Listen & Download
This teleforum will investigate the potential impact of the pending decision in the FTC's controversial Section 5 lawsuit against Qualcomm, brought days before the change in administration two years ago, with the incoming acting chair writing an unusual and biting dissent. Among other things, the FTC is seeking to permanently enjoin Qualcomm from engaging in certain industry-wide patent licensing practices, which the FTC alleges impair competition in violation of the antitrust laws. However, it has been argued that the FTC’s novel theory fails to meet the burden of proof by showing actual evidence of harm, as clarified recently by the US Supreme Court in Ohio v. American Express Co. The consequences of FTC's legal theory, if upheld by the court, could reach well-beyond patent licensing arrangements. Indeed, some experts fear that changes to Qualcomm's business model will undermine U.S. national security interests and cede American leadership in the 5G race to a foreign adversary—the same concerns echoed last year by the Committee on Foreign Investment in the United States (CFIUS) when it recommended that the President permanently prohibit Broadcom from acquiring Qualcomm.
Olivier Blanchard, Senior Analyst, Futurum Research
Geoffrey A. Manne, President and Founder, International Center for Law & Economics
Teleforum calls are open to all dues paying members of the Federalist Society. To become a member, sign up on our website. As a member, you should receive email announcements of upcoming Teleforum calls which contain the conference call phone number. If you are not receiving those email announcements, please contact us at 202-822-8138.
Operator: Welcome to The Federalist Society's Practice Group Podcast. The following podcast, hosted by The Federalist Society's Telecommunications & Electronic Media Practice Group, was recorded on Tuesday, February 13, 2019, during a live teleforum conference call held exclusively for Federalist Society members.
Dean Reuter: Welcome to the practice group's teleforum conference call as today we get a litigation update on the FTC/Qualcomm case. I'm Dean Reuter, Vice President, General Counsel, and Director of Practice Groups here at The Federalist Society. We're very pleased to welcome two experts to our call today. We're going to get opening remarks from each of about 10 to 15 minutes, and then maybe a little back and forth between the two of them, but then, as always, we'll be looking to the audience for questions, so please have those in mind.
We'll hear first from Geoffrey A. Manne. He's President and Founder of the International Center for Law & Economics. He'll be followed by Olivier Blanchard, who is a Senior Analyst at Futurum Research. With that, Geoff Manne, the floor is yours.
Geoffrey A. Manne: Thanks, Dean. And thanks to The Federalist Society for having us, and thanks to everyone for being here to listen in. I'm going to take a few minutes and talk a little bit about the antitrust issues presented by the FTC v. Qualcomm case. And I'll just admit upfront that I have my own view of where the right position lies, and I'm going present, largely, that view, although I will give the general sense of what I think the FTC's case is about. So what I'm really going to do is sort of present the case very briefly and then talk about why I think it's a bad case. And then, I think Olivier is going to talk about some really interesting insights he has into the nature of the industry, and why Qualcomm does what it does, and what the implications would be if the FTC were to win its case.
The core of the complaint that the FTC brings is, really, essentially, emerges out of complicated negotiations between Apple and Qualcomm. And these are two highly sophisticated, well-funded parties, but it's really the nature of this dispute between them that has blossomed into this larger FTC case. There're two separate and sort of interconnected components to Qualcomm's conduct that its counterparties, like Apple and now the FTC, have attacked it for, and that's the sale of communications chips that go into cell phones, and in particular, smart phones, that it directly manufactures. Alongside, its licensing practices border wireless relevant intellectual property. Qualcomm was, of course, a pioneer in developing CDMA technology and a host of other technologies that make smart phones into smart phones and not just really cool iPods. But it's important to note that Qualcomm also sells chips and increasingly what's called system-on-chips, which means that they're selling the chips that basically make the cell phones function, not just in terms of their wireless connectivity, but in terms of a host of functions that they perform.
The argument is that the FTC says that Qualcomm refuses to license its really popular chips, these chips that perform this wireless functionality and other functionality, if the licensees don't also agree to license Qualcomm's IP, its underlying technology. The FTC argues that Qualcomm has market power in the chip market and that it can leverage that market power to get supracompetitive rates for its technology. The FTC doesn't explicitly call it a tie, but they refer to this "no license, no chip" policy in a way that really makes it sound like what they're complaining about is the tying or bundling of two elements together.
There's a couple of problems with this. First, as then-Acting Chairman Maureen Ohlhausen pointed out when she dissented from the bringing of this case many years ago, the complaint fails to allege that Qualcomm actually charges more than a reasonable royalty and instead relies on allegations, characterizations by OEMs, by device makers, stating that the FTC says many OEMs regard Qualcomm's royalties as non-FRAND. So you would think at trial that they would demonstrate that it's not just the allegations of Qualcomm's counterparties, but that there really is reason to believe that they're leveraging this alleged market power in chips to extract supracompetitive rates for their IP.
And even if they weren't able to demonstrate that prices were high, the FTC said we don't even necessarily have to do that. It would be sufficient to show antitrust injury here if we could show that the conduct that Qualcomm undertakes here has foreclosed competition by other chip manufacturers, has reduced expenditure in R&D, has raised prices of devices even if not the prices of the IP directly, sort of sets of ancillary outcomes that you could point to and say, "Hey, look, this is what happens when you have anticompetitive conduct in a monopolized industry."
So at trial, I think it's pretty important -- I'm actually going to read a little bit of the testimony that came out at trial. So first of all, let me just say—I apologize, I meant to say this before—there was really no focus on exactly what appropriate or allegedly FRAND license rates would look like, and really, no evidence presented at trial that -- consistent with what the FTC alleged before trial, that the prices that Qualcomm was charging were actually supracompetitive. Instead, the argument was, as I said, that they were leveraging this market power and functionally leading to higher rates and lower output, evidence of coercion of the sort that you would expect if you were leveraging your dominance, use of nonstandard licensing practices that might be a proxy for coercion.
Okay, so here's this exchange that I want to read to you between Qualcomm's lawyer and the FTC's economic expert, Carl Shapiro. So the lawyer says:
"You have not quantified the effects of Qualcomm's business practices on any other chip maker in the relevant period, correct?"
"I have not quantified that. That's correct."
"You did nothing whatsoever to analyze whether rivals R&D spending, research and development, went up, or down, or remained the same, right?"
