Response to Harold Furchtgott-Roth.

• Author: Crawford, Susan
• Position: In this issue, p. 312

TABLE OF CONTENTS I. INTRODUCTION II. MARKET DEFINITION III. MARKET POWER IN WIRELESS IV. CABLE'S PROFITS FOR ITS DATA SERVICE V. INTERNATIONAL COMPARISONS VI. CONCLUSION--THE ROLE OF GOVERNMENT I. INTRODUCTION

I have great respect and affection for Harold Furchtgott-Roth, and it seems from his review of Captive Audience that he has respect and affection for me. Luckily for the rest of you, my view of his personality--and his view of mine--is irrelevant. I wanted to begin, though, by acknowledging his personal graciousness towards me.

What is relevant is the striking number of issues in his review on which he and I completely agree. We agree that U.S. presidential administrations for a long time have not thought of the FCC as an important agency or its role in the U.S. economy as vital. We agree that communications policy in the U.S. is not necessarily being made based on the merits of particular situations. We agree that the federal government would probably be terrible at running a nationwide network itself.

Where we differ is in our prescriptions, given this agreed-on background. Mr. Furchtgott-Roth's conclusion from these premises is that the only answer is to give up. (1) And my conclusion is that we cannot give up.

My conclusion, unlike Mr. Furchtgott-Roth's, is based on the reality of consumers' experience in America when it comes to high-speed Internet access. Based on how people actually use these connections and how much they are required to pay, consumers are being gouged; the rich are paying too much for services that are both noncompetitive and second-class, and not enough Americans are being served adequately or at all.

Mr. Furchtgott-Roth would have to agree with me, because it is a central tenet of his worldview, that competition is central to any functioning free market. My contribution to this conversation is that we have neither competition nor adequate oversight when it comes to the actual lived experience of Americans as purchasers of high-speed Internet access. As a result, the country needs a wholesale revision of both the manner in which policy is made and the details of those policies so that retail-level competition can in fact be unleashed in the places where it is possible. We need new investments in modern, competitive, wholesale fiber networks in cities across the country so that new retail providers can begin selling services. (Fiber networks can last for decades if they are the right quality, and photonics are getting better all the time.) The entire nation needs an upgrade, and new entrants into the marketplace--both public and private--are needed to change the status quo that is serving all of us so badly.

If we fail, and if the FCC cannot act because the agency fears that the incumbents will march on Capitol Hill and gut the agency's budget if they do, that will have major implications for our democracy. Nonetheless, we should certainly still try.

We will need leadership from every level of government; mayors, governors, and the President will need to take the current crisis in American communications networks seriously. I wrote Captive Audience to encourage these leaders to step up to this challenge and to encourage every American to make his or her voice heard on this fundamental issue when electing representatives to office. (2) World-class, high-speed Internet access should be available to each one of us at a reasonable price, and where it is possible to have retail choices, we should have competition. This issue isn't just about equality or dignity; it's also about economic growth.

2. MARKET DEFINITION

   Mr. Furchtgott-Roth says that we have a competitive "broadband" marketplace. (3) He gets there by saying that 4 Mbps down and 1 Mbps up is the right definition of that marketplace. (4) But that market definition is entirely unrelated to reality. Let's step back and look at what people in America actually do using high-speed Internet access capacity. What are typical, mainstream applications that people are using, and what download speeds and bit capacity (or volume--number of bits per month) do they need for such use?

   Here are the facts: Sandvine reports that Netflix is currently the largest source of online traffic in this country. (5) Netflix itself recommends that people accessing its application have 3.0 Mbps download service in order to get "DVD-quality" resolution. (6_ And for a HD-quality (high-definition) picture, Netflix recommends at least a 5 Mbps download service. (7)

   Let's say a typical American wants to watch one hour of Netflix in HD in one day. That is a conservative, sensible assumption. According to the FCC's most recent numbers (for the first half of 2012), most current DSL connections cannot handle this usage. At least 76% of DSL connections are slower than 6 Mbps downstream. (8) DSL fares reasonably well in slower-speed ranges. For download speeds of at least 768 Kbps and 1.5 Mbps, DSL has 33.7% and 31.0%, respectively, of total fixed connections. (9) But when we consider the speeds required by even moderate users of high-bandwidth communications, like Netflix, the picture changes dramatically: DSL's share of fixed connections falls to just 19.7% for download speeds of at least 3 Mbps (the minimum recommended speed to stream just a single DVD-quality Netflix movie), and to just 8.3% of fixed connections at the 6 Mbps threshold (as noted above, 5 Mbps is the minimum recommended speed to stream a single HD-quality Netflix movie). (10)

   Because American households often include more than one person, and people themselves often do more than one thing at a time, it is quite likely that more than one device is being used to access high-bandwidth communications at any given time in Americans' houses. In fact, a growing proportion of American households own multiple devices capable of supporting high-bandwidth communications, and the number of devices per household is increasing. (11) But DSL's shortcomings are even starker when it comes to connections capable of delivering at least 10 Mbps of downstream capacity. For connections of at least 10 Mbps, DSL's share as of mid-2012 was just 2.9% of fixed connections--less than one out of every thirty connections. (12)

   Speeds of at least 10 Mbps are, therefore, the threshold at which an Internet connection can become a true replacement for a pay-TV service. At 10 Mbps, a connection can carry two Netflix HD streams, or three different streams of DVD-quality. Or, a household can support one Netflix HD stream while another screen is being used to surf the Web.

   This explains why Americans signing up for new high-speed Internet access services are not choosing DSL. (13) What Americans are signing up for is their local cable monopoly service. (14) For high-capacity download speeds that are at least 6 Mbps and, more likely, above 10, 15, or even 25 Mbps, the vast majority of Americans have just one choice--their local cable incumbent. (15) According to Leichtman Research Group, during the first nine months of 2012, 88% of new high-speed Internet access subscriptions went to cable, a figure that exceeded 99% during the third quarter of 2012. (16) For 2012 as a whole, less than one out of every eight new high-speed Internet access subscriptions went to anyone other than the cable companies. (17) Even these figures understate the steep decline in users choosing DSL connections because they include not only DSL but also phone companies' fiber services, including Verizon's FiOS Fiber to the Home ("FTTH") service and AT&T's Fiber to the Node ("FTTN") U-Verse service. Focusing only on the DSL customer base of the two big phone companies, AT&T and Verizon, we see it shrinking nearly 23% for AT&T during 2012 and 12.5% for Verizon. (18) During that same period, the nation's top two cable distributors, Comcast and Time Warner Cable, increased their cable modem subscriber base by 6.7% (19) and 9.9%, (20) respectively. Verizon's FiOS fiber-optic service will be available when fully rolled out to about 10-15% of Americans (18 million people), (21) but Verizon's FiOS overlaps just 15% of Comcast's footprint and just 11% of Time Warner Cable's footprint. (22) These two cable actors face almost no competition for the download speeds that Americans want. And, of course, they never compete with each other. When it comes to connections that allow 10 Mbps or more in downloads and 200 gigabytes of data--a measure of capacity, or volume--per month, we are heading with increasing speed toward a series of regional cable monopolies in wireline high-speed Internet access service provision.

   Many ill-informed or possibly simply confused people are asserting that 4G wireless is a substitute method for access to the download speeds I have been writing about, but they are missing the reality of the situation. (23)

   Although it is true that both Verizon and AT&T claim that their 4G LTE services will have download speeds equal to a slow cable connection--say, 13-16 Mbps--these two companies also impose monthly data caps of 1-2 GB of data. (24) If you watch a movie in HD over a wireless connection, you're going to use about 3.5 GB of data, blowing through your monthly capacity allowance for a single-device plan costing $85-$100 per month in a single session. (25) These two dominant companies charge as much as $15 per GB of data for those who exceed the allotted usage cap. (26) This means that watching one movie would cost an additional $17.50-$35 (or $52.50 for those exceeding their monthly cap). And these amounts do not include the cost of the movie itself. In light of these usage caps, it does not seem realistic to claim that consumers see wireless connections as substitutes for a wire at home. (27) And, in fact, they do not: Mobile usage of Netflix is in the low single digits as a percentage of overall Netflix use. (28)

   Another inconvenient fact for people who claim that wireless high-speed Internet access is entirely substitutable for...