On today’s episode, Silicon Flatirons senior fellow Pierre De Vries joins student attorney Andrew Manley for a conversation about spectrum management, how usage of spectrum can be improved, and the value of interdisciplinary education.
Music of the show is provided by Gino and the Goons:
Opening theme: Troubles
Sign off music: On My Way
Licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International license.
[Editor’s note: we’re starting to experiment with transcripts for accessibility, searchability, and other reasons—but obviously have a ways to go. Please give us feedback on ways to improve the presentation!]
Andrew: Hey, folks. Andrew Manley back again with another episode of the Samuelson-Glushko Tech Law & Policy Clinic Podcast. For today’s episode, I spoke with Pierre de Vries, a senior fellow here at CU’s own Silicon Flatirons Center. Our conversation today revolves mainly around how we use and manage radio spectrum, focusing on the different schools of thought when it comes to spectrum management, and how we can manage it better as times change and technology advances. As I think you’ll hear in the interview, Pierre is a multi-disciplined specialist and he’s really good at being able to thoughtfully and clearly explain hard engineering concepts in ways that they’re digestible for us non-engineers. The interview is a little longer than some of our others, but I’m really glad Pierre took the time, because there’s a lot of good stuff in there. Thanks so much for tuning in. Enjoy the show.
Pierre, thanks so much for taking the time to speak with me today. To get started, how did you come to be involved with the Silicon Flatirons Center?
Pierre: It’s quite a strange story. I didn’t realize until quite recently, in fact, how the connection happened, because I just got a call out of the blue one day from Phil Weiser. But quite recently, I was chatting to Brad Bernthal, Professor Bernthal, and I learned that it was because of a paper I gave at a conference in Dublin in 2007. It was a paper about metaphors in spectrum policy and the fact that people who do spectrum policy seem to be fixated on [00:02:00] the metaphor that spectrum is like real estate; that there are, in fact, other ways to think about spectrum than just real estate; and that thinking about real estate is not necessarily always a good idea. It turns out that Brad enjoyed the talk. Actually, I met him there. We chatted afterwards, and I enjoyed talking to him, and he must have spoken to Prof. Weiser. One thing led to another, and so I’ve been associated with Flatirons for going on ten years now.
Andrew: So, you mentioned that there is this idea that spectrum is, like…it’s a metaphor for property rights, or that you could regulate it as analogous to property rights. What are the problems with taking that approach? Why is that ineffective?
Pierre: Every metaphor has strengths and weaknesses, because a metaphor is just a model. It says, “This thing is mostly like this other thing.” The problem with the “spectrum is real estate” metaphor is it makes a bunch of assumptions—for example, that spectrum is a thing, like a piece of land. And it leads to what I believe are really futile and fruitless arguments about this “resource,” which we now assume spectrum is. How do we manage it? How do we parcel it out? And the kind of arguments you get into is, Well, if spectrum is like land, should we treat it like a commons? Or should we treat it like real estate and private property? And there have been long political arguments about that with no real answer, because, as it turns out, there are [00:04:00] property-like ways of allowing people to use radios, and there are commons or park-like ways to allow people to use radios, and they both work.
So, for example, in your phone, when you make a call through Verizon or AT&T or T-Mobile or whoever, the voice part of your call probably goes over frequencies that AT&T, for example, has bought the right to use at auction. So, that’s property-like. But your phone also has a Wi-Fi connection, which is a piece of technology that everybody whose Wi-Fi radio meets the FCC’s criteria can just use—without getting any further permissions or paying anybody anything to use the radios. So, that’s much more like a commons. So, we use both. That means that this “spectrum is real estate”, or more generally “spectrum is a resource,” conversation makes us focus on unproductive problems.
I would also argue it distracts us from more important things. One of the most important is that it’s really not a question of spectrum, whatever spectrum is. It’s what radios do, how we use radios, how we develop the technology, and how we manage the technology. And if we argue about “spectrum is real estate”, we lose sight of that question.