"I'm not offering an opinion about their financial circumstances."
"You haven't done any quantitative analysis to determine how much of the surcharge you're opining about, if any, was caused by chip leverage as opposed to the threat of litigation, the fear of litigation, the fear of injunctions, and other factors, right?"
"I have not quantified the royalty surcharge."
"You have not quantified the effects of Qualcomm's conduct on handset prices?"
"So using Intel as an example, you actually predicted in your rebuttal report that Qualcomm will gain market share at the expense of Intel, right?"
"Again, everything else equal, that's the natural and inevitable economic consequence of the raising rivals' cost. I stand by that."
"But in fact, as you acknowledged two weeks ago here in this courtroom, Qualcomm's market share in what you call the LTE premium market has been declining since 2014, right?"
There's a lot of evidence like this in the trial that, to me, suggests that the FTC has not made out its really complicated case. It would have been a simple case if the FTC had said, "This leveraging theory leads us to believe that prices are too high," but they neglected to show that prices were too high. They pointed to this sort of more convoluted, complicated structural theory to say that the indicia of the abuse of market power here would appear in these other areas in terms of, as I said, sort of the R&D, the device prices, etc., etc. And yet, their own economic expert makes it very clear that he didn't even look into those. The FTC has a theory, and this was a theme that was hit on again and again by Qualcomm's lawyer, the FTC has a theory, but the theory is not enough to prove antitrust injury or to prove any competitive harm in a case like this. And they didn't actually advance any evidence that would actually prove that their theory was accurate.
A couple other things I want to just point to that are really quite baffling about the case -- there is a lot of evidence in the case -- so typically, in a case like this, if you're trying to demonstrate that certain complained of conduct leads to these anticompetitive outcomes, it is, indeed, often hard to find very clear evidence of that because it's hard to know what the "but for" world would look like. So of course, we can always say that prices are decreasing, R&D expenditure is increasing, number of competitors is increasing, and the theory could, indeed, point to all of those being better on their own dimensions if it weren't for this allegedly -- anticompetitive conduct.
The problem in this case is that we actually have a great deal of evidence as to what Qualcomm's practices were and what the effects of those practices were during periods where not only did they not have any market power because their market share was relatively low, but they weren't even manufacturing chips at all. In other words, in the time period in which the entire underpinning of the leveraging theory is impossible because they had nothing to leverage. All they had was IP. And what's very clear from the evidence at trial is that Qualcomm's licensing rates for its intellectual property was absolutely consistent over all of these periods in which both they weren't selling chips at all, and in periods in which even though they were selling chips, they clearly had nothing that one could identify as market power. And yet, exactly the same rates applied even when they did allegedly have market power.
It would be a really strange coincidence if the exact same terms, the exact same practices, and the exact same rates applied in a context in which the company both had and was leveraging market power as when it didn't. The argument would have to be that everything else was changing so substantially that even though royalty rates stayed the same, the quality of the product that licensees were getting was significantly lower or other terms of the agreement were detrimental. And yet, there's really no evidence of that. In fact, you can argue that the evidence points precisely the opposite direction, that over time, Qualcomm's patent portfolio only increased in size in terms of the relevant number of patents that were being licensed, in terms of the quality of the patents, in terms of the quality of the products that were being developed in implementing those patents. And yet, again, the relevant prices stayed the same over the entire term.
The one other point that I want to draw everyone's attention to before I turn it over to Olivier is to note that at issue here are—and I'm sorry, I should have said this at the outset—are what sort of standard essential patents -- that there are, in fact, constraints on the price that Qualcomm is allowed to charge and the negotiating process it's allowed to undertake for licensing the standard essential patents. And as I alluded to before, this is the FRAND concept: free, reasonable, and non-discriminatory terms is the standard -- I'm sorry, fair, reasonable, and non-discriminatory. The interesting thing is that we don't know -- there's a lot of disagreement over what FRAND terms contemplate.
And Carl Shapiro, in fact, the FTC's expert in this case, is one of the proponents of the very most restrictive method for estimating what would constitute FRAND terms. In his estimation, you try to figure out what negotiated terms would look like before a patent became standardized, became an essential patent, and limit the price that the holder of a standard essential patent can charge for that to that pre-standardization rate. As Carl has written, "Royalties that are or would be negotiated ex-ante with full information are a market benchmark reflecting legitimate return to innovation." And he has implied -- not just implied, he stated in this case, and the FTC claims in this case that Qualcomm is actually receiving a return that is greater than its legitimate return to innovation.
The problem is that we actually have that exact benchmark to look at in this case. We don't have to guess what the pre-standard terms of trade would look like. We know them because they're the same terms that Qualcomm offers that the FTC was critical of, that they were offering before they had any kind of market power and in other markets in which they clearly had no market power, again making the FTC's case really a difficult one to understand. It explains why the FTC didn't rest on the allegation that prices were simply too high because that would be almost impossible to demonstrate here. And yet, the more convoluted sort of structural arguments they made don't hold up either.
That's my brief, I think, even under 15-minute criticism of the FTC's case here, which would probably be a new record for me. I want to turn it over to Olivier who's going to, as I said, talk about the broader economic implications of Qualcomm's practices and the FTC's case, and then we'll have a little bit of discussion after that. Olivier, over to you.
Olivier Blanchard: Perfect. Thank you very much for that awesome brief. I almost don't need to be here. You kind of covered, really, most of the really important points about this and did a really amazing job of identifying the key points in the case. And hats off to you for even finding one of the most important pieces of testimony in the entire case. So I'm definitely not dealing with an amateur here. That was pretty fantastic.
So I'm going to try to not repeat any of the brilliant comments that Geoff just made and just share a few observations rather than adding monologue to what's already a pretty clear exposition of the case. I'll try to keep things short, and then Geoff and I will engage -- will answer your questions and then engage, probably, in a more dynamic dialogue about some of the broader implications of this case and also some of the details about this case that might be a little bit nebulous to people who aren't familiar with Qualcomm's business model or the wireless industry in general.
