I wanted to follow up on Eli Dourado’s excellent previous post (“Real Talk on Net Neutrality“) to reiterate the importance of a few points he made and add some additional thoughts about the issues raised in that New York Times article on Net neutrality and forced access regulation that lots of people are talking about today.

What Eli’s post makes clear is that there are those of us who think about Net neutrality and infrastructure regulation in economic terms (a rapidly shrinking group, unfortunately) and those who think it about in quasi-religious terms. The problem with the latter ideology of neutrality uber alles, however, is that at some point it must confront real-world economics. This is Eli’s core point: Something must pay the bills. In this case, something must cover the significant fixed costs associated with broadband investments if you hope to sustain those networks. Unless you are ready to make the plunge and suggest that the government should cover those costs through massive infrastructure expenditures and even potential nationalization or municipalization of broadband networks — and some clearly would be — then you have to get serious about how those costs will be covered by private operators.

Thus, we come back to the importance of business model experimentation and pricing flexibility to this debate. I have been harping on this point for a long time now, going all the way back to this 2005 essay, “The Real Net Neutrality Debate: Pricing Flexibility Versus Pricing Regulation.” And there’s a litany of other things I’ve penned on the same point, many of which I have cited at the end of this essay.

Here are the core points I have tried to get across in those earlier essays:

• For progress to occur in any economic system, firms must be able to freely set prices for goods and services without fear of government price controls or micromanagement of business models. Heavy-handing tech mandates — especially Internet price controls — could have a profoundly deleterious impact on investment, innovation, and competition. After all, there can be no innovation or investment without a company first turning a profit.
• The Net neutrality debate is about whether the government will allow broadband services to be differentiated or specialized for unique needs. Differentiated and prioritized services and pricing are part of almost every industrial sector in a capitalistic economy. (ex: airlines, package shipping, hotels, amusement parks, grades of gasoline, etc.)  Why should it be any different for broadband? Indeed, it is essential that such flexibility be allowed precisely because it is the key to making sure more populations get served with more diversified offerings. Of course, advocates of neutrality uber alles think this is heresy, even if it is based on sound and widely-accepted economics. They just figure you can ban all sorts of business practices without it having any consequences.
• But, again, there is no such thing as a free lunch. Something has to pay for ongoing Internet investment. It doesn’t just fall like manna from heaven. Differentiated business services and pricing can help in this regard by allowing carriers to price more intensive or specialized users and uses to ensure that carriers don’t have to hit everyone – including average household users – with the same bill for service.  Why should the government make that illegal through Net neutrality regulation?
• Net neutrality can have, and already has had, unintended consequences. Consider bandwidth caps, which critics paint as some sort of nefarious, anti-consumer plot. In reality, they are just a tool to manage capacity; a tool that has been necessitated by Net neutrality regulation. When the law says you are not allowed to differentiate or specialize service offerings, you have to find other ways to manage capacity and make sure you can recoup fixed costs. In a world without the omnipresent threat of Net neutrality regulation, things might have played out quite differently. Broadband providers might have found creative ways to have other downstream providers help defray the costs of specialized services so that consumers weren’t stuck picking up the entire bill or being forced to deal with caps. For example, video game developers like Electronic Arts and Activision might be willing to help subsidize the costs associated with online gaming by picking up that expense and then amortizing the expense over a diverse universe of online gamers. Similarly, some content companies or video services could help cross-subsidize new online video ventures to ensure those costs do not have to be spread across all customers but instead only those who most demand those services. Again, this is the alternate universe that might have played out if not for the hyperventilating of vociferous regulatory advocates who worship at the alter of perfect “neutrality” in all things. To reiterate, this is not the way any other sector of our capitalist economy works. Service differentiation and price discrimination are not some sort of bizarre anomaly; they are the norm.
• When it comes to industrial organization questions, infrastructure socialism simply isn’t a sustainable long-term alternative. Sharing is not competing. We’ve tried line-sharing and forced access regimes before and they didn’t end well. Creating networks built on paper is a dangerous endeavor. In the short-term, you can milk existing infrastructures for every drop of value they have left, but eventually the bills will come due and something must pay for sustained investment and upgrades. Facilities-based competition, not infrastructure sharing is the path forward if we want truly robust and sustainable networks and markets.

