An Unpersuasive Argument against Regulation

by on November 10, 2006 · 12 comments

Last week, my friend Brooke was kind enough to link approvingly to my post on the phantom threat of network discrimination. Brooke agrees with me that those who think “Verizon is just itching for the opportunity to detect and block every packet of data it carries that mentions the Second Amendment” are nuts.

She goes on to offer an example of a case where network discrimination would be beneficial:

One fear, however, didn’t make Tim’s list; it’s the fear that the ISPs will do exactly what we think they’ll do, which is to introduce tiered pricing for content delivery…

Suppose some new tech-tinkering über-geeks come up with a search engine even better than Google. Because they lack brand recognition, they need to keep expenses at a minimum while word of mouth slowly spreads about their better quality. In net neutrality America, they cannot keep expenses down by opting for lower quality delivery than that offered by Google. Delivery speed is not a viable option for competition; everyone has to ship at the $11 rate. Now imagine that one of the über-geeks is a trust fund baby. He’s so sure that his product is superior, he invests his trust fund in über-geeks, Inc. so they can buy higher speed delivery than Google offers, thus giving Google a serious competitive run for its money. Sadly this too is not an option in net neutrality land.

Prices and price flexibility are essential to competition. The fear that content competition will suffer without regulation is absurd on its face. Indeed, net neutrality regulation will rob new innovators and content creators of the very tools that would make challenging already established businesses possible. It’s little wonder then that the already established businesses–like Amazon, E-bay, and Google, to name a few–are fighting for net neutrality tooth and nail.


With all due respect to Brooke, I don’t find this example very compelling. She suggests that a web site might want to accept slower connectivity in order to save money. But bandwidth these days is incredibly cheap–so cheap that it’s a trivial fraction of most Internet companies’ budgets. Putting yourself at a competitive disadvantage in order to save a few dollars on bandwidth costs doesn’t make any sense.

Brooke also suggests that a site might pay Comcast extra for faster delivery. But Google already measures its page load times in fractions of a second. If Google’s home page loads in, say, 0.6 seconds, it’s not obvious how much of a competitive advantage it would be to have your home page load in 0.4 seconds.

Even if you don’t buy those objections, the kind of charges Brooke describes would be extremely impractical, because broadband ISPs like Comcast and Verizon don’t have any good way to collect them. A web site pays its own ISP for connectivity, and it’s the ISP’s job to negotiate interconnection agreements with other ISPs. Web sites don’t negotiate separate service agreements with every single ISP. Hence, even if Comcast wanted to charge individual web sites higher or lower rates for different levels of service, the administrative overhead of identifying and billing millions of tiny websites would make it more trouble than it would be worth.

(And that’s assuming they would even have the leverage to get web sites to cough up cash. I don’t have time to flesh out the argument here, but I think there are good reasons to think that they wouldn’t)

I feel bad picking on Brooke, because she’s far from the only person on the anti-neutrality-regulation side of the fence to make this kind of argument. I think free marketeers generally undermine their own case when they argue in favor of this sort of discrimination. It’s not a good idea to have every ISP charging every web site for access to its customers. In fact, it’s such a bad idea that I suspect that any attempts to do it will quickly fall flat. Indeed, that’s a big part of the reason I’m against new regulations.

There’s always a tension in libertarian thought between “X is good” and “X is bad, but it shouldn’t be illegal.” The drug war is an example of an issue where you’ve got libertarians (and libertarian-leaning conservatives) making both arguments. The first question–”Is X good or bad?”–is not an ideological or policy question. In this case, it’s a computer science question. And although there’s certainly some dissent, the overwhelming majority of the computer scientists seem to believe that network neutrality is an important principle of Internet architecture, and that it would suck if it changed. That includes critics of regulation like Brad Templeton and Ed Felten.

There might be limited circumstances in which there are plausible reasons for discriminating, but I’ve yet to see anyone make a serious argument that network neutrality should be abandoned as the Internet’s broad organizing principle. I think that the free-market side in the network neutrality debate undermines their own case when they argue otherwise.