My first observation is that just looking at the burden of proof that the FTC took on in this particular case, and was warned against taking on by Maureen Ohlhausen two years ago, or a little over two years ago when the FTC first filed this strange and ill-advised case against Qualcomm, is that it would be very difficult for the FTC to actually satisfy any burden of proof based on what it was trying to prove. And what we found out in testimony is, which we already suspected and which a lot of us analysts already knew from being familiar with the business model and the industry, is that Qualcomm may seem to be potentially, maybe, somewhere engaging in what could be construed as anticompetitive behavior. Qualcomm actually is not. Its licensing model, through its QTL arm of its business, is actually pretty straightforward.
The case, I think, ended up demonstrating that Qualcomm did not, for instance, increase royalty rates after obtaining, quote/unquote, "market power in CDMA". That did not happen. Qualcomm also did not stop providing chips to Apple or other OEMs during negotiations. That did not happen. There is no such thing as a premium LTE chip market or category. What happens is when a company like Qualcomm, or any leader in any space, develops the next iteration of a technology faster than its competition, that's not a new tier. That's just the latest model. The fact that it might be faster, better, cheaper, more powerful does not make it a premium chip. It doesn't create a premium category. It's just the company got to that particular level of performance ahead of its competition. So even that aspect of the case turned out to be incorrect.
Another aspect, or another argument that the FTC tried to present that completely backfired, I think, is that one of the arguments was that Qualcomm had, to a large extent, a monopoly over the chip market, and that, somehow, it precluded some of Qualcomm's competitors from being able to enter that market or be competitive in that market. And in a time period, we actually saw that Intel was able to completely take over for Qualcomm in Apple's business, and so if Qualcomm was blocking companies like Intel for being competitive in the mobile chip market, then how is it that Apple was able to very quickly replace 100 percent of its chip sets from Qualcomm to Intel chip sets? So even that just didn't bear out in the real world. It was just pure theory, and absolutely the opposite happened in the real world.
On the issue of "no license, no chip," I would point out that "no license, no chip" is not the same thing as "no chip, no license." And it's easy to get tangled in this if you don't necessarily understand the relationship between licensing and chip, or between software and hardware. If Qualcomm engaged in a "no chip, no license" scheme where, essentially, they required OEMs to buy their chips in order to have access to their licenses, that would absolutely be anticompetitive. That is not what's happening, however. What Qualcomm does with "no license, no chip" is it -- by requiring OEMs to purchase their chips to obtain licenses, they essentially unlock the functionality of the chips by allowing them access to all of their features as opposed to just selling them chips without the ability to activate those features, which would render those chips completely useless. So there's no anticompetitive behavior either. It's a practical matter that no one has ever had any problem with, until now, apparently, for some reason. All of Qualcomm's rates, especially with S-E-Ps, with SEPS, are market based and non-discriminatory. Evidence in the trial showed that.
And then I'm also going to bring up the fact that two of the FTC's lead expert witnesses, their star expert witnesses, were kind of destroyed during the case in testimony and counterarguments. One of them, Carl Shapiro, which Geoff mentioned in his segment and quoted, is a smart guy. He has some very interesting theories, and I don't want to impugn him in any way, but the reality of his theories is that they are just that, theories. And not only was it demonstrated in cross that Mr. Shapiro didn't actually apply his theories to this particular case, he didn't test his theories when he could have empirically with the facts of this case, what actually happened in the real world ran counter to what his theories suggested would happen.
I'm going to quote also from the record, to copy Geoff real quick. Edward Snider, who is a Professor of Economics and Management at the Yale School of Management, Professor Snider was asked, "How do you characterize the differences between the approach that you took and the approach that Professor Shapiro took?" And his answer was, "The approaches are very different. My approach is to conduct an empirical analysis informed by industrial organization principles, informed by understanding of what factors influence firm level success or failure in these kinds of industries and in this industry in particular, and also grounded in social science principles where the objective is to determine causality and to establish causality." Professor Shapiro's approach, quote, "is purely theoretical." So that fact that one of the FTC star witnesses could not actually prove that his theories were applicable in the real world, and the fact that what actually happened in the real world ran contrary to what he suggested, was, I think, damaging and definitely did not help the FTC's case. It didn't meet any kind of burden of proof. On the contrary, I think it was self-contradictory, and it worked against him.
And then the second expert witness, which again, I don't want to impugn in any way, something problematic and that leads us to a very different discussion about 5G, China, and role that Huawei might be playing in all of this is Michael Lasinski. And Michael Lasinski and his firm, 284 Partners, as it turns out, have done a lot of work for Huawei in the past. And let me see, let me consult my notes here so I get this right -- but some of the work that they've done includes commercial patent negotiations with/against Qualcomm in the past, and he has served as an expert witness for Huawei in other cases before. So here we have an expert witness who may be looking at a conflict of interest in this case, and here's why. And by the way, Huawei was also itself a witness for the FTC, which I found a little bit interesting.
Here's why all this matters, and why from a 5G industry standpoint, wireless industry standpoint, this really matters. There are currently in the world only two companies that are leaders in 5G development in the standards developments, in R&D, basically in all the 5G IP that makes the new technologies and future generations of digital technologies actually work. And those two companies are Qualcomm in the United States and Huawei in China. There's strong evidence to suggest that Huawei is a corporate arm of the Chinese government. And I think everyone here probably understands on some level the importance of digital technologies and the importance that 5G is going to play in the next 10 to 20 years of digital infrastructure, development, national defense, the Internet of Things, essentially, our infrastructure is moving into an IOT phase where everything is going to be connected, digitized, and controlled by microchips and software.
And so here we are with these two superpowers, the United States and China, and these two companies, Qualcomm in the U.S. and Huawei in China, who are the absolute leaders in this field. And there's no close third. Intel, Apple, Google, none of these other companies even, wireless companies like Ericsson, are not anywhere close to leading the research and development that goes into establishing 5G standards and developing 5G technologies. And those 5G technologies are vital to our infrastructure, to our national defense, to our communications, to our economy.
This is what happens if somehow Judge Koh, or any court of law, or any regulator finds a way to weaken Qualcomm by limiting its ability to derive revenue from its patents and from its research and development. If Qualcomm's revenue is severely limited by a judgement in favor of the FTC, a judgement against Qualcomm, the amount of revenue that will be able to be applied to R&D in the 5G space will be severely limited. The main beneficiary of that will be Huawei in China, and we could look at a nightmare scenario in which, essentially, Qualcomm is forced to relinquish its leadership position in 5G research and where Huawei and China become the dominant force in 5G, and 6G, and 7G development. And the risk there is that the United States and western countries may lose their ability to protect themselves from Chinese technology and how Chinese -- or Chinese developed technology could be used against us.
So I think that there's a bigger picture here besides just the antitrust case, which again, I think the FTC failed to prove, that we need to be aware of. And it's not limited to this case. The attacks on U.S. technology companies, and particularly Qualcomm because it's so important to the development of 5G, are pretty global and we're seeing them everywhere, from antitrust cases in Korea, Taiwan, China, to this particular case here in the United States. There seems to be a concerted effort by China, by Chinese agents, and by companies like Huawei to weaken Qualcomm and its ability to fund 5G research.
And the way that could be done with a case like this is currently, the way that a lot of Qualcomm's revenue or SEP licensing revenue is derived is through component level licensing. And so component level licensing means that when Qualcomm makes its portfolio of IP and patents available to OEMs, it does so as kind of wholesale. Essentially, it says here's one license, one access to all of our SEPs, and we're going to base the price of this license on a certain percentage of the device price. So if a smart phone, for instance, is going to cost $300, and the SEP license costs, or the entire portfolio costs 5 percent of that, what the OEM will pay is 5 percent of $300, which is the price of the device. And so depending on the device, the price per license might be $10, $15, $20 per device.
And it should be noted that Qualcomm actually has a cap, or at least at the time that it was framed by this case, the cap was at $500. So if a phone, for instance, was priced at $1000, the 5 percent licensing fee would still be based on $500 cap, not the $1000 price. So it's not a huge amount of money here for these licenses that -- there's thousands of them. They're not just licenses that touch on wireless connectivity, they're device-wide features that include power management and touchscreen features and all kinds of usability, so it's not just limited to the chip. It's really stuff that spreads to every part of the device and that makes this patent portfolio very flexible and agile for OEMs to use.
One of the objectives that Apple in some of its lawsuits against Qualcomm and that kind of found it's way into the FTC's case is the notion that because prices are too high right now, as Qualcomm has them, that one remedy would be to switch -- or to require Qualcomm to switch its model from device level licensing to component level licensing and essentially predicate that percentage not on the price of the device but on the price of the chip, which is on the one hand kind of counterintuitive since a lot of the SEPs actually are not limited to the chip. They go to the entire device. But also what is does is the chip is a lot cheaper than the device. So say that a chip is $20 or $30. Now, the 5 percent licensing fee would be predicated on a $30 component as opposed to the $300, $400, $500 device. And this would severely limit Qualcomm, or a company like Qualcomm's ability to derive revenue from very expensive investments in research and development that are essential to 5G and other critical technologies.
And so what we see here is a deliberate mechanism, I think, in my opinion, from Huawei, from China, from other actors, to try to weaken Qualcomm's revenue generating capacity in order to sideline it and give an advantage to Huawei, which doesn't have the same types of problems, can be propped up by the Chinese government, and can definitely take over when Qualcomm is forced to retreat from its very proactive role in developing 5G standards.
So with that, I think it's probably a good time to start our discussion and answer questions, as I am sure you want to probe into this topic a little bit more.
Dean Reuter: Very good. Let me go back to Geoff Manne and see if you want to augment your statement or if you have any questions that you might want to ask, I'll give you the first shot if you might want to ask of our other guest, Olivier Blanchard.
Geoffrey A. Manne: Thanks, Dean. There's actually one aspect of this that I want to draw out a little bit that I think will help in sort of understanding the underlying dynamic here, and actually, by the way, will also set up a further discussion about this really interesting Huawei issue. So I just want to go back, Olivier, to your great line that "no license, no chip" isn't the same thing as "no chip, no license." And I think this is consistent with what you were actually just talking about, but I just want to point out that my take on this has been, in addition to the sort of evidentiary and proof problems we've been discussing with respect to making an antitrust claim on this, that understanding why Qualcomm would adopt a "no license, no chip" model doesn't require recourse to any kind of antitrust analysis. I think I called it the "no duh" problem in the sense that of course Qualcomm conditions the purchase of its chips on the licensing of its intellectual property. How could it be any other way?
In the alternative, if, for example, Qualcomm were forced to license its technology at the component level, and in particular were forbidden from conditioning the access to its chips on the taking of a license, why would anyone ever license for anything more than an absolute pittance from Qualcomm? They would just say, "Listen, we're not going to license at anything like the terms you agreed to. By the way, you still have to sell us your chips." That's what the FTC in the court said. "Go ahead and sue us when we implement without a license. That's the way this is supposed to work." In order to sort of sidestep that, I argued that Qualcomm adopts this practice that says, "Look, if you're going to license our chips, you have to license our IP. That's just the way it works." And it's my understanding that that's perfectly consistent with how everyone in the industry works, whether they have alleged market power or not.
So I want to get your sense of whether that's accurate as a sort of description of the business logic behind what Qualcomm does, and also just your comment on what the sort of general industry practice is, which I think is helpful in understanding whether there's any reason to think that Qualcomm was doing anything illegal here.
Olivier Blanchard: Right. You know, one aspect of this, too, is also you can license IP from Qualcomm without buying chips. The access to IP is purely if you need the IP, it's there to license, and especially if it's a SEP, obviously, FRAND terms apply. It never occurred to me that anyone would even have a problem with this. In a way, it's a little bit like when consumers buy products or opt into a service, whether it's free or it's paid, and before they can use the product or the service, they click that little box, that little "I agree" box, that says they agreed to the terms and services, or a TOS agreement. Essentially, this is kind of the same way. To me, going ahead and getting a license to be able to use the chips that you're paying good money for is just par for the course. There is IP, there is -- it's also an agreement that sets the legal parameters for you to be able to use that chip.
And there's one aspect of this that people easily forget. The licensing agreements also indemnifies the chip purchaser, the OEM, and in case anybody ever wants to -- or even in case Qualcomm wants to sue them for using their chip at some later date in some imaginary universe in which they would want to do that. So the licensing doesn't just -- is not just a revenue mechanism for Qualcomm. It isn't just a mechanism for Qualcomm to open up its portfolio. It's also a mechanism that protects the licensees from lawsuits and other types of legal problems. So the notion that anyone would want to buy a chip without a license is kind of a headscratcher because unless you're going to use those chips for an art project and just glue them to your wall as a decoration, I just -- and that's a legitimate use for the chip where you wouldn't need a license. But if you're actually going to use the chip in a device, then you have to have a license for it for a plethora of reasons.