Where will this debate turn next? As we saw in today’s New York Times piece, the regulatory proponents are turnung up the heat and asking for more day-to-day Net neutrality controls, making it increasingly difficult for differentiated service offerings to develop. That leaves broadband providers in the unenviable position of telling their customers that they’ll either have to live with caps or some variant of metered pricing. But bandwidth caps are increasingly controversial and, quite honestly, completely unnecessary if the carriers are at liberty to freely price their offerings to account for traffic.

Thus, I’d be willing to bet that we’ll see more broadband providers gradually phase in metered or two-part pricing schemes. Pure metering is a harder sell since many consumers resent it and it also remains unclear how easy it is to meter bits and communicate usage patterns to consumers. This leaves two-part pricing and tiered pricing. Two-part pricing would involve a flat fee for service up to a certain level and then a per-unit / metered fee over a certain level. I don’t know where the demarcation should be in terms of where the flat rate ends and the metering begins; that’s for market experimentation to sort out. But the clear advantage of this solution is that it preserves flat-rate, all-you-can-eat pricing for casual to moderate bandwidth users and only resorts to less popular metering pricing strategies when the usage is “excessive,” however that is defined. Or you can just go with tiers of service like wireless operators already have. Of course, if you have enough graduated tiers of service, it very quickly starts to resemble a metering scheme.

In the end, there’s just no way of escaping basic economics. If the law doesn’t allow service providers to use creative schemes to more efficiently allocate fixed costs, the end user will have to pick up the full cost of service. The only interesting question left is whether Net neutrality regulation will make that illegal too.

Additional Reading:

• Netflix Falls Prey to Marginal Cost Fallacy & Pleads for a Broadband Free Ride (July 8, 2011)
• Smartphones & Usage-Based Pricing: Are Price Controls Coming? (July 12, 2011)
• Why Congestion Pricing for the iPhone & Broadband Makes Sense (October 7, 2009)
• The (Un)Free Press Calls for Internet Price Controls: “The Broadband Internet Fairness Act” (June 17, 2009)
• Free Press Hypocrisy over Metering & Internet Price Controls (June 18, 2009)
• Bandwidth Cap Hysteria & the Alternative (October 4, 2008)
• Once Again, Why Not Meter Broadband Pipes?  (September 7, 2007)
• Why Not Meter? (March 12, 2007)
• The Real Net Neutrality Debate: Pricing Flexibility Versus Pricing Regulation (October 27, 2005)

Infrastructure socialism isn’t a sustainable alternative.

Of all the shockingly naive and shamelessly self-serving editorials I’ve read by businesspeople in recent years, today’s Wall Street Journal oped by Netflix general counsel David Hyman really takes the cake. It’s an implicit plea to policymakers for broadband price controls. Hyman doesn’t like the idea of broadband operators potentially pricing bandwidth according to usage /demand and he wants action taken to stop it. Of course, why wouldn’t he say that? It’s in Netflix’s best interest to ensure that somebody else besides them picks up the tab for increased broadband consumption!

But Hyman tries to pull a fast one on the reader and suggest that scarcity is an economic illusion and that any effort by broadband operators to migrate to usage-based pricing schemes is simply a nefarious, anti-consumer plot that must be foiled. “Consumers and regulators need to take heed of what is happening and avoid winding up like the proverbial frog in a pot of boiling water,” Hyman warns. “It’s time to jump before it’s too late.”

Rubbish! The only thing policymakers need to do is avoid myopic, misguided advice like Hyman’s, which isn’t based on one iota of economic theory or evidence.

