US v American Express
An oral history of why lawyers should leave economics to economists at trial1
Cross-examining the opponent’s economic expert can present unique opportunities. Cross-examination can speak to all issues in an antitrust case, from market power and competitive effects to pro-competitive justifications. It can help the court understand the importance of certain facts to your case. But it can also hurt your case and credibility. Examining an opponent’s economist is akin to dancing the tango with a cobra: no matter how many skilful moves you have, there’s a better chance you’ll get bitten. For two days in the summer of 2014, I danced with a cobra.
On the morning of 6 August 2014, I was one month into the Department of Justice’s seven-week trial against credit card titan American Express, and had spent the past day or so cross-examining Amex’s economic expert, Professor Richard Gilbert. I asked him about “the quiet life”, a concept described in a 1935 paper by Sir John Hicks entitled “Annual Survey of Economic Theory: The Theory of Monopoly.” I immediately regretted it.
But, perhaps not surprisingly for those who know me, I stubbornly asked four more questions before I gained my senses and moved to a different topic. While those few questions might seem irrelevant in the hundreds of hours of trial, they had the potential to unravel the concessions I had won – and the credibility I had earned with the court – over the preceding days and weeks. All of the government’s hard work to challenge the company’s merchant rules could have been washed away in minutes.
This, then, is the story of those minutes: how seemingly minor moments can have significant consequences at trial, and how preparation can keep those minor moments from becoming major ones.
Beginning somewhere near the beginning
My part in this lengthy and complex story started soon after 4 October 2010, the day the DOJ (and seven states) sued the country’s three largest credit card networks (Amex, MasterCard and Visa). The story goes: every time a consumer uses a credit card, the merchant must pay a fee. These fees are a significant cost for merchants and, as such, the government alleged that merchants generally pass these fees on to consumers through higher retail prices.
Amex, MasterCard and Visa had rules that prevented merchants from trying to reduce those fees by playing the credit card networks against each other. For example, merchants could not encourage their customers to use a particular credit card, such as one with lower fees for the merchant. Because customers – not merchants – make the ultimate choice of which credit card to use, the credit card networks’ limitations on “merchant steering” all but eliminated merchants’ ability to reward the lower cost/higher quality credit card networks with greater sales, which in turn reduced the incentive for the credit card networks to be lower cost or higher quality. Thus, we alleged, these vertical agreements between the credit card networks and merchants directly restrained competition among the companies in violation of Section 1 of the Sherman Act.
The government sought straightforward relief: an injunction against the restraints. MasterCard and Visa settled on the day the lawsuit was filed. Amex did not.
On 7 July 2014, after more than three years of fact and expert discovery, trial began before Judge Nicholas Garaufis in the Brooklyn federal court. Judge Garaufis held trial over seven weeks during the summer of 2014, hearing over 30 fact witnesses and four expert witnesses. On 19 February 2015, Judge Garaufis issued his decision finding that the Amex rules did in fact violate antitrust law.
Judge Garaufis found that Amex’s rules “create an environment in which there is nothing to offset credit card networks’ incentives – including Amex’s incentive – to charge merchants inflated prices for their services. This, in turn, results in higher costs to all consumers who purchase goods and services from these merchants.”2 He expressly and repeatedly noted that the credit card industry is complex and that credit card networks, including Amex, “must balance the demands of two sets of customers – merchants and cardholders – in a market that is highly concentrated and distorted by a history of antitrust violations.”3
The Second Circuit overturned Judge Garaufis’s decision on 26 September 2016. It found that “[t]he district court erred here in focusing entirely on the interests of merchants, while discounting the interests of cardholders.”4
This essay does not delve into the debate of the proper analysis of antitrust enforcement, or “two-sided markets” under a rule-of-reason review. I encourage readers to closely study the trial and appellate decisions, and come to your own conclusions. Instead, I will regale you with a more modest tale: the perilous journey through my examination of Amex’s brilliant and world-renowned economic expert. For a short essay in a book about antitrust trials, I see that examination as a microcosm of the government’s case against Amex, the evidence adduced at trial and, possibly, antitrust analysis and trial practice generally.
