions. We appreciate (Whereupon, the proceedings concluded.) 22 23 24 25 For The Record, Inc. Waldorf, Maryland (301)870-8025 66 1 C E R T I F I C A T I O N O F R E P O R T E R 2 3 WORKSHOP ON SECOND REQUEST/MERGER 4 INVESTIGATION BEST PRACTICES 5 DATE: JULY 10, 2002 6 7 I HEREBY CERTIFY that the transcript contained herein 8 is a full and accurate transcript of the recording/stenomask 9 taken by me at the hearing on the above cause before the 10 FEDERAL TRADE COMMISSION to the best of my knowledge and 11 belief. 12 13 DATED: JULY 10TH, 2002 14 15 16 WANDA L. GRANGER, CVR 17 18 19 20 21 22 23 24 25 For The Record, Inc. Waldorf, Maryland (301)870-8025 --- Source: Frix Law Library, https://www.frixlaw.com/law-library/documents/agency%3Aftc%3A67e6526af395f187. Public record. Not legal advice.