Hyman’s economic illiteracy is evident from the get-go. He tries to spook people with the headline, “Why Bandwidth Pricing Is Anti-Competitive.” No it isn’t. Usage-based pricing is used in countless economic sectors every day and it is overwhelming viewed by economists as a sensible way to calibrate supply and demand while ensuring costs are covered. But Hyman says the laws of economics don’t apply to broadband!  No seriously, he says:

Cable and telecom companies argue that bandwidth is a scarce resource and that imposing caps and overage fees will relieve pressure on high-speed networks. Families pay more when they use more electricity, these companies point out, so why shouldn’t households pay more if they use more bandwidth?  The analogy is a false one. Wireline bandwidth is an almost unlimited resource due to advances in Internet architecture. Adding more capacity is easy. The marginal cost of providing an extra gigabyte of data—enough to deliver one episode of “30 Rock” from Netflix—is less than one cent, and falling. […] Consumer access to unlimited bandwidth is good for society. It fosters innovation, drives commerce, and advances political and social discourse. Given that bandwidth is cheap and plentiful and will only grow more so with time, there is no good reason for bandwidth caps and fees to take root.

Oh my goodness. Really? Hyman appears to be suffering from a rather serious case of marginal cost fallacy: the belief that prices should, as a rule, equal marginal costs. The problem with such thinking is that it leaves zero room for investment, innovation, and other real-world dynamics that get conveniently forgotten as “fixed costs.”  Of course, if you begin with the truly outrageous claim that “bandwidth is an almost unlimited resource,” and “bandwidth is cheap and plentiful and will only grow more so with time,” then it’s only logical that you’d fall prey to this fallacy!

Meanwhile, back in the real world, economists and financial analysts will explain to you that high fixed-cost goods like broadband networks don’t just grow on trees or fall like manna from heaven. Yes, of course it is true that “consumer access to unlimited bandwidth is good for society.”  But the same is true of countless other goods that we’d all like to have access to at zero cost. But that doesn’t invalidate the fundamental laws of economics. Someone financed and built those networks and someone has to keep building and improving them. You’d never get anything built if you adopted the view that scarcity was a myth and that prices must equal marginal cost.

On that point, I was tickled to see in the online comments to Hyman’s piece that one gentleman asked, “what happens when you allow unlimited access at.. marginal cost?” and for another to say in response, “Answer: You turn into Greece.”   Quite right. There is no free lunch. Something has to pay the bills, including the broadband bills. You can’t just free-ride on the future forever by pretending that bandwidth is an abundant good and holding prices at or below marginal cost.

Once you understand these facts you can point out what’s really wrong with Mr. Hyman’s reasoning: He basically wants average costs for all consumers to go up so that his costs (or the costs of any high-bandwidth use or user) will never go up. Shameful!  Indeed, let’s just call Mr. Hyman’s editorial what it is: a blatant attempt to get government to impose price controls on broadband providers to favor his company. End of story.  He could have spared us all the sloppy economic sophistry and just told us that. It would have made it a bit easier to take him seriously.

P.S. Incidentally, Mr. Hyman’s one and only suggestion for how to deal with network demand/congestion is this: “If Internet service providers really wanted to manage traffic efficiently, they would limit speeds at peak times.” Interesting. I wonder how the Net neutrality crowd feels about Netflix’s new-found love of broadband throttling!

This morning, the Technology Committee of the New York City Council convened a large hearing on a resolution urging Congress to pass a robust Net Neutrality law. I was supposed to testify, but our narrowband transportation system prevented me from getting to New York. Here, however, is the testimony I prepared. It focuses on investment, innovation, and the impact Net Neutrality would have on both.

“Net Neutrality’s Impact on Internet Innovation” – by Bret Swanson – 11.20.09

Interesting piece here from Slate’s Farhad Manjoo on why AT&T should dump unlimited data plans and end what he calls the “iPhone all-you-can-eat buffet.”  He notes that: “The typical smartphone customer consumes about 40 to 80 megabytes of wireless capacity a month. The typical iPhone customer uses 400 MB a month. AT&T’s network is getting crushed by that demand.” Because “some iPhone owners are hogging the network” and causing “a slowed-down wireless network,” Manjoo recommends a congestion pricing model as a method of balancing supply and demand:

How would my plan work? I propose charging $10 a month for each 100 MB you upload or download on your phone, with a maximum of $40 per month. In other words, people who use 400 MB or more per month will pay $40 for their plan, or $10 more than they pay now. Everybody else will pay their current rate—or less, as little as $10 a month. To summarize: If you don’t use your iPhone very much, your current monthly rates will go down; if you use it a lot, your rates will increase. (Of course, only your usage of AT&T’s cellular network would count toward your plan; what you do on Wi-Fi wouldn’t matter.) To understand the advantages of tiered pricing, let’s look at AT&T’s current strategy of spending billions to build more network space. Why won’t this work? For the same reason building more roads doesn’t reduce traffic—more capacity increases the attractiveness of driving, which brings a lot more cars to the road, which leads to more gridlock.