  • James Gattuso

    Tim — Frankly, I’m with Brooke. Why wouldn’t a challenger potentially want to compete based on speed? As you say, it may be a matter of fractions of a second, but competition takes place on the margin — firms compete on hundreds on factors much more seemingly minor than that. This is just one more potential — and quite legitimate dimension. I’m also unpersuaded by the claim that charging for such differentiated treatment is “impractical.” There seem to be plenty of network owners who see it as quite doable — would they be fighting this battle if it wasn’t?

    I’m not saying that this sort of differentiation definitely will or will not develop. But that’s the point — in markets like these (or any market for that matter) we simply don’t know. And we also don’t know — no matter how many engineers say it would “suck” – what is best for consumer welfare. That’s what markets are there to discover. In the meantime, I wouldn’t dismiss the potential outcome Brooke outlined.

  • James Gattuso

    Tim — Frankly, I’m with Brooke. Why wouldn’t a challenger potentially want to compete based on speed? As you say, it may be a matter of fractions of a second, but competition takes place on the margin — firms compete on hundreds on factors much more seemingly minor than that. This is just one more potential — and quite legitimate dimension. I’m also unpersuaded by the claim that charging for such differentiated treatment is “impractical.” There seem to be plenty of network owners who see it as quite doable — would they be fighting this battle if it wasn’t?

    I’m not saying that this sort of differentiation definitely will or will not develop. But that’s the point — in markets like these (or any market for that matter) we simply don’t know. And we also don’t know — no matter how many engineers say it would “suck” – what is best for consumer welfare. That’s what markets are there to discover. In the meantime, I wouldn’t dismiss the potential outcome Brooke outlined.

  • AK

    I’ll dismiss it. If this hypothetical company wanted a speed advantage, they could pay their ISP for a faster connection. They could set up multiple sites or go with a a company like Akamai who has created a business around creating a speed advantage for their customers. What they shouldn’t need to do is go around and pay baksheesh to each ISP to go a little faster to their customers.

    I am not in favor of net neutrality legislation, yet. The major cases of net neutrality violation have been overseas and the FCC has properly slapped down ISPs who block VOIP, so I don’t see the need. For one thing, I think telco lobbyists would twist it horribly. But this is a ridiculous argument against it for anyone who actually understands how the Internet works.

  • AK

    I’ll dismiss it. If this hypothetical company wanted a speed advantage, they could pay their ISP for a faster connection. They could set up multiple sites or go with a a company like Akamai who has created a business around creating a speed advantage for their customers. What they shouldn’t need to do is go around and pay baksheesh to each ISP to go a little faster to their customers.

    I am not in favor of net neutrality legislation, yet. The major cases of net neutrality violation have been overseas and the FCC has properly slapped down ISPs who block VOIP, so I don’t see the need. For one thing, I think telco lobbyists would twist it horribly. But this is a ridiculous argument against it for anyone who actually understands how the Internet works.

  • http://www.techliberation.com/ Tim Lee

    Obviously, all else being equal, a website would probably be happy to pay to have his web site load faster. The question is whether a pay-for-speed scheme would make any sense.

    I’m just having a lot of trouble imagining how it would work. I was actually under-stating my case when I suggested an ISP could reduce latency from 0.6 to 0.4 seconds. In fact, I just did some testing, and the average time to load Google’s home page is about 0.3 seconds. Presumably my ISP only has control of about half that round-trip time, at most. So even if they could reduce their half of the delivery time by 2/3 (which is almost certainly infeasible), that would still mean they were able to reduce load times by 0.1 seconds.

    More likely, ISPs are already delivering those packets as fast as they can, and they couldn’t speed the delivery time up if they wanted to. Occasionally, prioritizing during periods of congestion could marginally improve performance, but for applications like web browsing your Internet connection is usually far below capacity.