Andrew: So, is it possible, then, that as technology gets better and technology improves, you know, the ways that we’re going to manage spectrum is obviously going to change? So, you think back, the original Wi-Fi routers, which were… And that is the same band of spectrum that’s used in a microwave, a microwave oven. [00:06:00] The original ones, they wouldn’t be able to cut through a microwave, which is just blasting out all this radiation. But, as we’ve gotten better at it, they’re more able to cut through. So, as this technology changes, does that necessarily mean that we have to change the way we approach spectrum management, as a result?
Pierre: That is one of the key questions in spectrum policy debates, and it’s also one of the key challenges for anybody who tries to manage spectrum. Before I give you a few examples, let me just say that when I say manage, I don’t only mean regulate. I don’t only mean the government. I also mean the University of Colorado managing the radios on its campus, or me in my house trying to manage all the radios in my house. So, management occurs at lots of different levels.
As technology changes, to your question, some of the assumptions that you make no longer hold, and you need new ways of regulating. For example, at the government level, the old way and the traditional way and still the dominant way for the government to manage radio operation is to say, “You can use this kind of radio to do this kind of thing in these frequencies in this place.” So, for example, “You can broadcast television programming. You have to use this particular broadcast technology. You have a license to use these frequencies in Boulder.” So, that’s called command and control. The regulator decides everything up front. [00:08:00]
That doesn’t work over the very long term, because there are new things you can do with radio. So, for example, one could, using the same channel that you broadcast a TV station, broadcast or serve cellular telephones. But the original broadcast license only allowed somebody to broadcast TV programming. So, we’re in a huge, complicated auction transaction right now called the Incentive Auction to find a way to transfer the use of some of the TV bands from TV broadcasting to cellular. And the new kinds of licenses—the licenses that Verizon, AT&T, and T-Mobile have—are flexible use licenses, so that the FCC doesn’t specify what technology they have to use, or, in fact, what they do, whether voice calls or streaming video or something else. It’s completely up to the operator. So, that is a big change that was inspired by technology.
Another change that, again, took place because of technology was the new rules back in the mid-1980s that allowed the kinds of unlicensed that we see now, like Wi-Fi and Bluetooth and ZigBee and on and on. The engineers at the FCC (specifically, a guy called Michael Marcus) realized that new technologies were being developed that enabled radios to be resistant to the kind of noise that you get from microwave ovens. [00:10:00] Nobody knew what anybody was going to do with them, but the FCC made these new rules and said if you use just this very generic kind of technology with very generic conditions on the power levels that you can transmit (so that you don’t interfere with other people) anybody can do it. And that has led to Wi-Fi and other unlicensed technologies.
Andrew: And so, this is actually getting to be an issue with the unlicensed technologies, right? Because they become so ubiquitous that, in the Wi-Fi band in particular, we’re starting to see conflicts with, you know, something like an LTEU, right. So, how should we address that if they’re unlicensed, but we don’t want to, I guess, rock the boat and possibly disrupt these unlicensed technologies that we’re all relying on, even though our current, I guess, regulatory scheme is more of the commons?
Pierre: It’s a question of which boat you don’t want to rock, and the Wi-Fi LTE argument (LTE is a technology that is used in cellular phones almost ubiquitously now) is pretty typical of the kinds of debates that regulators have to deal with.
So, for example, with unlicensed, one kind of boat that you want to keep stable is the boat that brought us to this wonderful destination, which is that anybody can use any technology that they like, as long as it meets very simple [00:12:00] maximum power transmission constraints in this band. And they can create any kind of technology, from baby monitors to Bluetooth to anything else they can conceive of. Part of the rules for the unlicensed bands that have made it successful is that you have to accept interference from anybody else. That’s why you don’t have to pay for your license.
One of the things that Verizon and AT&T and so on pay billions and billions of dollars to get in their license is the protection from interference. If anybody interferes with Verizon’s service, Verizon can stop them. That allows Verizon to provide the kind of quality that they feel their customers need. In the unlicensed bands, you don’t pay, therefore you don’t get protection, but it’s led to great innovation. So, that’s one boat. And, of course, one side of the argument will say that’s the boat we shouldn’t rock.