Congestion pricing and metering is something I’ve written quite a bit about here in the context of wireline broadband (1, 2, 3), but Manjoo is equally correct that it could be applied for wireless data plans.  It has the added value of taking pressure off lawmakers to impose Net neutrality regulation since pricing of the pipe becomes an effective substitute for most other forms of network management. In other words, price, don’t block bandwidth-hogging customers and applications.  The problem, Manjoo explains:

Of course, users would cry bloody murder at first. The traditional criticism of tiered pricing on telecommunications systems is that it’s too expensive and too annoying for customers; people don’t know how much they’re spending during the month, and then they’re smacked with huge bills. Most Internet companies aren’t big fans of tiered pricing, either. They worry that adding a meter to Internet time will reduce people’s propensity to try out new stuff online—killing innovation on the world’s most innovative communications platform. But tiered pricing on the iPhone doesn’t have to be onerous. I’d call on AT&T to create automatic tiers—everyone would start out on the $10/100 MB plan each month, and your price would go up automatically as your usage passes each 100 MB tier. The key to implementing this policy is transparency. The phone should have an indicator—sort of like the battery bar—that changes color as you pass each monthly tier. That way, people can adjust their usage to suit how much they’d like to pay—limiting surfing if they approach the next tier, or deciding to press on if money’s no object.

What Manjoo is getting at here is what economists refer to as a “Ramsey two-part tariff.” A two-part tariff (or price) would involve a flat fee for service up to a certain level and then a per-unit / metered fee over a certain level. It is widely regarded by most economists as the most efficient and pragmatic solution to high-fixed cost, low marginal cost investment conundrums.  It’s hard to know where the demarcation should be in terms of where the flat rate ends and the metering begins, but that’s for market experimentation to sort out. But the clear advantage of this solution is that it preserves flat-rate, all-you-can-eat pricing for casual to moderate bandwidth users and only resorts to less popular metering pricing strategies when the usage is “excessive,” however that is defined.

Some companies have shown signs of embracing it, but few have formally adopted congestion pricing or metering.  Worse yet, some of the regulation-happy activist groups in D.C. (like the neo-Marxist charlatans as the UnFree Press) have already made ridiculous accusations that metered pricing is somehow “unfair” when, in reality, it is the fairest system under the sun. There’s even been legislation introduced by Rep. Eric Massa (D-NY) that would forbid the practice through the imposition of Internet price controls.  Foreclosing experimentation with such innovative pricing schemes would be a real innovation-killer.

I hope we get there eventually for all high-speed data services, whether we are talking wireline or wireless. Although I generally try to be agnostic about business models, I think this one is worth doing a little cheerleading for because it helps take regulatory pressure off the marketplace.  Pricing also acts as a signal for others innovators and entrepreneurs in the market regarding how to adjust investment strategies or enter new markets.

Those who advocate regulating Internet service providers as common carriers subject to “open access” mandates (a/k/a “Net Neutrality”) want us to believe that their cause is the “Civil Rights” issue of the digital age, with huge popular support and opposed only by self-interested cable companies and their henchmen. In fact, such regulations would actually harm consumers, increase broadband prices, retard the heretofore-explosive growth of bandwidth, and dramatically increase government control over the Internet. Of course, the degree of public interest in a cause doesn’t actually tell us anything about its justice and, fortunately, we live in a democratic oligarchic republic, not a pure democracy. But it’s worth asking whether Americans are really up in arms about the need for “Net Neutrality” regulations. Google Trends suggests not:

This kind of comparison should dispel once and for all the myth of a popular groundswell for net neutrality regulation—especially since online search volumes heavily over-represent the interests of the digerati, thus over-stating general interest in web-related topics.