    Indeed, as Ed Felten has argued, the way a two-tiered Internet would most likely work is that the network owners would build a high-speed network and then artificially throttle the traffic of companies that didn’t pay up. I don’t understand how that’s in the interests of consumers. As a consumer, I want my ISP delivering all of my packets to me as fast as possible. I don’t want them introducing artificial bottlenecks or delays in an effort to extort more money from website owners. (Obviously, ISPs have the right to charge consumers different prices for different amounts of bandwidth, but if I’ve paid for a given size of pipe, I want all of my content coming to me at full speed within that bandwidth limit)

    Imagine if a grocery store required all of its customers to stand in line for 5 minutes before they were allowed to check out–even if there were plenty of cashiers available to ring up their orders–unless the customer happened to be buying products that had paid the grocery store extra for “premium” chuck-out service. That would obviously be absurd and harmful to consumers. I don’t see how this case is different.

  • http://www.techliberation.com/ Tim Lee

    Obviously, all else being equal, a website would probably be happy to pay to have his web site load faster. The question is whether a pay-for-speed scheme would make any sense.

    I’m just having a lot of trouble imagining how it would work. I was actually under-stating my case when I suggested an ISP could reduce latency from 0.6 to 0.4 seconds. In fact, I just did some testing, and the average time to load Google’s home page is about 0.3 seconds. Presumably my ISP only has control of about half that round-trip time, at most. So even if they could reduce their half of the delivery time by 2/3 (which is almost certainly infeasible), that would still mean they were able to reduce load times by 0.1 seconds.

    More likely, ISPs are already delivering those packets as fast as they can, and they couldn’t speed the delivery time up if they wanted to. Occasionally, prioritizing during periods of congestion could marginally improve performance, but for applications like web browsing your Internet connection is usually far below capacity.

    Indeed, as Ed Felten has argued, the way a two-tiered Internet would most likely work is that the network owners would build a high-speed network and then artificially throttle the traffic of companies that didn’t pay up. I don’t understand how that’s in the interests of consumers. As a consumer, I want my ISP delivering all of my packets to me as fast as possible. I don’t want them introducing artificial bottlenecks or delays in an effort to extort more money from website owners. (Obviously, ISPs have the right to charge consumers different prices for different amounts of bandwidth, but if I’ve paid for a given size of pipe, I want all of my content coming to me at full speed within that bandwidth limit)

    Imagine if a grocery store required all of its customers to stand in line for 5 minutes before they were allowed to check out–even if there were plenty of cashiers available to ring up their orders–unless the customer happened to be buying products that had paid the grocery store extra for “premium” chuck-out service. That would obviously be absurd and harmful to consumers. I don’t see how this case is different.

  • http://www.ceiopenmarket.org Brooke

    I’m not going to pretend that I’m as familiar with the technical operation of the Internet as you are, Tim and AK. But what I’ve tried to do is paint a picture of the sorts of network evolutions that net neutrality regulation would prevent. Sure, I’ll concede that maybe I’ve made a ridiculous argument based on how the Internet currently works. I’m unconvinced though that it’s a ridiculous argument based on what might be possible with networking technologies in the future. What I am convinced of is that net neutrality regulation shuts the door on the possibility of such evolutions before the market and technology can figure out whether they are desirable to consumers and/or technologically feasible.

    I also wholly disagree with the contention, Tim, that it would necessarily “suck” if Internet architecture changed. For a number of reasons I won’t go into here, cybersecurity chief among them, perhaps reassessing the importance of the firmly held principles of the Internet is precisely the right thing to do. There it is: I’m making a make a serious argument for why abandoning network neutrality as the Internet’s broad organizing principle might be a good idea.

    I think the concept of the dumb network and the end-to-end principle make less sense now than they did when there was less variety in applications and less variety in content. As you pointed out to me, right now, identifying the kind of content (video, voice, email, web, etc) being sent over the package-switched dumb network is a difficult technical challenge. Well, maybe some applications need a network to evolve on which such differentiations aren’t a difficult technical challenge. And maybe some consumers want a network that can keep objectionable content out of their homes without having to rely on end-technologies or pop-up blocking software. Wayne has made the argument in favor of the “splinternets” on several occasions, and it is a very reasoned and seriousÃ?¢â?‰?though unpopularÃ?¢â?‰?argument that network neutrality should be abandoned as the Internet’s broad organizing principle.