Another boat, as you describe, is, well, we’ve now created this world where we’ve got Wi-Fi services which we all depend on, and we don’t want that to change. We don’t want to feel threatened by any other new technology, and therefore, we need to protect the Wi-Fi service. That actually says we are moving from engineering rules for unlicensed, which says no more than a certain transmit power maximum, to essentially an economic incumbency rule which says once a service has become an incumbent, we need to protect it. And, for my money, of the two, I am more inclined to favor [00:14:00] innovation than protecting the incumbent, and for two reasons.
One is it’s very hard to predict the future. There are no facts in the future. So, you never know whether there is, in fact, going to be any interference and whether this incumbent that we have will continue to deliver the value that we have had so far.
But, more importantly—and one of the main problems, what’s ineffective about how spectrum or radios are currently managed—is the disproportionate power that incumbents have. The way it works with radio policy is you get a license or you start delivering a service, you build a business around that, and that business includes a legion of well-paid lobbyists, and you then protect what you have built up at all costs. Those are costs that everybody bears, and the costs include new services not being allowed. For example, let’s go back to the mid-1980s. The incumbents at the time said, “No, don’t allow this kind of unlicensed. It’s going to harm our services, and there’s no point.” If we had protected those incumbents back then, we wouldn’t have had Wi-Fi now.
The last point, of course, and this is where I put on my proto-engineering hat—I’m not an engineer, but I spend a lot of time with them—is there is a factual question about whether these other technologies like LTE would, in fact, cause more harm [00:16:00] to Wi-Fi than more Wi-Fi radios. I’ve just done some work with engineering colleagues at Aachen University in Germany, and the results of that paper essentially say that there’s nothing to worry about. In most cases, there is so much frequency capacity available that Wi-Fi won’t even be aware that LTE is there, and when you force everybody together into the same channel, sometimes Wi-Fi does better with other Wi-Fi, and other times it does better with LTE. So, that means, as an engineer looking at it, my recommendation to the FCC would be there’s no basis for changing the rules.
Andrew: And I think I’ve seen similar said… Like, Qualcomm had put out something similar, saying that you put LTEU next to the Wi-Fi, then you’re actually going to get better Wi-Fi performance, versus putting two Wi-Fis next to each other.
Pierre: Yes. Qualcomm did a very cute video that essentially told the story that LTE is a better neighbor, is a better friend to Wi-Fi than Wi-Fi itself. And, of course, the Wi-Fi interests said, “Oh, no, no, no, no. LTE is the enemy. It’s going to destroy us.” And it turns out, when you do the engineering analysis, sometimes LTE is a problem, and other times, Wi-Fi is a problem. It’s a wash.
Andrew: So, to go back to the issue of the overbearing power that incumbents seem to have in our current state of regulation inspector management, [00:18:00] you brought up the Incentive Auction, and to me, that seems like it’s a very property rights approach. So, basically, what we are looking to do is we have these TV broadcasters, and we want to pay them to leave. That’s the same thing as paying someone to get off their land so you can rent it out for someone else. This seems like there could be problems, though, with creating inefficiencies effectively through spectrum squatters. Is that something we should be concerned about?
Pierre: I think it depends on your ideology. Of all the things that the FCC regulates, spectrum is probably the least ideological. But, unfortunately, that doesn’t say very much.
So, people on the right, who take a free market, libertarian, property rights view, believe and can make strong arguments that treating radio operating rights like property rights is economically efficient. People on the left will make equally strong and beautiful arguments that say, no, property rights have all sorts of costs and market failures and lead to squatting—as you say, people who have licenses and then prevent other people from using those licenses—and therefore, we need something which is much more like a commons. And that, in a way, is the socialist view.
In practice, what the FCC has done is to allow both. Both have been successful and both have their problems. [00:20:00] I think with the Incentive Auction we’re dealing with a couple of things. One, technology has moved on. Cellular data is now a “higher use,” a more valued use of any given radio channel in a particular place than a TV broadcast would be. But the way the licenses of the broadcasters have been written, they cannot use those licenses for anything other than TV broadcast. It then becomes a public policy question of how to move those licenses from broadcasting to, let’s say, flexible use cellular.
People who don’t like property rights and people who don’t like big companies will then say that is a windfall for the broadcasters, and they don’t deserve it. “The spectrum is a natural resource that is owned by the people.” (Footnote: I think that’s an unintelligible statement. I’ve never been able to figure out what the word “spectrum” means, but we might or might not come back to that.)