5 August 2014: 2.15pm
We had just came back from lunch.
During the trial session that morning, Gilbert had completed his direct examination with the legal team from Amex. That testimony bears no further mention, as I asked him no questions specifically about his opinions in favour of Amex. The questions I did ask led to this important lesson: if you tussle with an economist (especially one like Gilbert) about their economic opinions, you, the lawyer, will always leave the loser. I have never seen, heard or read about a top-tier economic expert changing his or her opinion on cross-examination. Going directly at economic experts’ strongly held opinions merely allows them to further explain why they are right and you are wrong.
So what do you do? First, you can try to discredit them. But that is very difficult when you are presented with a prominent and careful economic expert, supported by an army of economic evidence. Suffice it to say, I did not do that. Second, you can nibble at the edges of his or her opinion: “you don’t have an opinion on X” or “you didn’t consider Y.” I did some of that, but it would have made for a pretty dull time for all and would have been a real lost opportunity. Third, you can say: “I’m going to talk about what I want to talk about.” That’s what I did. It worked almost to perfection, with one glaring exception: “the quiet life.” And we can never forget the exceptions.
Start with a bang: obtain a critical concession
The government contended that Amex’s rules were anticompetitive because they thwarted a fundamental principle of competitive markets: that a company can gain more volume by reducing prices. So, my examination of Gilbert started right there:
Q. Would it be competition amongst networks if a network decided to lower its price to a merchant so that it would attract that merchant to steer towards the lower price network?
A. It’s a form of competition . . .
. . .
Q. Is that competition between the networks prohibited by the [Amex rules]?
A. If it’s adverse to American Express, it is, yes.5
Score. For a case that was all about how Amex’s rules block competition, I felt pretty good. I then wanted to weave in a White Paper Gilbert wrote and submitted to the government four years earlier during the government’s investigation:
Q. Would it be fair to say that competing to take market share away from rivals is the essence of the market economy?
A. Yes. I might have even written that myself.6
I then introduced the White Paper, entitled “An Economic Analysis of American Express Contracts Limiting Steering.” I pointed him to two statements on page 15 of the paper:
“If merchants had unfettered freedom to steer customers at the point of sale, it is likely that they would encourage customers to use the card that has the lowest merchant discount fee.
If such steering were pervasive, it might force Amex to charge a lower merchant discount fee . . .”7
Ok, now we’re cooking with oil. For those keeping score at home, in the 10 minutes of my cross-examination, Gilbert had admitted that credit card networks compete by lowering prices to attract more sales; merchants would use steering to promote that competition; that competition would result in lower prices; and the Amex rules block that competition.
This point clearly was not lost on Judge Garaufis. He expressed it loudly and clearly in his opinion: “By preventing merchants from steering additional charge volume to their least expensive network, for example, the [Amex rules] short-circuit the ordinary price-setting mechanism in the network services market by removing the competitive ‘reward’ for networks offering merchants a lower price for acceptance services. The result is an absence of price competition among American Express and its rival networks.”8
Pretty good for a tango with a cobra, huh?
Remind the judge about prior concessions
I could have stopped there and declared victory. And I actually considered it. A few questions, a few concessions, boom, sit down. Instead, I moved into a series of questions under the umbrella of “you didn’t consider . . .”, which were also designed to remind Judge Garaufis about earlier concessions in this long trial. For example, I asked Gilbert about testimony I elicited at trial from a senior Amex executive two weeks earlier:
Q. Are you aware that [senior Amex executive] also testified that one outcome of the non-discrimination provisions being removed could be that American Express cardholders get more rewards?
A. I think I’ve seen the testimony, yes.9
Then I read to Gilbert the trial testimony from that same senior Amex executive:
Q. Now, the reason that American Express didn’t adopt a strategy of trying to lower its price to track more American Express sales was because American Express does not compete on cost versus Visa and MasterCard, and the intent of American Express programmes is not to make American Express cheaper than Visa and
A. I don’t think anybody’s business strategy is to be cheaper than the next guy.
. . .