    Most of the opposition is that these changes would violate the “spirit” of the Internet. I find arguments against change that are based on the preservation of the “spirit” of something to be largely uncompellingÃ?¢â?‰?I just never drank the kool-aid at the altar of Internet worship.

    Finally, Tim, your grocery store analogy misidentifies who the consumer is in the scenario I laid out in my original post. The purchasing of faster transmission speed is a factor of production, not a final good. You and I aren’t the direct purchasers of the transmission speed, just as the people in line at the grocery store didn’t directly “purchase” the delivery of their groceries to the store by suppliers. The content producer (or food manufacturer) is the consumer of an intermediate service, faster transmission, in this scenario.

    If we want to use a grocery store analogy for my post, imagine General Mills trying to decide how to get their Cheerios into the grocery store. Should they send them by truck in three days, train in 36 hours, or airplane in 24 hours? They might choose the truck to keep costs down. Now imagine a strawberry farmer; his product is a little more time sensitive. He chooses a train, so he can sell at reasonable prices and pretty good freshness. His competitor may want to compete on freshness but not price, so he chooses a plane.

    I, as a consumer of the final good, want my strawberry suppliers to be able to choose between trucks, trains, and airplanes; if I’m making a smoothie, I don’t really need the freshest berries, so I’ll pick the cheapest ones there. If I’m making a strawberry shortcake though, boy howdy, I’ll upgrade to the freshest looking ones there. Similarly, I don’t necessarily want all my data packets delivered as fast as possible. I want my video packets sent faster. I want VOIP sent faster. I want my i-tunes sent faster. I don’t much care if there’s a couple of minute delay with my e-mail, but I want my IMs to get there fast. And I want companies to be able to compete with the speed at which they send me my content. If i-tunes wants to charge more for the convenience of wicked fast transmission and downloads so I can hear the new JT album on my i-pod RIGHT NOW, that’s fine with me. If I’m downloading Rod Stewart ballads so I can make a mix CD later for my mom, I don’t really care how fast it goes, so maybe I’ll pick some cheaper, slower service. Right now, that’s not an option.* In the future, maybe it could be.

    As James points out above, we simply don’t know where the markets and technology will lead; maybe the things I’m talking about will be possible and maybe they won’t. But I don’t want to dismiss the possibility of doing away with net neutrality–or dismiss the discussion of what might happen if we did–just because doing so violates the “spirit” or grand tradition of something that’s only about as old as I am. I haven’t seen any net neutrality proposals with sunset provisions on them, so I’m not going to confine my thinking to the way networks work right now.

    * Admittedly, there may be slower or faster downloads between companies as a result of capital investment in servers or whatever, but I sure want the people competing for my business to be competitive in as many ways as possible to ensure as many options in speed, quality, and price that the market can support.

  • http://www.ceiopenmarket.org Brooke

    I’m not going to pretend that I’m as familiar with the technical operation of the Internet as you are, Tim and AK. But what I’ve tried to do is paint a picture of the sorts of network evolutions that net neutrality regulation would prevent. Sure, I’ll concede that maybe I’ve made a ridiculous argument based on how the Internet currently works. I’m unconvinced though that it’s a ridiculous argument based on what might be possible with networking technologies in the future. What I am convinced of is that net neutrality regulation shuts the door on the possibility of such evolutions before the market and technology can figure out whether they are desirable to consumers and/or technologically feasible.

    I also wholly disagree with the contention, Tim, that it would necessarily “suck” if Internet architecture changed. For a number of reasons I won’t go into here, cybersecurity chief among them, perhaps reassessing the importance of the firmly held principles of the Internet is precisely the right thing to do. There it is: I’m making a make a serious argument for why abandoning network neutrality as the Internet’s broad organizing principle might be a good idea.