The people on the other side of the argument say, look, clearly there is a higher use. The cellular companies are likely to pay more for the right to deploy cellular systems than the broadcasters will demand to be paid to be go away. Therefore, let’s do a trade and get on with it. And that’s the deal that we’ve made as a country through our regulator, the FCC, and that’s what we’re trying to do with the Incentive Auction.
Andrew: So, looking at possible changes, then, to ways that we can [00:22:00] regulate and manage spectrum, I know in a paper that you and Phil Weiser, who wrote for…I believe it was the Hamilton Project, you had a couple of proposals, and among them, we want to move away from our kind of ad hoc way of adjudicating spectrum disputes and taking a look at more fact-based and actually analyzing the problems of interference, things like that. So, how would that look, then, if you had your druthers, how would we address and manage spectrum going forward?
Pierre: There are a number of things I think we need to do to manage radios better going forward. (You’ll notice that I say manage radios and not manage spectrum. As an exercise, I once tried to see if I could avoid using the word spectrum altogether, and I find it’s quite easy. You just have to say radios.)
So, how can we manage radio operation better? One of the fundamental problems with radio policy is that it’s hard to have overlapping uses of radios in the same place, at the same time, in the same frequencies. (Another way of saying that, if you’re more of the spectrum language persuasion, is “spectrum is scarce.” I say “overlapping radio use is hard.”) The question is, how do you resolve those disputes? I think we need to do things up front and [00:24:00] after the fact, that is, ex ante and ex post.
So, up front, I think it would be very helpful if the rights that people have to protection from interference—if they have any rights to protection from interference—are stated clearly and up front, before anything happens. I’ll give you two examples of how it’s problematic right now and then I’ll talk about how you might resolve it.
One way in which it’s difficult to be clear about what rights you have is that the definition of harmful interference in the FCC rules is a legal definition, not an engineering definition. Essentially it says, among other things, that harmful interference is interference which “severely degrades, obstructs, or repeatedly interrupts” another service. And you can argue about each of those terms for as long as you have money to pay your lawyers—what does “severely degrade” mean, what does “obstruct” mean, and so on. So, therefore, it’s very difficult to know before you deploy a new service whether there is going to be harmful interference or not.
Another example: What kind of protection does a radio service or a radio operator have against claims for interference going forward? This is what we see in the Wi-Fi/LTE case. What the rules currently say is that any unlicensed device needs to accept any interference from any other device, including other unlicensed devices. [00:26:00] What the FCC has done (de facto) through its LTE proceeding is to say, “Well, that’s true, except if you’re Wi-Fi, in which case we’re going to give you some extra protection.” It’s not protection in the rules: the FCC has not done a rulemaking about what extra protection Wi-Fi devices might have. It has used other techniques, shall we say, to do that. That meant that the kinds of protections that Wi-Fi had three years ago were much less than they now have in practice, and the rights that a newcomer to the unlicensed band had to cause interference now are much weaker than three years ago. And in this case, this all happened without anything changing in the rules.
So, how does one address these kinds of problems? One way is to add some engineering hoops to the harmful interference definition. For example, I’ve been working on something called harm claim thresholds. This doesn’t replace the legal definition of harmful interference, but it says a radio service needs to be able to cope with a certain amount of interference before it can go to the FCC and claim harmful interference. It says, “Just get by with a certain amount of interference, and if it exceeds it, come and complain.” When that radio [00:28:00] operator then goes and complains to the FCC, the legal process starts.
The second part of the solution, I believe, and this is something that Phil Weiser and I talked about in a paper for the Hamilton Project, is that the adjudication of claims of harm needs to be more fact-based, transparent, and timely. At the moment, if I have an interference problem and I think you are causing it, all I can do is go to the FCC’s Enforcement Bureau, ask them to do something about it, and then wait. And the Enforcement Bureau may or probably may not have the time and resources to do anything about it, and if they don’t do anything about it, I’m out of luck. What Professor Weiser and I suggested, in that paper and previous work, was that the FCC should use its administrative law judges—in other words, the adjudication function that it already has but doesn’t use—to adjudicate these disputes. So, if I have a problem and I think you’re causing the interference, rather than go to the Enforcement Bureau and wait, I can take my case directly to a judge, the administrative law judge, and ask that judge to resolve the dispute. That’s an option that we think the FCC hasn’t used, and if it did use it, would give parties additional avenues for getting problems solved, which would help them both to have more certainty about what their rights are, and to get on with whatever they’re doing that’s currently being disrupted.
Andrew: Do you think something like that would bring it closer to kind of the ways or methods behind, like, an [00:30:00] STA or ESR? So, if you wanted to expedite the process, you would bring your claim of interference before the commission, and then the alleged interferer would have to stop until they can show otherwise, or if they show otherwise, then your claim would be put on hold, or something like that.
Pierre: You’re raising legal questions which I’m not competent to answer, which is, do you have a liability claim, or can you actually get an injunction to stop some kind of interference?
That actually goes back to the nature of the operating rights. And by the way, I say operating rights: these are not actually property rights, I would suggest, not being a lawyer. But the FCC has, in fact, never been clear about what the nature of these rights are. So, we don’t know.
You also mentioned STAs and experimental licenses. These are Special Temporary Authority (STA) to operate a radio, for example, for product development or testing; or experimental licenses to do the same kind of thing. What should the burden be for obtaining these kinds of temporary permissions?
Thinking about special temporary authority and experimental licenses points to a bigger problem, or a bigger opportunity. When the FCC has to make the decision about whether to allow some new operation—and it doesn’t have to be experimental; it could be a new service—on what basis does it make that decision? At the moment, when the FCC tries to make a decision about a new service, the incumbents [00:32:00] will come to the FCC crying bloody murder. They will say, “Look: this new service is going to cause interference, and here is what could happen,” and then they will then adduce a worst case, or a parade of horribles. But, essentially, this is the worst possible thing that can happen. In practice, what happens is that the FCC thanks them, closes the door, rolls around on the floor laughing for a little while, gets up, and then shaves a little bit of the badness off the worst-case scenario and then goes with that. The FCC will use words like, “Well, you know, there’s a low risk of interference,” or, “The interference is not going to be very severe,” but this is all qualitative.
What I’ve been suggesting is that the FCC becomes more quantitative in doing a risk assessment when it tries to make these judgments. So, you’ve got an incumbent, you’ve got a new entrant. The new entrant could cause harm. That means there are hazards. A risk assessment means the FCC would look at not only how bad it can get—that’s what worst-case is—but also how likely it is that it gets that bad. It may only be a one in a million or a one in a ten million chance that it gets that bad, in which case, the FCC might choose to discount that objection.
In practice, behind closed doors, I suspect that the FCC engineers already do that, but because that’s not stated in their reports and orders, it’s very hard for the community to know the basis on which the FCC makes these decisions, which turns every new allocation into [00:34:00] a from-scratch argument with no precedent. If the FCC made more use of risk assessment and published these results, the whole community would, a) have the benefit of more evidence-based decision-making, and b) would see the evidence that the decisions were being made on and could adjust their behavior to that in the future.
Andrew: So, for people that are listening to this and maybe aren’t big on spectrum or radio policy and technology and all that, why does this matter? Why is this so important that we get to an effective way of managing this technology and regulating it effectively?
Pierre: Spectrum, or radios, are pretty invisible. And, of course, radio waves themselves are entirely invisible. But we all depend on it more and more every day. Everybody wants their wireless. And I will just give you a couple of the most obvious examples. We’ve talked about these innovations like Wi-Fi. We haven’t talked about all the innovation on the LTE side, as well, to increase the amount of data that your cellular provider can deliver to your phone. There are new kinds of technologies that benefit us. For example, medical monitoring equipment: there’s a whole raft of new devices that are body worn, sometimes in the body, that depend on radios — in other words, that depend on the radio rules. And, since medicine’s involved, you might want to make sure that there isn’t interference so that the information that you’re having a heart attack [00:36:00] actually gets to where it has to be.
When people talk about radio regulation risks, usually they talk about safety-of-life applications, and there are many of them. I’ll just give you a few. One of them, of course, is 911 calls. So, when you make a 911 call, your location matters. It used to be, in the days when we had wire line phones, that when your 911 call is routed to the public safety access point, your address was delivered to the first responders. That gets a bit harder with cellular phones, and so the FCC has, over a number of years, developed rules and requirements about the kinds of information that your cellular phone should deliver to first responders when you call and say there’s an urgency.
