What is Broadband?

Chris:  Hello. Welcome to the maiden voyage of TPI’s new podcast – Two Think Minimum. I’m Chris McGurn, TPI’s Director of Communications. Each week on this podcast, we’ll facilitate a conversation between TPI fellows and special guests on some of the most pressing and important issues in tech policy and tech politics.

This week’s episode features a very important conversation on a topic that couldn’t be more relevant.  That is broadband.  What it is, what its definition is, and how policymakers define it.  Today we have TPI Fellow, Sarah Oh, and TPI Senior Fellow, Scott Wallsten, who will discuss this issue and shed some light on what broadband is.

Sarah:  Hey, Scott. Do you have time for a question or two?

Scott:  I do, but shouldn’t you be working on your journal article revision?

Sarah:  Yes, I’m working on it. I just got so many comments, it’s taking some time. I thought I’d take a break and bug you. So, what do you think about broadband definitions? Do you think 25 megabits down and 3 megabits up is a good definition for broadband? What’s the significance?

Scott:  That’s a good question, why are you interested all of a sudden?

Sarah:  I read a headline that people were talking about that number for the 2018 Broadband Progress Report. They used it for the 2016 Report[1] and there is some talk of change. Why now?

Scott:  I think people put way too much emphasis on the definition of broadband. I don’t think we really need a definition in order to talk about it. It’s not the case that 25 megabits per second is broadband and 24 megabits per second is suddenly not broadband and you can’t do anything with it.

Broadband – what we’re interested in – is the distribution of speeds across the country and more importantly, what you can and can’t do with different speeds and how much people value those different speeds.

Chris:  Hi, Scott.

Scott:  Chris.

Chris:  Oh, hey, Sarah. How are you doing?

Sarah:  Hey, Chris.

Chris:  What you guys talking about?

Sarah:  We’re talking about broadband definitions. What do you think broadband is?

Chris:  Broadband is how I stream all this stuff slowly on my computer at home. What are your definitions of broadband?

Scott:  I was just giving a long history. Unfortunately you interrupted me so nobody had to listen to all of that. But the FCC has gone up to 4 megabits per second downstream, at 1 megabit per second upstream, to most recently, 25 megabits per second downstream and 3 megabits per second upstream.[2]

The only reason it needs these definitions are, first, it needs to be able to define a minimum quality of service for our Universal Service Funds, and, then, it’s also used for competition analyses.

Chris:  What would those definitions mean for the just usual user at home, someone without a PhD in this stuff? What would those upstream and downstream numbers mean?

Scott:  What users care about is what they can do with their connections. And a typical user is going to want to do very low bandwidth things like send email and Twitter and so on, and some higher bandwidth things like stream HD video.

An HD Netflix stream takes about 5 megabits per second; a super HD Netflix stream can take 2 or 3 times that. And if you have multiple people in your house then you could have multiple screens going at one time. But most people aren’t going to notice much of a difference beyond say 20, 25 megabits per second.

Sarah:  For video gamers who need real-time response, wouldn’t latency be more important than just a number like broadband, the volume of data?

Scott:  That’s a really important point. If you’re a gamer you want to make sure that when you press the button on your controller everybody else who is playing the game sees you shoot the gun at exactly that moment. And latency is of course how long it takes a packet to go from your computer to a server and back.[3]

Gamers are going to care more about latency than about bandwidth per se, and that’s pretty important. And so far that’s been ignored in the definitions, except at one of the upcoming auctions, which we can talk about later.

Sarah:  I remember broadband being defined as 10 down and 1 up for Universal Service, the Connect America Fund. Why does rural broadband get defined down and what kind of policy implications are there for setting levels like that?[4]

Scott:  That’s a really good question, because one of the main reasons to have a definition in the first place is for policy purposes like the Universal Service Fund. If you’re going to give subsidies to companies to provide a service you have to define the minimum level of service that they’ll provide.

To that extent the 25/3 doesn’t really have policy implications beyond doing competitive analyses. The one that does have policy implications or implications in the real world is the 10/1 standard, because rural companies will have to meet that standard.

That’s a nod to the realities of providing service in more expensive areas that it’s more costly to provide faster service and it’s recognizing some kind of trade-off between building-out and bulking-up.

Sarah:  While we’re talking about other cases, how about mobile broadband? The 2016 Broadband Progress Report didn’t define mobile broadband.[5] LTE is around 15 megabits down, what do you think? Should mobile broadband be defined?

Scott:  The real question is how much of a competitor is mobile to fixed wireline broadband. And the way to do that is to measure the cross elasticities of substitution. Mobile broadband is getting better and better, and nearly all providers have various unlimited plans and that begins to take care of the data cap problem with mobile broadband.