Q: But since the American Express non-discrimination provisions exist today, American Express does not need to either reduce its discount rate or provide these extra rewards. Correct?
A: Well, we would be fighting to retain the business by any means necessary, right. So, yes, we may need to increase incentives to consumers. We may need to reduce pricing to merchants. All of that is economically negative to American Express. And all of that, we believe, is unnecessary because we believe our non-discrimination provisions are there for a good reason.10
This line of questioning was about two-sided markets. The point of the government’s case was that Amex’s rules interfered with the way the market would work if merchant fees and cardholder rewards were competitively set. Amex did not necessarily help its cardholders – many of whom value the rewards they earn when they use their Amex cards – by impeding competition on merchant fees. If Amex were forced to compete with lower merchant fees charged by the other credit card networks, it might lower its own merchant fees. But it had other options. It could keep its merchant fees the same – or even increase them – and use its higher merchant fees to increase cardholder benefits. Giving cardholders better rewards would make them less susceptible to merchant steering.
I also directly addressed the consequence of Amex’s two-sided market argument: that increased merchant fees mean increased cardholder rewards. The people who pay for higher prices at merchants (all consumers) are not the same as the beneficiaries of those higher prices (Amex cardholders). So, while we all pay prices at the store that accounts for the merchant’s credit card costs, not all of us receive the benefits Amex said it provides based on its higher fees.
Q. In your direct testimony, you talked about how cardholder rewards and cash-back are kind of like a coupon or a discount, economically, when a cardholder uses their rewards cash-back card. Is that fair?
Q. So, economically, when I go to the store and use my rewards card, it’s like I’m paying a lower price for whatever I’m buying at the store. Is that fair?
Q. But that’s only true for the people who have rewards credit cards. Is that correct?
Q. So anyone who does not use a rewards credit card does not get that lower price. Is that fair?
. . .
Q. So if I am a consumer who doesn’t have high enough income or who otherwise can’t qualify for an American Express rewards card, I don’t get the benefit at the retailer of the lower price paid by American Express cardholders. Is that correct?
A. That’s correct . . .11
That concept also found its way into Judge Garaufis’s opinion: “Even if American Express passed through every cent of its premium or the incremental revenue realised from its . . . price increases to cardholders – which it does not – customers who do not carry or qualify for an Amex card are nonetheless subject to higher retail prices at the merchant, but do not receive any of the premium rewards or other benefits conferred by American Express on the cardholder side of its platform.”12 He wrote that in the extreme example, every retail customer not using an Amex “is subsidising, for example, the cost of the premium rewards conferred by American Express on its relatively small, affluent cardholder base in the form of higher retail prices.”13
Risk it all for “the quiet life” (or why to leave economics to economists)
All is well and good for Uncle Sam, right? I thought it was going so well that I then violated my first principle and engaged Gilbert in hand-to-hand economics. I went there not once, but twice.
First, I had Gilbert confirm that the seminal papers by Jean-Charles Rochet and Jean Tirole on platform competition in two-sided markets cited not only work by Professor Michael Katz – the government’s expert in the Amex case – but also Carl Shapiro, the deputy assistant attorney general for economics at the time the Amex complaint was brought.
I established with Gilbert that e-books is a two-sided market, but he did not consider the second side of the market in finding that Apple had market power – a finding adopted by a federal judge and affirmed by the Second Circuit.
Gilbert agreed that computer operating systems are two-sided markets under Rochet and Tirole’s definition. Then I walked Gilbert through each concession. Microsoft was found to have market power in the government’s case against the software company? Yes. Microsoft’s vertical agreements with computer manufacturers were anticompetitive? Yes.
Three-for-three: all singles and not home runs, but still constructive.
But then I raised “the quiet life”.
The silliest thing, in retrospect, is that the point I was making wasn’t even that good. In my defence, I was trying to respond to an argument Gilbert had made in his direct testimony that, if the government’s arguments about how Amex had market power were correct, then Amex would be charging even higher merchant fees than it did – they would be charging the so-called “maximum rational price”. From an economic perspective, there are lots of reasons why this is not right. Amex could have had imperfect information, or just thought its information was imperfect, and so it could not fully maximise its pricing power. Amex could have received other benefits from merchants in addition to price. And
Or, as “the quiet life” describes, even monopolists do things that look like competition – usually innovation, but maybe lower prices too. Hicks wrote in 1935: “The best of all monopoly profits is a quiet life.”14 Gilbert himself had used that principle to say that “competitive market structures are more innovative than are protected monopolies, although econometric evidence on the general linkage between concentration and innovation is inconclusive.”15 Even now, I remain uncertain why I thought it was important enough to make.
Nonetheless, I started down that dangerous road.
Q. And one of the potential factors could be the risk of regulation if a firm raised price to the maximum rational price. Is that fair?
A. I don’t think that was something that was considered in the Microsoft case, and I don’t think that was particularly relevant in my analysis here.16
To be candid, my question was stupid. It made no sense, certainly not to Judge Garaufis or any other normal-thinking person, and it took me right where I didn’t want to go: economic theory. It prolonged a discussion about Microsoft, a case a decade earlier that merited no further mention. While I should have never asked the question, I should have stopped there and turned to more constructive pursuits. But I didn’t.
Q. Another reason is what economists call “the quiet life”. Do you recognize that term, “the quiet life”?
A. Well, it depends upon – I recognize those words. It depends on the context.
Q. So let me attribute them to an author. So, Sir John Hicks in 1935 described the quiet life. Is that correct?
A. I don’t think he described the quiet life. He, again, wrote those words.
Q. “The best of all monopoly profits is the quiet life.” Is that what he said?
A. I believe he said something like that.
Q. The point that Sir John Hicks was making about a monopolist enjoying the quiet life is that it’s more profitable to keep waters calm and not raise the price to the maximum rational level than it is to try to get every last penny out of customers. Is that correct?
A. I mean, I actually interpret the Hicks – Hicks’ comment and the term in the context of innovation where, you know, one issue is whether – does a monopoly innovate more or less than a competitive firm. And I think he was – that quote is generally considered in the context of, well, would a firm with a lot of market power sort of disrupt its own industry by innovating.17
I got lucky. This could have been my Waterloo. As it turns out, it was just a waste of time, but please take this as a caution. It could have undermined all the good work I had done. If you leave this essay with anything, leave it with the warning that no matter how well you know your case, how much you think you know economics and how great a trial lawyer you think you are, there are just some things that must be edited from your examination. For me, it was the quiet life.
It would have been quite a let-down to end with the quiet life. But I had saved my best questions for last. You see, a long time ago, I realised that virtually every examination (direct or cross, fact or expert) has the same structure: start with the second-best points; end with the best ones; and try to keep from pissing off the judge and the jury in between.
To understand the importance of my final questions to Gilbert, we must first discuss what happened before he took the stand. For the entire case-in-chief, trial was going well for the government. It felt like we won or tied every single day. And then we rested. Amex called as its first witness a long-time senior executive. This witness was dynamic, one of the most compelling executives I have ever come across. Not only is he a fantastic businessperson, as it turns out, he is an awesome witness. He literally inspired awe in me.
He took the stand and gave his eloquent testimony. He ended with his belief that Amex will go out of business if it loses its case. (Recall how you end any examination with your best point?) At that moment, I thought we were done. Regardless of what you think about Amex, you have to concede that it is an important part of credit card competition. If it goes out of business, that would be bad for everyone (well, everyone except for MasterCard and Visa, I guess). I recall this moment like it was yesterday. He said, “So I am very, very clear. If the [rules] go away, we will go away.”18 My stomach dropped.
It was absolutely devastating testimony that captured the attention of everyone in the courtroom. Making this worse for the government, this testimony came on a Thursday, and we did not have court on Friday. The next two witnesses were other Amex executives with Amex Direct, who were certainly not going to disagree with another executive’s testimony. The first opportunity to try to recapture the momentum and narrative would be the cross-examination of Gilbert, days later.
On that fateful Thursday, after the devastating Amex testimony, I returned to the office to scour Gilbert’s literature for something in his writing useful enough to bring to court. At first, I found articles written by Gilbert that led me to a line of questions that basically said: “Who cares if Amex goes out of business, if it can’t compete without blocking competition?” But I cared; everyone cared. So, even if Gilbert would admit it, that was not an option.
The best I could find in his writings was an analysis he did of the Microsoft trial, interestingly co-authored with Michael Katz, the government’s expert in the Amex case. In that paper, Gilbert and Katz wrote: “In its defense, Microsoft contended that the company is a vigorous competitor that benefited consumers by supplying high-quality, innovative products. According to Microsoft, antitrust action against it would dampen incentives for competition and slow software innovation.”19 The paper also references Bill Gates’ concern that competition from Netscape threatened to “commoditize the underlying operating system”.20 You can see why I didn’t jump for joy. This paper gave me a potential point – Microsoft lost an antitrust case it thought was existential, and it turned out OK – but it was an unfulfilling one.
I had to solve the puzzle about how to prove Amex would continue to exist and thrive even if the government won. In my quest to figure it out, I began reading the US v Visa decisions from early in the past decade, the district court opinion and the Second Circuit affirmance. Those decisions addressed the government’s challenge to Visa and MasterCard rules that kept banks from issuing Amex cards. Not surprisingly, the US v Visa case had been discussed at length with Judge Garaufis, including at summary judgment and during trial.
I started with the Second Circuit opinion. Nothing.
Then I turned to the lengthy trial decision. There it was, on page 74: a single, wonderful sentence. Although Visa said changes to its rules would be “a fatal blow to its system . . . such was not always its view,”21 the court said. Eureka!
At that point, I was running around in the middle of the night, screaming and interrupting others who were trying in vain to actually prepare for their parts of the trial. But then, as I showed the decision to each and every one of my colleagues, it dawned on me: what is a “fatal blow to the system”? Why did the Visa court cite the testimony? Is this really all that good? So I pulled the trial transcript from US v Visa. And that effort – and good fortune – produced this part of my cross-examination of Gilbert:
Q. Do you recall the testimony . . . that “if the [rules] go away, we go away”?
. . .
Q. Are you aware that the United States sued Visa in 2000?
Q. And, in fact, you cite it in your reports?
Q. And you’re familiar with the trial testimony in that case. Is that correct?
A. Oh, I wouldn’t say I’ve read the trial testimony. I might have read parts of it.
Q. You’re familiar that it’s called US v Visa, or the exclusionary rules case?
Q. And you’re aware that the exclusionary rule that Visa had was called Bylaw 2.10(e)?
A. I know it was a bylaw with a number on it, and it also had a letter. But whether it’s those numbers and that letter, I couldn’t tell you.
Q. And that bylaw basically blocked banks that issued Visa cards from also issuing other network cards, like American Express and Discover. Is that correct?
Q. And are you aware that, just like [the Amex executive] in this case, there was a senior executive at Visa who was Visa’s first defense witness?
A. I’m not aware of that, no.
Q. Are you aware that that witness, his name was Mr Beindorf, testified that if Visa lost its exclusionary rule, it would be a fatal blow to Visa?
A. I’m not aware of that testimony.
Me: Your Honor, may I approach?
The Court: Yes.
Me: Your Honor, I have handed out what’s been marked as PX 2732. It is the transcript from the US v Visa case for the day – the trial day of 1 August 2000.
. . .
Q. Professor Gilbert, let’s turn to page 4,403.
. . .
Let’s read starting . . . on line 19:
“Question: . . . So, we’re shifting here away from advertising for the moment. I think you said – certainly correct me if I’m wrong – that you thought Visa might be destroyed – did you say that – if there were no 2.10(e)?
Answer: Yeah. I think it could be a fatal blow to the Visa association.”