Another one that’s really important to all of us that travel is airplane safety. Planes use all sorts of radios to communicate with air traffic control, to navigate in the rain when they’re landing. One that’s been in the news in recent years is wind shear detection radars. At more than 45 airports in the United States, including Denver, there are radars that detect wind shear, which is an invisible meteorological phenomenon that’s particularly dangerous when a plane is about to land and can whack a plane into the ground unexpectedly. So, these radars are important for aviation safety—in other words, our life safety—but there have been cases where there has been interference to these radars from unlicensed devices, and that has caused a lot of [00:38:00] concern and additional regulation from the FCC in order to protect those services.
Andrew: Okay. So, I always want to end on kind of a fun question. Something that people might not immediately grasp when they take a look at your work is that you actually spent some time at art school. What would say… Is there… What would you say is the benefit of having that kind of interdisciplinary experience? How has that helped you throughout your radio and spectrum work?
Pierre: I work as an interdisciplinary person. So, for example, when I’m in Germany at this engineering group, I’m the expert on—or I’m seen to be the expert on—FCC policy. When I’m talking to folks at the law school, relatively speaking, I’m the expert on the engineering. But, actually, I’m neither a lawyer nor an engineer. So, I think that the ability to understand what other disciplines are talking about is important not only for people like me who make a business of translating from one to the other or interpreting from one to the other, but for everybody. So, when a lawyer is trying to make a contract or draft a rule or write a filing to achieve a good outcome in radio policy, it helps to be able to understand, at least at a first approximation, what the engineering considerations are. And likewise, for an engineer, it helps to understand the [00:40:00] legal and policy context in which your inventions could be used because some inventions are very difficult to get into the market given current rules, and current rules are very hard to change, and others are easier.
For me, studying art was useful, I think, because it was a big change from my original training as a physicist. When you’re a physicist, you spend months and months in a room, sitting at a desk with a pencil and paper writing stuff, and then at the end you write a paper and you publish the paper, and it might be only four or six pages, and that’s all anybody sees. When you’re in art school, everybody can see what you’re doing every day. Anybody can walk past what you’re doing. And so you are baring your soul, if you will, a lot more than you are if you’re doing physics and all you present is the result at the end.
I think that making policy is a lot more an art than a science. It’s more an art for the obvious reason that it’s subjective. There are a lot of value judgments involved. But it’s also an art in that it’s iterative. Making policy is a collaborative process, where you try something out, somebody else tells you why that’s a dumb idea. That person might be your client; or it might be the regulator that you go and talk to, to try your ideas out; or it might be the party that opposes your ideas. But the direction that you go keeps changing. I think that [00:42:00] there clearly is scope for zealots, people who are absolutely passionate about a particular position and will advocate for that position no matter the counter-arguments and the circumstances, and they can be very creative in coming up with arguments to bolster their position. I think in the long run, though, because the technology keeps changing and the policy environment keeps changing, one has to be adaptive and creative, and I think the people who are really effective in the long run are people who are not only willing to change their minds, but are willing to be seen to change their minds. And that’s something which, I think, making art, doing art, and, in fact, the humanities in general helps you to do.
Andrew: All right. Well, thank you so much, Pierre, for speaking with me today. Really appreciate it. I was glad to have you on the podcast.
Pierre: Thank you very much.
Andrew: How about that? As I said, great stuff. Pierre’s one of those guys where it’s always interesting to hear him talk and pick his brain, and I’m really glad that we got to cover so much ground in one sitting. Up next we have Lorrie Cranor, the Chief Technologist of the Federal Trade Commission. I sat down and spoke with Lorrie while she was in town for our privacy conference. We talked about different attempts to make online privacy mainstream, how to get more young girls interested in science, and, of course, her famous “bad password” dress. As always, music in the show is provided by Gino and the Goons through a creative commons license. Check those guys out on Facebook, ReverbNation, and the Free Music Archive. Thanks for listening, and until next time, dance like no one’s watching, but encrypt like everyone is.