    I think the concept of the dumb network and the end-to-end principle make less sense now than they did when there was less variety in applications and less variety in content. As you pointed out to me, right now, identifying the kind of content (video, voice, email, web, etc) being sent over the package-switched dumb network is a difficult technical challenge. Well, maybe some applications need a network to evolve on which such differentiations aren’t a difficult technical challenge. And maybe some consumers want a network that can keep objectionable content out of their homes without having to rely on end-technologies or pop-up blocking software. Wayne has made the argument in favor of the “splinternets” on several occasions, and it is a very reasoned and seriousÃ?¢â?‰?though unpopularÃ?¢â?‰?argument that network neutrality should be abandoned as the Internet’s broad organizing principle.

    Most of the opposition is that these changes would violate the “spirit” of the Internet. I find arguments against change that are based on the preservation of the “spirit” of something to be largely uncompellingÃ?¢â?‰?I just never drank the kool-aid at the altar of Internet worship.

    Finally, Tim, your grocery store analogy misidentifies who the consumer is in the scenario I laid out in my original post. The purchasing of faster transmission speed is a factor of production, not a final good. You and I aren’t the direct purchasers of the transmission speed, just as the people in line at the grocery store didn’t directly “purchase” the delivery of their groceries to the store by suppliers. The content producer (or food manufacturer) is the consumer of an intermediate service, faster transmission, in this scenario.

    If we want to use a grocery store analogy for my post, imagine General Mills trying to decide how to get their Cheerios into the grocery store. Should they send them by truck in three days, train in 36 hours, or airplane in 24 hours? They might choose the truck to keep costs down. Now imagine a strawberry farmer; his product is a little more time sensitive. He chooses a train, so he can sell at reasonable prices and pretty good freshness. His competitor may want to compete on freshness but not price, so he chooses a plane.

    I, as a consumer of the final good, want my strawberry suppliers to be able to choose between trucks, trains, and airplanes; if I’m making a smoothie, I don’t really need the freshest berries, so I’ll pick the cheapest ones there. If I’m making a strawberry shortcake though, boy howdy, I’ll upgrade to the freshest looking ones there. Similarly, I don’t necessarily want all my data packets delivered as fast as possible. I want my video packets sent faster. I want VOIP sent faster. I want my i-tunes sent faster. I don’t much care if there’s a couple of minute delay with my e-mail, but I want my IMs to get there fast. And I want companies to be able to compete with the speed at which they send me my content. If i-tunes wants to charge more for the convenience of wicked fast transmission and downloads so I can hear the new JT album on my i-pod RIGHT NOW, that’s fine with me. If I’m downloading Rod Stewart ballads so I can make a mix CD later for my mom, I don’t really care how fast it goes, so maybe I’ll pick some cheaper, slower service. Right now, that’s not an option.* In the future, maybe it could be.

    As James points out above, we simply don’t know where the markets and technology will lead; maybe the things I’m talking about will be possible and maybe they won’t. But I don’t want to dismiss the possibility of doing away with net neutrality–or dismiss the discussion of what might happen if we did–just because doing so violates the “spirit” or grand tradition of something that’s only about as old as I am. I haven’t seen any net neutrality proposals with sunset provisions on them, so I’m not going to confine my thinking to the way networks work right now.

    * Admittedly, there may be slower or faster downloads between companies as a result of capital investment in servers or whatever, but I sure want the people competing for my business to be competitive in as many ways as possible to ensure as many options in speed, quality, and price that the market can support.

  • http://www.techliberation.com/ Tim Lee

    Brooke, thanks for taking the time to comment. I have to apologize that I don’t have time to respond to your comment in the detail it deserves, as I’m headed out of town in a few minute.

    It seems to me that you’re making two distinct arguments here:

    <ol>
    <li> Technology changes in unpredictable ways, so I don’t want government regulations unnecessarily limiting those changes.

    </li><li> Network neutrality (as an architectural principle, not a regulatory scheme) isn’t such a great idea.
    </li></ol>

    The former is a question of political philosophy and economics–about which you know a great deal. I wholeheartedly agree with it, and I think it’s great that smart and articulate people like you to be making it.