It’s still not a perfect competitor, you can’t use your mobile connection to, say, stream video on your TV at high definition without it counting against your data cap. Since that’s a major component of what people use their home broadband for it’s clearly not a perfect substitute. But it’s a substitute for some people, and it’s getting better, it’s becoming an increasingly a substitute all the time.

Sarah:  I heard that 5G technologies will offer up to 20 gigabytes per second, but maybe in a decade.

Scott:  We’ll see. 5G is going to be a big deal, but it’s going to take a lot to get there that the companies are especially worried about trying to get access to rights-of-ways and poles to put their small cells on. As far as I understand it seems like you’re going to need a small cell every five feet, we’ll be tripping over them to provide the 5G service.

Sarah:  Chris, do you have any thoughts about your broadband speeds at home? Do you check what you have or is it what you use?

Chris:  I’m just curious of the comparison between the fixed and the mobile like you were talking about. In our house there’s my wife and I, we each have at least four devices at any given time – our iPads, our mobile phones, our laptop computers. There definitely times where there’s more of a lag in our service that we notice. How can people bounce from the fixed to the mobile more easily? Will this new standard have any bearing on that?

Scott:  That’s an interesting point because, as you’re implying, what people really care about is how they can use their devices, how many devices they can use, what they can do on them, and how well they work.

When you’re at home and you’re using Wi-Fi, that’s a part of your home connection. Usually that’s going to be based on your fixed wireless [oops – meant to say “wireline”] connection, rather than the cellular mobile connection.

How well your cellular connection works is going to depend on a lot of things, it’s going to depend on how strong the cellular signal is where you are, how many other people are trying to connect in that particular beam, the equipment that you’re using, and what you’re trying to do. Those depend on lots of factors.

Sarah:  Part of the debate is that, well, 25 down and 3 up, is too low, we should be demanding higher broadband speeds. We should have a gigabit of data. If we’re defining it so low then the pressure on providers isn’t there.

What do you think about the trade-offs for building out networks? Should we be deploying broadband to people who don’t even have 25/3? Or should we be building it out for a gigabit for everyone?

Scott:  Even though I know you’re playing devil’s advocate, tell us a little more about what, at least, people believe to be the benefits of gigabit service.

Sarah:  It’s really just a simple idea that more broadband is better. We need just more of it, because it’s a good thing. But a gigabit of service even when I use big data on the cloud, when I send commands to my server somewhere else, I only need a trickle of broadband speed. I can send one line through a terminal and I don’t need a gigabit. Only if I’m downloading a lot of data really quickly would I ever need a gigabit.

Scott:  We did some research recently. I did a paper recently with Jeff Prince [and Yu-Hsin Lu] at University of Indiana [sorry, that should be “Indiana University”] looking at how much people are willing to pay for different speeds. What we discovered was that the incremental amount people are willing to pay increases pretty quickly as you go up to about 50 megabits per second.[6]  But then the increase starts to decline. By the time you get to 100 megabits per second, people aren’t willing to pay almost nothing to go beyond that.

For example, to go from 100 megabits per second to a gigabit, people value that very, very little, whereas going from say 10 megabits per second to 25 megabits per second they value a lot. And that’s a function of what we need to bandwidth for. We understand that people recognize that there’s not a lot of value in going beyond 100 megabit level.

Sarah:  Speaking of trade-offs, if we’re defining broadband as 25 down and 3 up, I remember the 2016 Broadband Progress Report saying that 10% of Americans didn’t have access to 25/3 yet, and actually 39% of rural Americans didn’t have 25/3.[7]

Wouldn’t it be a good thing, it’s a good thing, right, to have some sort of broadband definition so then we can compare, let’s say, the 2018 Report to the 2016 Report? I’d like to know if fewer than 10% of Americans don’t have access to 25.

Scott:  That’s a great example of why the arbitrariness of 25 megabits per second ends up having implications, because what’s important isn’t really what share of Americans don’t have 25 megabits per second. We want to know a better, a more full picture of what broadband looks like. So, those people who don’t have access to 25 megabits per second, do they have access to 20 megabits per second? That’s why we want to know the distribution. And we can still compare how the distribution changes from year to year.

When you have a distribution, then when you’re thinking about how to improve the situation, you can also begin to think about how much additional investment would be required for different areas to get to increase the speeds by some amount.  Then you can compare it to how much people actually value those speeds, and ultimately make a calculation as to whether the benefits are worth the cost.[8]

Sarah:  Great. Thanks, Scott. This is a good conversation. It was a good break. But I need to get back to my paper now.

Scott:  That’s probably a good idea.

Sarah:  Thanks, Scott. See you.

Scott:  Bye.

Chris:  Thank you, Sarah, and thank you, Scott, for a very lively conversation on the topic of broadband for our inaugural Two Think Minimum. Please join us next week when we have another topic that relates to tech policy and politics for you guys to listen to.