In fact, “Net Neutrality” regulation is a niche cause trumpeted incessantly by the blogosphere with about the same level of broad popular interest online as “housing rights”—a topic about which most of us probably don’t often fall into conversation (unless we happen to live in Bakuninist Berkeley or the Bolivarian Caliphate of Cambridge, MA, ground-zero of American Chavismo). “Net neutrality” currently seems to attract about the same level of interest as the term “end the Fed,” the title of Rep. Ron Paul’s call for abolishing America’s central bank—something I’ve been ranting about for years but which, until recently, most people found about as bizarre and irrelevant as my (sincere) insistence that President Jefferson should have obtained a constitutional amendment rather than simply assuming the power to execute the Louisiana Purchase.

So just how much do Americans care about Net Neutrality? About 83% as much as they care about “kibble,” which usually refers to the ground meat used in dog food and other forms of animal feed—but about fifty times less than about “dog food.”

Finally, since we all write a lot about privacy, “online privacy” gets 1% as many searches as “privacy.”  “Internet privacy” and “privacy Internet” each get about 4% as many searches as “privacy,” for a total of about 9% as many searches as “privacy” or about three times as many searches as “net neutrality.”  Americans seem to be far more concerned about “identity theft,” which gets 30% as many searches as “privacy”—or 3.33 times more than the three online-privacy terms mentioned above. This is consistent with Tom Leonard and Paul Rubin’s findings that identity theft, not online data collection for advertising purposes, is the real harm facing consumers, and regulating online data collection and use in the name of “protecting privacy” isn’t likely to benefit consumers, while the costs to consumers from such regulations are likely to be significant, as Adam Thierer and I have noted here, here, here, here and here.

Make sure to read George Ou’s two recent articles over at the Digital Society blog setting the record straight about broadband usage caps: “Putting American Bandwidth Caps into Context” and “We Need to be Reasonable about Broadband Usage Caps.”   George is one sharp cookie. I particularly like the way he takes apart Free Press for their hypocrisy on this issue, something I have commented on here before after George brought it to my attention. See:

• Free Press Hypocrisy over Metering & Internet Price Controls; and
• The (Un)Free Press Calls for Internet Price Controls: “The Broadband Internet Fairness Act”

… and here’s some older material on the issue…

• Why Not Meter?
• Once Again, Why Not Meter Broadband Pipes?
• Bandwidth Cap Hysteria & the Alternative
• Bandwidth Cap Worries (by Tim Lee)
• Net Neutrality Rules = Price Controls (by Ryan Radia)
• The Real Net Neutrality Debate: Pricing Flexibility Versus Pricing Regulation

In several of our previous podcasts (see episodes 34, 35,and 37), we’ve discussed what we’ve called the “Comcast Kerfuffle,” which was the controversy surrounding the steps Comcast took to manage BitTorrent traffic on its networks. Critics called it a violation of Net neutrality principles while Comcast and others called it sensible network management.

This week we saw a new kerfuffle of sorts develop over the revelation in a Monday front-page Wall Street Journal story that Google had approached major cable and phone companies and supposedly proposed to create a fast lane for its own content. What exactly is it that Google is proposing, and does it mean – as the Wall Street Journal and some others have suggested – that Google is somehow going back on their support for Net neutrality principles and regulation? More importantly, what does it all mean for the future of the Internet, network management, and consumers. That’s what we discussed on the TLF’s latest “Tech Policy Weekly” podcast.

Today’s 30-minute discussion featured two of our regular contributors at the TLF, who both wrote about this issue multiple times this week. Cord Blomquist of the Competitive Enterprise Institute wrote about the issue here and here, and Bret Swanson of the Progress & Freedom Foundation wrote about it here and here.  To help us wade through some of the more technical networking issues in play, we were also joined on the podcast by Richard Bennett, a computer scientist and network engineer guru who blogs at Broadband Politics as well as Circle ID and he also pens occasional columns for The Register.  Also appearing on the show was Adam Marcus, Research Fellow & Senior Technologist at PFF, who wrote a “nuts and bolts” essay full of excellent technical background on edge caching and net neutrality.

You can download the MP3 file here, or use the online player below to start listening to the show right now.

[display_podcast]

The venerable Economist magazine has made a hash of my research on the growth of the Internet, which examines the rich media technologies now flooding onto the Web and projects Internet traffic over the coming decade. This “exaflood” of new applications and services represents a bounty of new entertainment, education, and business applications that can drive productivity and economic growth across all our industries and the world economy.

But somehow,  The Economist was convinced that my research represents some “gloomy prophesy,” that I am “doom-mongering” about an Internet “overload” that could “crash” the Internet. Where does The Economist find any evidence for these silly charges?

In a series of reports, articles (here and here), and presentations around the globe — and in a long, detailed, nuanced, very pleasant interview with The Economist, in which I thought the reporter grasped the key points — I have consistently said the exaflood is an opportunity, an embarrassment of riches.

I’ve also said it will take a lot of investment in networks (both wired and wireless), data centers, and other cloud infrastructure to both drive and accommodate this exaflood. Some have questioned this rather mundane statement, but for the life of me I can’t figure out why they deny building this amazingly powerful global Internet might cost a good bit of money.

One critic of mine has said he thinks we might need to spend $5-10 billion on new Net infrastructure over the next five years. What? We already spend some $70 billion a year on all communications infrastructure in the U.S. with an ever greater portion of that going toward what we might consider the Net. Google invests more than $3 billion a year in its cloud infrastructure, Verizon is building a $25-billion fiber-to-the-home network, and AT&T is investing another $10 billion, just for starters. Over the last 10 years, the cable TV companies invested some $120 billion. And Microsoft just yesterday said its new cloud computing infrastructure will consist of 20 new “super data centers,” at $1 billion a piece.

I’m glad  The Economist quoted my line that “today’s networks are not remotely prepared to handle this exaflood.” Which is absolutely, unambiguously, uncontroversially true. Can you get all the HD video you want over your broadband connection today? Do all your remote applications work as fast as you’d like? Is your mobile phone and Wi-Fi access as widespread and flawless as you’d like? Do videos or applications always work instantly, without ever a hint of buffer or delay? Are today’s metro switches prepared for a jump from voice-over-IP to widespread high-resolution video conferencing? No, not even close.

But as we add capacity and robustness to many of these access networks, usage and traffic will surge, and the bottlenecks will shift to other parts of the Net. Core, edge, metro, access, data center — the architecture of the Net is ever-changing, with technologies and upgrades and investment happening in different spots at varying pace. This is not a debate about whether the Internet will “crash.” It’s a discussion about how the Net will evolve and grow, about what its capabilities and architecture will be, and about how much it will cost and how we will govern it, but mostly about how much it will yield in new innovation and economic growth.

The Economist and the myriad bloggers, who everyday try to kill some phantom catastrophe theory I do not recognize, are engaging in the old and very tedious practice of setting up digital straw men, which they then heroically strike down with a bold punch of the delete button. Ignoring the real issues and the real debate doesn’t take much effort, nor much thought.

Last week I discussed Barbara Esbin’s new PFF paper about the FCC’s absurd investigation into how the cable industry is transitioning analog customers over to digital. This is an essential transition is the cable industry is going to free up bandwidth to compete against telco-provided fiber offerings in the future. The faster the cable industry can migrate its old analog TV customers over to the digital platform, the more bandwidth they can re-deploy for high-speed Net access and services. Mark Cuban helps put things in perspective:

1. the only thing that cable companies, and satellite for that matter have to sell is bandwidth and the applications they can run on that bandwith. More bandwidth means more digital everything. 2. For Basic Cable subscribers that get say, 40 analog channels, they are consuming 40 x 38.6mbs or 1.54 Gbs. Let that sink in. 1.54 Gbs of bandwidth. Compare that to how fast your internet access is. That more bandwidth than your entire neighborhood consumes online, by a lot. Thats also the equivalent of 500 standard def digital channels. If you convert that to revenue per bit for cable companies, or cost per bit for basic cable consumers, the basic cable customers are getting the best deal in town. By a long shot. Digital cable customers, not so much. Digital customers are paying multiples of analog customers for bandwidth. In reality, analog customers are getting an amazing deal, and the cable companies have been hesitant to convert them only because of the potential FCC backlash. I’m as cynical as the next guy when it comes to cable rates and motivations, but the reality is that the longer analog remains, the fewer opportunities to leverage the freed up bandwidth to create next generation bandwidth hog applications. Will the cable companies charge us an a lot for that bandwidth, probably. But when we start to see applications built on top of 250mbs per second and more, it will have far more value to society than watching USA Network on your old analog TV. And Net Neutrality?  Well if everyone had that 1.54gbs available to them, net neutrality would be a non issue. We wouldn’t be arguing about access or pre-emption, we would be arguing about quality of service.

Once again we are reminded that all regulations have opportunity costs and in this case the FCC’s actions could cost consumers the loss (or at least delay) of higher-speed broadband offerings in the near-term.

A new report from TeleGeography finds that bandwidth prices for backbone transit continue to decline rapidly across the globe. In San Francisco, for instance, the price per mbps of Gigabit Ethernet transit has dropped 38 % in the past 12 months. Developing countries are also enjoying substantial price cuts in 15 to 20% range.

But if the Internet’s core is controlled by an oligopolistic cartel—as Tim Wu argued in a recent New York Times essay—then why does bandwidth keep getting cheaper?

Perhaps it’s because the fourteen or so firms which offer backbone IP transit are competing fiercely to win over business from smaller carriers and enterprises. And as businesses of all sizes demand faster connectivity, more dark fiber is being lit, creating an expansion in network capacity. In DC, for instance, a price war has made high-speed commercial data services much more affordable, with one communications provider offering converged 10mbps full-duplex dedicated Ethernet over copper for less than the market price of four bonded T1 lines.

Why are some providers moving towards data transfer caps if bandwidth prices are dropping ? In part, it’s because backbone transit is not the only usage-variable expense that residential ISPs face. Last-mile bandwidth remains a highly contested resource in many neighborhoods, and the cost per megabit of bringing faster speeds to the doorstep far exceeds the cost of adding more wavelengths to a long-distance fiber optic line. As consumers demand greater speeds, providers are investing heavily in network upgrades—these costs are adding up, and there’s a strong case to be made that heavy users ought to shoulder a larger portion of the burden than light users.

Web Pro News’ Jason Lee Miller seems to think he’s hoisted my colleague Bret Swanson, and The Progress & Freedom Foundation in general, on our own collective  petard.  Bret had responded to Tim Wu’s NYT op-ed by questioning Wu’s argument for developing “alternative supplies of bandwidth” to free us from the tyranny of the OPEC-like broadband cartel:

Unlike natural resources such as oil, which, while abundant, are at some point finite, bandwidth is potentially infinite. The miraculous microcosmic spectrum reuse capabilities of optical fiber and even wireless radiation improve at a rate far faster than any of our macrocosmic machines and minerals. It is far more efficient to move electrons than atoms, and yet more efficient to move photons. Left unfettered, these technologies will continue delivering bandwidth abundance.

Miller suggests that this response to Wu destroys arguments Bret and others at PFF have made against net neutrality regulation–a crusade led by Wu (who taught me Internet law, as it happens):

So what [Swanson is] saying is bandwidth scarcity is a notion invented by internet service providers and wireless providers to jack up prices and provide excuses for interfering with competing services on their networks. Nice. In a weird way, Swanson focuses so hard on disproving Wu’s analogy one way, he misses how the analogy is proved in another: a few organizations (government or not) controlling an important resource and forcing artificial scarcity in order to control the market for that resource is called a cartel.

Miller’s “Gotcha!” rests on the seemingly undeniable premise that broadband can’t be both abundant (as Bret argues) and scarce (such that ISPs must management traffic on their networks, however non-neutral that may be).   But in fact, this seeming contradiction is inherent in the very nature of the Internet–and the way Internet access is currently priced.

On the one hand, Bret is right that broadband is “abundant” in a way that resources in the real world cannot be:  Continued investments in broadband networks by network operators have dramatically increased the amount of bandwidth available–causing prices to plummet for both wireline and wireless broadband.  Consumers today enjoy greater download speeds while paying constantly decreasing prices per bit.  So much for Wu’s OPEC analogy.

But contrary to those defenders of net neutrality regulation who think we can somehow grow our way out of the problem of network congestion merely by increasing the amount of bandwidth available, the demand for bandwidth is also infinitely elastic.  Making more bandwidth available simply encourages the development of new services and content whose use and consumption requires more bandwidth.  The significant advances in bandwidth available to U.S. broadband consumers in recent years have made it possible for us all to share huge amounts of data through peer-to-peer file-sharing services, view essentially infinite amounts of video back up hundreds of gigabytes on online storage services like Amazon’s reasonably-priced S3, and even begin moving our most basic computing tools like email and word processing into the “cloud.”  One has only to contemplate the kind of bandwidth that will be required when YouTube goes hi-def (something the less-popular Vimeo has already done) to realize that, from the network operator’s perspective, trying to solve network congestion problems simply by increasing the amount of bandwidth available is like a pie-eating contest where the prize is… more pie.

Thus, broadband can be increasingly “abundant” in the sense that there is always more of it available than ever before and, in the narrow and particular sense of Internet network congestion, “scarce” ( i.e., unlimited) at the same time.  This apparent contradiction stems from three facts:

1. Internet content and services are increasingly free to the user, either supported by advertising revenues or by some “up-sell” of additional features beyond the basic, free version.  This means that consumers have no economic reason not to gobble up the “new, new thing”–which usually consumes more bandwidth than whatever content or service it replaces.
2. Similarly, and more importantly, data use is priced on an all-you-can-eat basis.  Consumers pay a flat monthly fee for essentially “unlimited” broadband.
3. The secret to the Internet’s efficiency lies in its architecture as a packet-switching network of networks:  Unlike the circuit-switched traditional telephone network, the Internet works precisely because only a small fraction of its users are sending or requesting bits at any particular moment.  If everyone tried to watch a hi-def video at once (or watch any video, for that matter), the Internet would simply crash to a screeching halt.  Thus, even “abundant” bandwidth is necessarily scarce in terms of how many people can try to use it at any particular moment for a particular application.  Yes, it’s conceivable that in some future world with orders of magnitude more bandwidth than exists today, every person could indeed watch a classic YouTube video from 2008, but if they all tried to host holographic conference calls…  the same basic limitation would apply.

Thus it is that a tiny number of network users can consumer the vast majority of its bandwidth.  Since fact #1 is essentially a law of the Internet universe, and since no amount of additional bandwidth will overcome the constraints inherent in fact #3, ISPs must find some way of dealing with the problem of network congestion if they are to satisfy the vast majority of their customers whose network use is degraded by those who use more bandwidth than they do.  Proponents of network neutrality regulation like Wu would limit the ability of ISPs to deal with this problem through traffic management by putting government bureaucrats in charge of deciding which forms of management are benign and which are not.  (On this very day, the FCC is about to hold Comcast in violation of a non-binding 2005 policy statement for throttling, but not blocking, certain bandwidth-hogging users of the peer-to-peer file-sharing system BitTorrent.)

While some amount of traffic management will always be necessary, the need for it can certainly be reduced by changing fact #2:  moving to a different pricing structure for Internet access.  As my PFF colleague Adam Thierer has explained,

a “Ramsey two-part tariff” … would involve a flat fee for service up to a certain level and then a per-unit / metered fee over a certain level. I don’t know where the demarcation should be in terms of where the flat rate ends and the metering begins; that’s for market experimentation to sort out. But the clear advantage of this solution is that it preserves flat-rate, all-you-can-eat pricing for casual to moderate bandwidth users and only resorts to less popular metering pricing strategies when the usage is “excessive,” however that is defined.

Experiments in this area are indeed underway (and see further discussion here).  Their ultimate success will likely depend on setting the all-you-can-eat threshold high enough that ordinary users (say, 90-95% of all users) are not affected.

In the meantime, those of us who defend the accelerating abundance of broadband as the result of ongoing investments by network operators while opposing net neutrality regulation as an impediment to such investments clearly have our work cut out for us in explaining the apparent contradiction of “scarcity”-in-abundance that is unique to Internet bandwidth.