Q. Now, on page 4,404, line 9, the questioning continues:
“Question: And so what you’re saying is that American Express has not harmed Visa outside the United States, but it would not only harm it, but kill it in the United States. Is that what you’re saying?
Answer: What I said was that I believe that if Bylaw 2.10(e) was eliminated, it would lead to the destruction –”
The Court: It could lead.22
(A quick digression: I knew right here that I did it. While I had misread the transcript by accident, Judge Garaufis’s correction showed me that he was following word for word. He knew I knew this was a critical point, and he was looking to see what I had to say in response.)
Q. “– it could lead to the destruction of the Visa association in the US marketplace.” Do you see that?
Q. Now, let’s turn back to what – the previous testimony this is referring to. Page 4,373, please. Now, this was Mr Beindorf’s direct examination. What I just read to you was his cross-examination.
On 4,373, the page number in the upper, left-hand corner, line 4, on direct:
“Question: Do you believe – and I’m addressing your attention now to Bylaw 2.10(e) – do you believe that that rule, that bylaw of Visa, do you believe that Visa needs such rules in order to be able to run an effective business?
Question: Is Bylaw 2.10(e) one such rule?
Answer: Yes. Without Bylaw 2.10(e), I think that the association, as we know it, would disappear.”
Q. Do you see that?
Q. And today, does Visa exist?
Q. No further questions, Your Honor.23
Microphone proverbially dropped.
* * *
So there you have it. A few moments in time from one of the Obama antitrust trials. While I hope you find this essay informative, on macro, micro and in-between levels, this essay is neither complete, nor the end of the story. Throughout our eight weeks in Brooklyn, there were many highs and lows, successes and failures, fights and concessions. While this essay focuses on one small aspect of trial from my perspective, it cannot be lost that dozens of hard-working employees at the DOJ gave almost four years of their lives to prepare for, and win at trial, one of the most difficult and interesting antitrust cases ever. It also cannot be forgotten that the Second Circuit reversed Judge Garaufis’s decision. And the week this book was sent to press, the United States Supreme Court granted 11 state attorneys general leave to appeal the Second Circuit’s ruling.
- This essay is solely on my own behalf and does not attempt to articulate a position of either the Department of Justice or Quinn Emanuel Urquhart & Sullivan.
- U.S. v. American Exp. Co., 88 F. Supp. 3d 143, 150 (E.D.N.Y. 2015).
- U.S. v. American Exp. Co., 838 F.3d 179, 206 (2nd Cir. 2016).
- U.S. v. American Exp. Co., Trial Tr., 5129, 5133, Aug. 5, 2014.
- Id. at 5136.
- Id. at 5138.
- American Exp., 88 F. Supp. 3d at 151.
- U.S. v. American Exp. Co., Trial Tr., 5144, Aug. 5, 2014.
- Id. at 5142-45.
- U.S. v. American Exp. Co., Trial Tr., 5252-54, Aug. 6, 2014.
- American Exp., 88 F. Supp. 3d at 216.
- Id. at 217.
- John R. Hicks, Annual Survey of Economic Theory: The Theory of Monopoly, 3:1 Econometrica, 1, 8 (1935).
- Richard Gilbert and Michael L. Katz, An Economist’s Guide to U.S. v. Microsoft, U.C. Berkeley, Center for Competition Policy, Working Paper No. CPC01-19, at 23 (February 2001).
- U.S. v. American Exp. Co., Trial Tr., 5247, Aug. 6, 2014.
- U.S. v. American Exp. Co., Trial Tr., 5247, Aug. 6, 2014.
- U.S. v. American Exp. Co., Trial Tr., 4634, July 31, 2014.
- Gilbert, An Economist’s Guide to U.S. v. Microsoft, at 3.
- Id. at 6.
- U.S. v. Visa USA, Inc., 163 F. Supp. 2d 322, 401 (S.D.N.Y. 2001).
- U.S. v. American Exp. Co., Trial Tr., 5255-59, Aug. 6, 2014 (emphasis added).
- Id. at 5259-60.