    The latter is a technical question, about which, as you admit, you don’t know very much. I think your take is off-base, and in my experience, the vast majority of computer scientists agree with me. The people who build and maintain these networks–people like Tim Berners-Lee and Vint Cerf as well as thousands of ordinary network engineers and programmers–believe that the end-to-end principle is important. I don’t think you should be deriding their opinions–developed over decades of practical experience–as mere “Internet worship.”

    By making Argument #2–especially without the technical expertise to defend it properly–you undermine your credibility with a lot of technically-minded people who are sympathetic to Argument #1.

  • http://www.techliberation.com/ Tim Lee

    Brooke, thanks for taking the time to comment. I have to apologize that I don’t have time to respond to your comment in the detail it deserves, as I’m headed out of town in a few minute.

    It seems to me that you’re making two distinct arguments here:

    1. Technology changes in unpredictable ways, so I don’t want government regulations unnecessarily limiting those changes.
    2. Network neutrality (as an architectural principle, not a regulatory scheme) isn’t such a great idea.

    The former is a question of political philosophy and economics–about which you know a great deal. I wholeheartedly agree with it, and I think it’s great that smart and articulate people like you to be making it.

    The latter is a technical question, about which, as you admit, you don’t know very much. I think your take is off-base, and in my experience, the vast majority of computer scientists agree with me. The people who build and maintain these networks–people like Tim Berners-Lee and Vint Cerf as well as thousands of ordinary network engineers and programmers–believe that the end-to-end principle is important. I don’t think you should be deriding their opinions–developed over decades of practical experience–as mere “Internet worship.”

    By making Argument #2–especially without the technical expertise to defend it properly–you undermine your credibility with a lot of technically-minded people who are sympathetic to Argument #1.

  • http://www.ceiopenmarket.org Brooke

    First, I was very careful not to make argument #2; I simply made the case that I’m unconvinced that the end-to-end principle is essential and unconvinced that blind adherence to it is the best way to continue.

    Second, I don’t think argument #2 is necessarily just a technical question, but also a philosophical one.

    Third, I did not and would not admit to not knowing very much about the technical aspects; I said I didn’t know as much as you, and I tried to do it graciously since I was about to disagree with you, just as you keep graciously thanking me for my comments before disagreeing with them. I don’t have the technical expertise to defend my assertion that net neutrality might not be the best idea moving forward, but unless you can see into the future, you certainly don’t have the technical expertise to defend the assertion that it is the best idea. As we have agreed, technology changes in unpredictable ways.

    Fourth, there are a number of people whose opinions–such as yourself–I value greatly on these issues, who I don’t think drank the kool-aid and I don’t think are engaging in Internet worship. You haven’t argued about the “spirit” of the Internet here, and I in no way mean to deride you or your opinions. But my friend, if you ever take the title of Chief Internet Evangelist, I will mock you mercilessly and you will deserve it.

  • http://www.ceiopenmarket.org Brooke

    First, I was very careful not to make argument #2; I simply made the case that I’m unconvinced that the end-to-end principle is essential and unconvinced that blind adherence to it is the best way to continue.

    Second, I don’t think argument #2 is necessarily just a technical question, but also a philosophical one.

    Third, I did not and would not admit to not knowing very much about the technical aspects; I said I didn’t know as much as you, and I tried to do it graciously since I was about to disagree with you, just as you keep graciously thanking me for my comments before disagreeing with them. I don’t have the technical expertise to defend my assertion that net neutrality might not be the best idea moving forward, but unless you can see into the future, you certainly don’t have the technical expertise to defend the assertion that it is the best idea. As we have agreed, technology changes in unpredictable ways.

    Fourth, there are a number of people whose opinions–such as yourself–I value greatly on these issues, who I don’t think drank the kool-aid and I don’t think are engaging in Internet worship. You haven’t argued about the “spirit” of the Internet here, and I in no way mean to deride you or your opinions. But my friend, if you ever take the title of Chief Internet Evangelist, I will mock you mercilessly and you will deserve it.

Previous post:

Next post: