Innovation Files has moved! For ITIF's quick takes, quips, and commentary on the latest in tech policy, go to

Live Different: Susan Crawford’s Broadband Blinkers

Horse with Blinkers
An op-ed by Susan Crawford in last weekend’s New York Times Weekend Review paints a very peculiar picture of America’s broadband networks. Crawford claims, as she has for many years, that American broadband is a dire mess that can only be fixed by the government’s enacting sweeping new regulations on the way the networking business is structured. In particular, she wants Congress to require each network owner to sell wholesale services at a price set by the FCC under a system of “structural separation” that depresses investment in next-generation networks. Her diagnosis is faulty and her snatch-and-grab analysis – motivated more by a desire to reinforce an existing point of view than to understand current developments – causes her to misconstrue major new trends in network adoption, usage, and deployment.

The polemic begins with a superficial look at on-line retailing, arguing that the poor are unable to enjoy shopping deals because they lack wired broadband connections in their homes. The fact that the poor are not notoriously big spenders doesn’t perturb the conclusion: Crawford insists that poor people’s relatively high reliance on mobile networking cuts them off from digital shopping’s nirvana, Cyber Monday.

In fact, the biggest story in retailing this holiday shopping season is the emergence of bargain-hunters’ friends on mobile broadband and social networks. According to IBM’s Smarter Commerce initiative, Black Friday shoppers are increasingly using smartphones to help them research products and make purchases. The IBM data show that mobile sales tripled this year, now accounting for ten percent of all Black Friday retail. Cyber Monday, traditionally the biggest day for on-line shopping, will soon by eclipsed by Thanksgiving Day, up an amazing 40% from last year. Shopping chat on social networks, primarily Facebook, was up 110%, as consumers shared tips on deals, out-of-stock items, parking, and checkout lines.

An IBM press release summarized the story:

“This year marked Thanksgiving‘s emergence as the first big spending day of the 2011 holiday season with a record number of consumers shifting their focus from turkey to tablets and the search for the best deals,” said John Squire, Chief Strategy Officer, IBM Smarter Commerce. “This momentum continued into Black Friday where the big winners were those retailers that delivered a smarter commerce experience with compelling, relevant deals that people could easily access from their channel of choice.”   

A more detailed report shows that the early trends carried over to Cyber Monday. People aren’t just shopping more this season; they’re shopping differently than they have in the past. Instead of waiting for Cyber Monday to shop from work, bargain-conscious shoppers are using their smartphones and tablets to get a head start while the stores are closed, and those who need to see and feel their purchases before they buy – especially clothing – are using their smartphones as they shop to check prices and find in-stock merchandise. Cyber Monday shoppers tend to use employer’s networks, which are less important as more people get connected at home and through smartphones. The rise of mobile broadband is changing the way we shop, as well as the way we work, play, entertain ourselves, and communicate.

Crawford sees mobile broadband as nothing more than a second-rate substitute for “real” broadband: “High-speed access is a superhighway for those who can afford it, while racial minorities and poorer and rural Americans must make do with a bike path.” This is wrong on several levels.

Poor people can get basic wired broadband for as little as $9.95 per month thanks to the “Connect to Compete” program created by the FCC in cooperation with major networking providers.  This program also provides low-cost PCs and computer training, and thus addresses the major barriers to wider use of broadband in American homes, low rates of computer ownership and digital literacy. Even without subsidies, wired broadband is less expensive than mobile, with typical plans starting at $15 per month for wired and $30 for mobile.

Crawford sneers at mobile’s technical advances and preys heavily on matters of race and class. She employs these classically divisive, hot-button issues to provoke favorable responses for the regulatory system she’s been promoting since the early days of the net neutrality controversy, the “structural separation” regime that segregates wholesale and retail broadband services. The logic of structural separation relies on establishing two claims:

  1.  The claim that Americans pay high prices for inferior broadband; and:
  2.  The claim that we do so because of insufficient competition in the retail broadband marketplace. 

Establishing the first claim is done by comparing some aspect of American broadband – price, speed, build-out, technology, adoption, it doesn’t really matter – to conditions in some other nation or city, preferably a newly-connected one that’s skipped a generation or two of American technology. These locales don’t have as many lower speed legacy systems in use as we do, so their statistical averages look great. Former Soviet republics such as Latvia and Romania are good for this purpose as their technology history is so short; they’ve jumped from having no broadband at all to the current state of the art with few legacy systems to depress their averages.

As more Americans get advanced vDSL+ systems capable of reaching speeds of 50 Mbps per connection, (such as AT&T U-verse,) and cable companies convert to DOCSIS 3 systems that share 160 Mbps across two to three hundred homes, it becomes harder to make the case that we’re lagging. The U. S. is certainly on an upward path in international rankings over the past few years: Akamai has us up from 22nd to 16th in average speed over the past 18 months, and in other rankings we’re higher.  To a large extent, these rankings are constructed in such a way as to measure the speed of the installed base of PCs and not simply their network connections, so countries with longer technology histories and more old computers are always going to lag smaller nations with shorter histories.

Crawford’s claim that cable networks are so vastly superior to the new vDSL+ networks operated by the telephone companies as to represent a “looming monopoly” rests on an apples-to-oranges comparison. As we’ve explained before, cable has a higher raw data rate than DSL, but cable is shared by many homes while DSL provides an exclusive connection.

Whether a connection that’s six times faster and shared with 250 others is better than one that’s not shared depends on the applications we and our neighbors are using. When Web surfing was the dominant Internet application, cable had a clear advantage; as the mix shifts to long-running video applications such as Netflix, the advantage trends in the other direction. Since most users employ a mix of applications, neither is the clear winner. This analysis is widely accepted by technologists, but Crawford, sadly, isn’t one of us.

Neither cable nor DSL is standing still, so any prediction of “looming” conditions has to honestly assess the potential for improvement in both systems. In any event, the mix between cable and DSL market shares has been roughly 55 percent cable and 45 percent DSL for a decade and shows no sign of changing. 

It’s also important for the mobile-hater’s analysis to ignore new trends in mobile broadband technology, lest it discover the inconvenient truth that the U. S. leads the world in the adoption of LTE, the next generation mobile system that’s capable of supporting speeds up to 300 Mbps. According to Akamai’s most recent “State of the Internet” report, mobile broadband peak rates are in fact competitive with wired broadband in many countries, frequently clocking in above 15 Mbps and ranging as high as 42 Mbps in Croatia. Most of us would be satisfied with that much speed, even for the couch-potato usage that Crawford touts over mobile apps.

The Cisco Visual Networking index forecasts that wireless Internet access will exceed wired access by 2015. This figure suggests that those who regard wireless as second-rate service aren’t on the same page as the American consumer.  A great deal of this wireless use is video streaming to tablets via Wi-Fi – a fairly un-demanding application that has great appeal – but the fastest growing portion of it is mobile traffic, much of it connected with new mobile apps.

The digital divide has different dynamics in the mobile app space than it does otherwise; according to the Pew Center, “African-Americans and young adults are more likely than others to download apps that help them communicate with friends and family” and “minority cell owners are significantly more likely than whites to use most non-voice data applications on their mobile devices.” This finding includes the use of smartphones to access Crawford’s Web-centric Internet. 



Image credit: Pew Center

Newer smartphones have the ability to connect laptops and tablets to the Web with their tethering and Wi-Fi hotspot applications, so such activities as filling out on-line job applications, declared impossible from smartphones by Crawford, are more accessible to mobile broadband users than she knows. Once again, the barrier to such applications is computer ownership, not broadband prices.

The bottom line is that the diffusion of broadband to more users through mobile devices is an advance in the connected lifestyle, not any kind of a detriment. If the spread of mobile broadband is troublesome to some long-held policy agendas, perhaps it’s better to take a broader view of the policy space than to hide the data that doesn’t fit the old models.

The major cable companies have just announced that they’re selling currently unused wireless spectrum to Verizon in a deal that will enable them to resell Verizon’s LTE services to their own customers (public safety should take note of this development.) This will bring cable companies into the world of quad-play services, where they will sell bundles of wired Internet access, TV and phone service, and mobile phone and data services to their customers at reduced rates. It will also provide Verizon with wholesale access to the cable networks, which will expand consumer choice.

Differentiated wholesale and retail markets naturally develop for many network services, without regulatory mandates, simply because it makes sense for companies to focus on core competencies. The virtue of such arrangements, when arrived at freely, is the rationality of pricing set by providers who can accurately value their services and reasonably assess opportunities. In the strict structural separation that Crawford favors, prices are set by bureaucrats in a haphazard and arbitrary way, and analysis shows that these regimes depress investment in advanced networking technology. These are things we’d rather see companies doing of their own volition, with appropriate government supervision, rather than under the gun of broad and heavy-handed regulatory mandates that can have chilling effects on investment.

The primary shortcoming of the old-school analysis of the modern mobile Internet is its refusal to accept changing patterns of usage and advanced technologies such as LTE as beneficial. Technical change is often disturbing, but it’s generally for the good. Crawford’s particular style of analysis is heavy on vague generalizations and light on substance. She talks about “high-speed connections” without ever specifying a rate, and refers to a number of common applications – such as filling out job applications on-line – as if they required massively reengineered networks when they clearly don’t.

Her vision of future applications is utterly pedestrian. She says: “Within a decade, patients at home will be able to speak with their doctors online and thus get access to lower-cost, higher-quality care” without acknowledging that doctors already use e-mail, can use video calling in many cases, and have been reachable by phone for several generations. When she says: “High-speed connections will also allow for distance education through real-time videoconferencing,” she at least acknowledges that “thousands of high school students are earning diplomas via virtual classrooms” today.

There’s no question that improvements in broadband networks can bring about advances in medical care, education, and in dozens of other areas; nor is there any doubt that new broadband technologies are being developed and deployed. The question that Crawford refuses to confront in a serious way is why we should kill the goose that lays these golden eggs and seems intent on producing even more.

photo credit: Alex E. Proimos

Print Friendly, PDF & Email

  • groupeintellex

    An interesting viewpoint, Richard, but one that, in the same way that Susan Crawford seems to favour a ‘fixed line’ regulatory solution, doesn’t fully recognise the interdependency of Fixed and Mobile access infrastructures.The advent of LTE has only served to make this interdependency stronger – as I noted in an op-ed back in June .… It was written for a UK readership but translates easily for other audiences.On the broader question of Crawford’s enthusiasm for structural separation in fixed line coms I would note that something similar is happening in Mobile – albeit not driven by regulators but more by the distinction between basic Access (a utility) and Services (competitive content). There are already several examples worldwide of mobile infrastructure companies providing a common wholesale platform to make better use of spectrum and reduce investment costs for Services providers.

  • Richard Bennett

    Fixed and mobile broadband share backhaul at some level, but it’s generally not copper and is hardly a utility service. Fiber emanates from Internet & Ethernet exchanges as far toward homes and offices as it needs to go in order to carry the load, but microwave is a player in that portion of the system as well, bridging the gap between cell towers and exchanges. Neither you nor Crawford address the issues of middle mile buildout as far as I can tell.It’s appealing to think about the middle mile as a “utility” but fundamentally misleading. Utilities exist outside the scope of innovation, and that’s not a characteristic of high-capacity data networks.

  • Nice article. I agree FCC must regulate cable and broadband providers but it must not be too much. And why in this era also we are having racism? The data is categorized here in terms of races. Every person in this world is equal. We must not give important to racism.