Geoffrey A. Manne: I think it's also worth highlighting something you said which is really important here that Qualcomm offers the license independently of the chip. And that particularly goes in my mind to the sort of ersatz tying theory that the FTC is trying to bring here because it's a really ineffective tie because anybody who doesn't want to purchase the bundle of Qualcomm chip plus Qualcomm license can buy the relevant chips that they want from someone else and still license Qualcomm's IP, and as I understand it, on precisely the same terms; that there's no difference in the terms whether you buy the chip from Qualcomm or not.
And so to some extent the FTC's entire case turns on the idea that nobody can function without Qualcomm's chips, that this is a false choice – the idea that they could buy the chips from someone else and also license Qualcomm's technology. And they point to the fact that most OEMs are licensing Qualcomm's chips as evidence for this. What they don't do, I don't think, and none of this is a big dispute at issue at trial, but I don't think they really do a great job of distinguishing between some sort of unjustifiable market power that leads everyone to purchase Qualcomm's chips. And the fact that Qualcomm's chips are just better, which of course is not a basis on which you can rest and draw [inaudible 35:03] stability.
Olivier Blanchard: Right. And we can --
Geoffrey A. Manne: And the fact that everyone buys their chips from Qualcomm is not the same thing as saying everyone has to buy their chips from Qualcomm.
Olivier Blanchard: Exactly. And it's honestly what I see here is actually a really healthy market conditions. So you have Qualcomm, which on occasion, comes out with superior chips, right? So let's say they're kind of like the Apple of the chip world. They come out with the best chips ahead of everybody else. Obviously, if you want the best possible chip you have the choice to buy that chip, especially for a premium market phone or a phone that has very high-end features or you want to be very high performance. You're not obligated to buy those chips. You can buy less or lower performing chips from Qualcomm or from another chip maker. You can buy any chips you want and build any kind of device that you want. The trick is that you do need some SEPs for your phone to actually work, and some of those patents are owned by Qualcomm. And so Qualcomm makes those available to you. But, again, the pricing is not anticompetitive. Qualcomm has shown no actual behavior in the real world of using its market position to push the price or to force anybody to pay exorbitant prices for these licenses.
In my opinion, I think they should actually charge more for what you get. And the fact that the portfolio licensing model has been so effective, or the reason why it's been so effective, is that it's practical. You have access to all of the patents that you need in one transaction, and you can use them any way you want based on the agreement, as opposed to constantly going back and forth and saying, "Well, I need this patent but not that one. Hey, can we get a license for this one? Can we get a license for that one?" Six month later, "You know, we were wrong. We need a license for this other patent over here." That's untenable for devices that uses thousands of patents simultaneously that have an immense amount of inoperability challenges to overcome. The fact that a company like Qualcomm that spends so much time solving complicated engineering problems is just -- it's common sense to go to a model -- or to go from a model that's simple and affordable today to an unbelievably complex piecemeal and unmanageable model to build devices just because the FTC has gotten in its head that this model might be anticompetitive or might become anticompetitive in the future, or because Apple wants to pay less than it pays now for IP that everybody else is happy to pay for is -- it makes no sense.
Actually, I think that if the FTC were to prevail with its remedies and to impose its theories of how it believes that the business should work as opposed to how it actually works, that is what actually would cause harm to consumers, to technology companies, to OEMs, and what would cause harm ultimately to the market. And so I find this particular case, independently of anything else the FTC does, I find this particular case ill advised, ill articulated, and just completely nonsensical. It makes no sense on any level whatsoever – not on a political level, not on an economic level, not on an industry level, not on a technological level. I just don’t understand how it even got this far, honestly.
Geoffrey A. Manne: Tell us how you really feel, though.
Olivier Blanchard: [Laughter] Right? Again, nothing against the FTC. I think that they have an immensely important role to play. I just find it unfortunate that in this particular instance they're wasting talent and resources on a fool's errand that could ultimately not only harm a really important industry, which affects the U.S. economy and technology development as a whole, but also could ultimately jeopardize national security. And not just for the United States but for Canada, for Western Europe, for all of the U.S. allies.
Dean Reuter: I've got a couple questions online. So we'll get to those questions momentarily. One question I wanted to ask myself. It's a little confusing to me. I know this case involves Apple. We've talked a lot about Huawei, but they—Huawei, that is—would seem to be a clear winner if the FTC prevails in this case. But what really does Apple have to gain or lose one way or the other?
Olivier Blanchard: I think I can take this one, or at least start, and then, Geoff, you can add your comments to that. Apple wants to pay less for everything that it buys. So everything in this supply chain costs X, and if Apple can impress upon the court that the licensing rates that Qualcomm asks for are too high because they are predicated on a device level licensing model as opposed to a component level licensing model, then Apple can trick the court into potentially forcing Qualcomm to switch to a component level licensing model, which would be a lot more cost effective for Apple.
So, essentially, say that Apple pays, whatever, $7 or $10 per phone for Qualcomm's licensing portfolio, switching to a component level licensing model could drop that amount from $7 or $10 to, say, $2. And so that cost savings per phone per device is very advantageous to Apple. So independently of anything else, I think what triggered a lot of the dispute between Qualcomm and Apple is that Apple just wanted to pay less than it pays now for IP licensing. That's it. It's purely a cost-reduction initiative on Apple's part.
Geoffrey A. Manne: I want to just add one thing to that, really in the form of a question for Olivier because he's better at this than I am. There's a kind of weird irony to this, which is that if successful and if it's correct that a consequence of reduced revenue for Qualcomm is, as it has said and seems sort of self-evident, less R&D, less ability to participate with Huawei in particular in 5G and beyond, it seems like it is potentially a real mistake for Apple because they're always going to need to license this technology and probably also buy the actual component. And they may be setting up a world in which they're dependent on a real, true, monopoly provider with some government protection on top of that. Which in balance in the not too distant future, could be really a lot worse for Apple. So my question to Olivier is is that dynamic accurate? Am I saying that accurately? And if so, how do you explain, or what do you think is going on in Apple's thinking? Are they just thinking too short term, or am I wrong in that there's a real risk here?
Olivier Blanchard: No, no, no. You're absolutely right. I think that your observation of short-term thinking versus long-term thinking is spot on. And I'm not going to try to guess what's going on with Apple and how they're thinking about it. But I think the initial step was, "Let's cut cost out of the model," which is something that Apple is very good at. Tim Cook has always been really good at streamlining and kind of squeezing his supply chain. So this is par for the course for Apple, to try to -- and in a way I understand. If I were Apple, or any other device maker, and I saw an opportunity to pay $3 for something that I'm paying $10 for now, I would probably pursue that opportunity and see if I can make it work. So I think that was the inception, and it's something that once that dog got a bite, it just didn't let go. And Apple's just not letting go of this dream. It's still pursing that dream. But now all of these other things have attached themselves to Apple's battle against Qualcomm and have kind of snowballed into this mix of Huawei, and China, and 5G, and all of these other elements.
But there's also something else that you need to realize about Qualcomm, chips, and Apple. Qualcomm is, essentially, the IP engine for not just the wireless industry, but for the Android space. And now that Apple has kind of moved on from Qualcomm and has put a lot of its eggs in the Intel basket, what you kind of have -- you have a little bit of a fork in the road where Qualcomm has become much more of the Android power base and then Intel and Apple are over on their own corner. And so by weakening Qualcomm, Apple is also able to weaken innovation, or slow down innovation in the Android space, which gives it an advantage since, in my opinion in the last few years, Apple has kind of slowed down when it comes to innovation and has been overtaken by the Android space in a lot of ways – faster modems, better cameras, a lot of features in Android that are much better than Apple's. So I think it's an interesting and it's a smart strategy on Apple's part to weaken Android's ability to compete or to develop innovative new features faster than Apple can.
In the long term, you're correct. If everything goes Apple's way and nothing changes, or they don't deviate from the course that they're on currently, what you may end up with is a world where you have, essentially in the wireless industry, two giants. You have Huawei and Apple. Huawei as a handset maker is a competitor of Apple, but Apple is not a competitor of Huawei in the 5G and wireless infrastructure space. Huawei is a huge maker of antennas and base stations and cell stations. They build and implement infrastructure products that Apple and Qualcomm don't. There're aspects of this where Apple might have done a calculation whereby they're okay with competing against Huawei on the handset level, but they don’t really care if Huawei dominates the world in cell tower production. And they may depend on the U.S. government to at some point say, "Okay, we're not going to accept Huawei's equipment in the United States, Canada, Western Europe, wherever." And that's a different battle, and Apple can stay completely out of it, and only focus its energies on competing against Huawei for consumers' dollars. So it's not just as simple as -- it's not like this linear or this binary Apple versus Qualcomm, or Apple versus Huawei. The only binary equation here is Qualcomm versus Huawei. Apple has its own little fiefdom, its own little schemes that don’t necessarily play as explicitly as Huawei versus Qualcomm and U.S. verses China.
Dean Reuter: We've got a couple callers who've been waiting patiently. Let's turn to them now.
Olivier Blanchard: I'm sorry.
Dean Reuter: No, this is a great discussion. Let's turn now to our first caller if we could.
John Shu: Hi, this is John Shu in California. Two quick questions. First, do you really think that Judge Koh, Lucy Koh, will take into consideration the larger aspects of other parts of the U.S. government prosecuting Huawei, for instance, when the FTC representing the U.S. basically has not done so? I mean, Judge Koh has never issued a ruling that was unfriendly to Apple in her career.
And then the second question is what are Qualcomm's options if it loses at trial, which it seems kind of likely at this point?
Geoffrey A. Manne: Let me answer the first one --
Olivier Blanchard: Yea, go for it Geoff.
Geoffrey A. Manne: -- in two parts, and I think the second one is definitely best for you, Olivier. I don't think that Judge Koh will take account of the Qualcomm versus Huawei, U.S. versus China concerns in her opinion. Nor do I think that antitrust law, as it's currently constituted in the U.S., would give her any basis for doing so. Olivier and I talked about this a little bit before this call, and I think Olivier will make some interesting points in a second about why that might be a problem. I'm not so sure whether it's a problem or not in the narrow sense -- sorry, in the broader sense. I think it could, perhaps, set a bad precedent. I think it's really interesting, though, to think about these broader issues and to be having a conversation about how, and whether an antitrust decision like this that may actually have such broad implications, how those broad implications can actually, in a sensible way that comports with the rule of law, be incorporated.
John Shu: Well, except for the credibility of the witness because Huawei was a witness.
Geoffrey A. Manne: Oh, that's true. That's an excellent point. She could take that into account. There are plenty of other witnesses too, of course. But that would be a hook, you're right, for bringing it into her decision. So sorry, I take it back. You make an excellent point there.
I think as a more general matter, however, she would be hard pressed to find a way to say, "This is an antitrust violation because it creates these national security concerns." I just don’t think that's part of the law as it stands.
Olivier Blanchard: I would actually agree with that. Yeah, I think in a perfect world and in a world in which this is actually an antitrust case and can remain just within -- very narrowly, defined as an antitrust case, then there's no reason why Judge Koh would ever take this into consideration. It's good background noise. It can inform some of her opinion, but I don't think it should really guide it.
The thing is I don't think that this is a cut-and-dry antitrust case. I think that there's more to it than that. And so Judge Koh should at least be aware of some of the circumstances under which the case came to be and how it developed into what it is today.
Secondly, I think that as an antitrust case alone, if Judge Koh just ignores all this stuff and just looks at the case on its merits, I wouldn't be normally particularly concerned. I think that the FTC completely bundled this and failed to make its case. The only concern that I have is Judge Koh, as you said, tends to be very pro-Apple, tends to rule in favor of Apple. I don't think it's entirely an accident that this case landed on her desk. And so you almost feel an invisible thumb, whether it's Apple's amazing amount of power, politically, or even Huawei's very clever shenanigans in the background, kind of weighing down the scales in this particular instance.
My preference would have been for the FTC, not Judge Koh, to consider the broader ramifications of this and to listen to the Justice Department and to listen to, or pay attention, to some of the opinions run by CFIUS and other parts of the U.S. government that warn against things like this. And that should've prepared the FTC to ask itself, "Are we being accidentally weaponized, or are we allowing ourselves to be weaponized by foreign adversaries by weakening -- or going after U.S. businesses of critical importance?" So I think to answer your question I don't think that burden should be on Judge Koh. I think that burden should've been on the FTC.
John Shu: So, Olivier, would you please address the second question, which is if you would please handicap Qualcomm's chances at appeal if they lose at trial?
Olivier Blanchard: You know, I've talked to people at Qualcomm about this and they're concerned, obviously. I'm less concerned than they are, or that they seem to be. I think that Judge Koh will do the right thing. I don't see how she could rule in the FTC's favor in this case. And I might be naïve, but it's just, on its face, there's no way the FTC can win this.
If, for whatever reason, Judge Koh does rule in the FTC's favor and by default in Qualcomm's favor, I think that Qualcomm has a very strong chance of winning on appeal. I really don't think -- there's one Judge Koh. I don't think that they're going to get unlucky twice, and I don't think an appellate judge or a commission would look at the FTC's case and say, "Yes, this is strong. We believe the FTC in this particular instance." They completely failed to prove every single one of their assertions. It's a little bit of a head-scratcher for me that we're even here at this point. It should be a no-brainer.
Dean Reuter: We've still got two questions pending. I know both of you said -- both of our guests said they could go beyond the hour if need be, so let's try and get at least one more question in and we'll see how quickly we can do this. Go right ahead, caller.
Aurora Peterson (sp): Hi. Aurora Peterson in Washington D.C. A question, first, for Geoff. Geoff, can you identify a theoretical model for calculating FRAND rates that's decent and thus inhospitable to successful innovation than Carl Shapiro's model? And then a quick question for Olivier. Olivier, to what extent in this industry are asset-specific investments involved, for example, when Apple switched to Intel as a supplier? What sort of specific investments did Apple have to make, and what sort of investments did Intel have to make to supply chips to Apple? Looking to that question, of course, to examine any supposed potential for, quote/unquote, "hold up". Thanks.
Dean Reuter: Geoff Manne.
Geoffrey A. Manne: Yeah, sure. To the first question, I think that there're a number of economists who've developed alternatives to the kind of restrictive, pre-standardization hypothetical negotiation model – Anne Layne-Farrar and Jorge Padilla, among others. For purposes of this case, I guess I go back to what I said before, which is that in a way, it doesn't matter. We could even take Carl Shapiro's more restrictive approach, and I think there's evidence in this case to suggest that we kind of have a good sense for what the pre-standardization negotiation might look this -- or I should say in this case, perhaps it's the pre-monopoly leveraging negotiation might look like because we actually know what those negotiations look like in the time period in which there's no argument that could be made that Qualcomm had any kind of monopoly leveraging here. So at least in this case, I don't mind using Carl's approach because I think even using that very restrictive approach, there's no basis for arguing that the terms here are anything but FRAND.
And I should add before I turn it over to Olivier, I think the same goes for the more structural arguments. The FTC -- it’s a complicated set of claims and they kind of mash together, for example, as we kind of discussed a little bit, that this sort of inherent sense that device level licensing is a part of this convoluted Qualcomm alleged scheme to extract monopoly rents. Or they referred to the portfolio licensing as part of this as well. And as Olivier pointed out and certainly it's been my understanding, and I think actually FTC testified at trial to this, too, that's the norm for everyone in the industry whether they have anything approaching market power or not, whether they are vertically integrated in the way that Qualcomm is or not. So we have kind of a lot of evidence here about what negotiation would look like and what rates might look like in a sort of analogous but-for worlds, and they look exactly the same as what we have in the allegedly anticompetitive world. It makes me feel like, "Hey, let's use Carl's own approach here," and even under that approach, I think everything looks copacetic.
Olivier Blanchard: To answer the second question as to the amount of investment that goes into development of these chips or chips specifically designed for a particular device, this actually came up in trial. And I don’t have the exact figure in my head, but I think what happens -- what you have to realize is when a chip maker like Qualcomm or Intel designs a chip, and essentially, an SOC or an entire system for a new device, so say the iPhone 11, or the iPhone 12, or Samsung Note 13 when it arrives, the chip maker is going to, essentially, build an entire system and a lot of code onto the chip and onto the SOCs to be able to work with a device.
And what I think we found out in trial with the testimony about Apple claiming that Qualcomm had refused to sell it chips for some time and had demanded certain minimum requirements of purchase numbers is that it probably cost roughly around a billion dollars per new device, or per new device family, for a chip maker to actually design a chip and the technologies around the chip to be integrated into that phone. And that's money that, essentially, comes out of the pocket of the chip maker. They make an investment in the development of those chips and that technology for that vendor, for that phone maker. And so the business model there is that the expectation is that they will be able to sell enough chips and enough licensing to make up for that investment and hopefully make a profit out of it based on their own projections.
And so to answer your question, I think that's the best way to look at it. Just think of it as basically a billion dollar investment from a chip maker for every phone, or phone generation, that they're going to put their IP into.
Geoffrey A. Manne: I just want to -- I think this came out at trial, too, and maybe you remember, Olivier, is there any reason to think that that billion dollar price is any different for Qualcomm versus Intel, or Media Tech, or Samsung, or anyone else? At least at minimum, I'm certain that licensing is not a part of that because there is no licensing rate at the component level.
Olivier Blanchard: Right. I think it's -- yeah, if you think of economies of scale, I think the billion might actually be more -- and I could be wrong about this. I'm just kind of -- this is tugging at my memory strings, but I don't have the full picture in my head right this moment, but I think where it's particularly relevant is with Apple. Apple has its own little ecosystem, its own stack. And so when you're designing specifically for Apple, you're designing for Apple's stack and nothing else. Whereas when you're designing for the Android space, you're designing for everybody. Everybody, essentially, uses the same technologies in the same model. For Qualcomm to be designing for Samsung, and Motorola, and everybody else, it's kind of like the investment is roughly the same, or it's mitigated by the fact that everybody is using this same language, whereas --
Geoffrey A. Manne: The return is less risky, though.
Olivier Blanchard: Right, right. So by agreeing to design for Apple, you're taking on an additional billion dollar investment that you wouldn't normally take on if you were only dealing with Android handset makers. I think it might be oversimplification --
Geoffrey A. Manne: I think a billion dollars was the right number because I think that was the number that Qualcomm and Apple agreed to for Qualcomm to design chips specifically for Apple when they switched from Infineon. I think it was a billion dollars.
Olivier Blanchard: I think that's the case. That's why that number keeps coming up to me. Again, it might be an oversimplification, but it's a good way to kind of generally understand how it works.
Dean Reuter: Let's check in with our final caller of the day. We're in danger of setting a record for the length of a teleforum call. Go ahead, caller.
Caller 3: Hi, thank you so much. I'll be brief. Could both panelists please touch upon the impact of the remedy of the FTC's requesting on the overall mobile industry?
Geoffrey A. Manne: I think that was implicit in one of the earlier questions, and I was going to try to push Olivier to answer that a little bit more. And rather than take up time with me giving my relatively uninformed opinion about this, I actually really want to hear what Olivier has to say on this. This is really his specialty.
Olivier Blanchard: Yeah, specifically, what remedy are you referring to?
Geoffrey A. Manne: I think -- sorry, just --
Olivier Blanchard: Component level licensing?
Geoffrey A. Manne: -- component level licensing or at minimum, putting a bar to vertical integration for Qualcomm. So at minimum it's, essentially, saying you can't both sell -- in my mind, it's effectively saying you can't both sell chips and license technology. But I think the broader sense it becomes this question, as you raised before, that they're really sort pointing to requiring component level licensing.
Olivier Blanchard: Honestly, I'm not that concerned about the impact that it would have on the smart phone industry or the cell phone industry. We already have really good cell phones. People are not buying them at the rates that handset makers would like them to, which is they'd like them to buy a new phone every year. We're moving away from two-year replacements to two-year or three-year periods between upgrades. So we could actually as consumers withstand a negative judgment against Qualcomm in this case. And we wouldn't necessarily feel the pinch of a slowdown in innovation with handsets and wireless technology for years to come. We might not even feel it until 2023, and then we might start getting impatient and wonder why smart phones aren't getting any better as the same rate as they were before.
My concern is actually more with regard to 5G leadership. The fact that the entire digital and technological fabric of our country's infrastructure, of our economy, of our markets, of everything that we do is going to be jeopardized, potentially, by a negative judgment against Qualcomm. Phone makers will still make money. We will still be able to buy phones. Whether we end up spending more time on 4G networks than 5G networks or it slows down the proliferation and deployments of 5G services and really 16 5G, which is going to be much more mesh than -- and interconnectivity base than about faster download times.
The consumer concerns or the consumer feature concerns are very, very secondary to the infrastructure and technology security concerns that I have about a negative judgment against Qualcomm.
Dean Reuter: And this is Dean Reuter. Let me just tie that back into the question we heard at the very outset from I think it was John Shu about the consequences of a decision here. This is for Olivier Blanchard, I suppose. You seem pretty confident that Judge Koh can't find in favor of the FTC, but let's assume for the sake of argument she does, it feels to me like the damage might be done at that point, even if Qualcomm gets another bite at the apple, so to speak, in the Ninth Circuit and maybe even the Supreme Court, that once -- if Apple prevails or if the FTC prevails, then all the other licensees of Qualcomm might stop paying or require Qualcomm to take them to court in order to extract payments, and then the whole business model sort of comes apart. Is there a danger of that?
Olivier Blanchard: Yeah, of course there's a danger of that. As a matter of fact, Apple and Apple's contract manufacturers have refused to pay Qualcomm for, what, the last year and a half, two years? So to a certain extent, this has already been happening, and Qualcomm has been able to weather that storm fairly well. My concern is that a negative judgment would send investors running. It would create a lot of uncertainty in the market for the future of Qualcomm, even if it turns out that Qualcomm is a good case and that whatever changes would be required by a negative judgment against Qualcomm would not be immediate.
And, in fact, 0they couldn't be. There's some practical reasons why they couldn't be. All these licensees have contracts in play right now that would have to be renegotiated. That wouldn't be interrupted the day of the judgement, right? So there would have to be a time period for that adjustment to take place. And I expect, in the hypothetical nightmare scenario in which this would happen, that grace period, that transition period, during which the details would have to be worked out, all of these contracts would have to be reworked, would be roughly equal to the time period it would take for Qualcomm to have its case heard in appellate court. And you're right, I think it is the Ninth Circuit.
So I think barring investor panic in the first 48 hours after a hypothetical judgment against Qualcomm, I don't think that we see a -- we have a very real threat of anything really horrible happening. I think that Qualcomm's business is solid. I think that the industry itself is solid, and I don't think it's all going to fall apart overnight.
Geoffrey A. Manne: I just want to add one quick thing. --
Dean Reuter: Yep. Final thought.
Geoffrey A. Manne: -- I have to credit Olivier with this. I think it is really important to -- and I don't know exactly what Qualcomm's balance sheet looks like, so I don't know how much of their R&D, for example, is funded by licensing and chip sales directly versus other sources of investment revenue. But while the former might not be immediately impacted, the later could be, right? It would have to be because it's probabilistic. If they get a bad judgment, even if the remedy is stayed pending appeal, and even if there's some likelihood of reversal on appeal, to the extent Qualcomm is reliant on sources of investment funding other than its existing contracts and those sales of IP and chips, I think that revenue would be affected immediately. I don't know by how much, of course, because it would be adjusted or discounted by the likelihood of an eventual bad outcome. My point is just to say, and I think this is consistent with that Olivier was just saying, that that effect could happen very quickly, even if the direct revenue effect weren't as immediate.
Dean Reuter: Very good. I'm afraid we're going to have to leave it right there. I want to thank our two guests for your patience and your endurance here. We certainly appreciate it. Thanks to Olivier Blanchard and Geoff Manne. To our audience, thanks as well for dialing in, and for hanging in, and for your questions. A reminder to our audience to check The Federalist Society's website and check your emails as well for information on our upcoming teleforum conference calls. But until that next call, we are adjourned. Thank you very much, everyone.
Operator: Thank you for listening. We hope you enjoyed this practice group podcast. For materials related to this podcast and other Federalist Society multimedia, please visit The Federalist Society's website at fedsoc.org/multimedia.