As the Internet becomes more widely used by consumers, more households want a fast connection to it (usually called broadband) at home. Although that use of high-speed Internet access has grown as fast as the use of any recent consumer good has, many observers argue that the broadband market is not functioning properly, either for reasons internal to it or because of government regulation. This Congressional Budget Office (CBO) paper--prepared at the request of the Senate Budget Committee--reviews the recent trends in the market for residential fast Internet access to determine whether there are impediments to the market's growth. In keeping with CBO's mandate to provide objective, impartial analysis, this report makes no recommendations.

Philip Webre of CBO's Microeconomic and Financial Studies Division prepared the paper under the supervision of Roger Hitchner and David Moore. The paper benefited from the comments of Steven Wildman, Andrew Odylzko, Harold Furchtgott-Roth, Gerald Faulhaber, John Berresford, and Coleman Bazelon. Within CBO, Tom Woodward, Mark Lasky, and Kathy Gramp contributed helpful suggestions.

Leah Mazade edited the manuscript, and Christine Bogusz proofread it. Maureen Costantino designed the cover. Angela Z. McCollough prepared drafts of the paper, Lenny Skutnik produced the printed copies, and Annette Kalicki prepared the electronic versions for CBO's Web site.

Douglas Holtz-Eakin
Director
December 2003

Households and small businesses have adopted high-speed Internet access, also known as broadband, at a rapid rate over the past three years. Data published by the Federal Communications Commission (FCC) show that as of December 2002, 17.4 million households and small businesses subscribed to high-speed lines for Internet access, up from only 1.8 million in December 1999 (see Summary Table 1). A regional view of the data also indicates rapid growth. In 1999, more than 40 percent of the zip codes in the United States did not have even a single subscriber to a broadband service, according to the FCC. At the end of 2002, only 12 percent of zip codes had no such subscribers (see Summary Figure 1). Indeed, broadband's rapid rate of growth is rivaled only by that of the most successful of recently introduced consumer electronics products, such as cellular telephones.

Summary Table 1.
High-Speed Internet Access Lines in Residences and Small Businesses
(Thousands)
		Dec. 1999		June 2000		Dec. 2000		June 2001		Dec. 2001		June 2002		Dec. 2002
Asymmetric Digital Subscriber Linesa		292		772		1,595		2,491		3,616		4,395		5,529
Other Wired Technologies		47		112		176		138		140		224		213
Cable Modem Services		1,402		2,215		3,295		4,999		7,051		9,157		11,343
Fiber-Optic Cable		1		*		2		3		4		6		15
Satellite and Other Wireless Technologies		50		64		102		182		195		202		257
	Total	1,792		3,164		5,170		7,812		11,005		13,984		17,357
Source: Congressional Budget Office based on Federal Communications Commission, High-Speed Services for Internet Access: Status as of December 31, 2002 (June 2003), Table 3.
Note: * = rounds to zero.
a. With asymmetric lines, the speeds at which customers can receive data are by design much greater than the speeds at which they can send information.

 

Summary Figure 1.

Percentage of Zip Codes Nationwide with High-Speed Internet Access Subscribers, by Number of Providers

Source: Congressional Budget Office using data from Federal Communications Commission, High-Speed Services for Internet Access (various years).

Note: The data are for lines that provide high-speed Internet access to residences and small businesses.

Yet some observers maintain that the market is expanding too slowly. They contend that flaws in the market's structure impede even more rapid growth and the benefits it would bring to the economy, and they argue that the federal government should intervene--by providing financial incentives--to encourage such growth. Advocates of federal action have thus introduced two dozen bills to accelerate the deployment of high-speed Internet access--in many cases, by means that would increase federal spending or decrease revenues. This Congressional Budget Office (CBO) report analyzes the development of the residential broadband market (a category that covers both households and small businesses) to assess whether structural features or regulatory obstacles (or both) impede its further rapid growth.

The local networks that supply telephone and cable television services also provide most of the high-speed access to the Internet that is purchased by small businesses and households. Among markets nationwide, the combined share of broadband service provided by local telephone and cable companies averages more than 90 percent. That feature of the supply side of the residential market for high-speed Internet access raises the possibility that the two dominant firms in each individual market might be able to keep prices above the cost of providing service. If that occurred, too few people might subscribe to a broadband service at too high a price relative to the prices that would prevail in a more competitive market--a situation known as a market failure. Many advocates of federal intervention maintain that that is what has happened. Claims of a failure on the demand side of the market are more diverse and in general are less important for the market's operation.

CBO's analysis concludes, however, that nothing in the performance of the residential broadband market suggests that federal subsidies for it will produce large economic gains. That finding is relevant in addressing claims about broadband's future contribution to the economy and the support that some advocates of federal intervention draw from those claims. Specifically, if the market is functioning adequately, economic gains from interventions to increase the number of subscribers are likely to be offset by losses in other markets, leaving the economy as a whole worse off.

Economic efficiency (allocating resources to the uses that will most enhance society's well-being) is not the only reason that federal policymakers might decide to subsidize telecommunications services. Concerns about equity and regional economic development may also enter into their considerations. However, those issues are beyond the scope of this analysis, which addresses only the question of economic efficiency.
 

Overview of the Residential Broadband Market

Small businesses and households can connect to the Internet through broadband or dial-up connections.⁽¹⁾ Broadband access to the Internet is distinguished from dial-up access provided by a conventional telephone line on two dimensions: speed and availability. Broadband access allows subscribers to connect (with either outgoing or incoming transmissions--usually incoming) at a speed of at least 200,000 bits of information per second. By contrast, dial-up services typically operate at less than 56,000 bits per second. Broadband's performance also exceeds that of dial-up access in that, typically, it is always "on" (dedicated and immediately available).

As noted earlier, local telephone and cable television companies currently dominate the residential broadband market. In December 2002, high-speed Internet access through a cable modem accounted for 65 percent of all broadband subscriptions carried by households and small businesses, and digital subscriber line (DSL) service accounted for 32 percent. Telephone companies deliver DSL service through their existing wired telephone networks using a special modem, which transmits the computer signals at frequencies above the range of human hearing and thus does not interfere with voice telephone service. Cable television companies deliver their broadband service by using an unused channel on their cable network and a special modem.
 

The Supply Side: Competition Now and in the Future

Currently, the major cable and telephone companies maintain their large share of the broadband market at the expense of the growth of their competitors' share (mainly satellite companies and competitive providers that use the telephone companies' wired networks; see Summary Figure 2). Both the incumbent local telephone company and the local cable franchise deliver their broadband service over extensive and costly wired networks that were originally designed and built for a different purpose. Those networks create a particular structure of costs. Delivering any service is expensive to begin with because the network has to be built. But once it is in place, service can be delivered at a relatively lower cost. That combination of high startup and low operating costs makes it particularly difficult for would-be competitors to enter the market. As a consequence, some observers are concerned that telephone and cable companies will continue to increase their share of the market and keep prices above the level they would reach if the market was competitive. In that circumstance, too few people would buy broadband services and those who did would pay too much for them.
 

Summary Figure 2.

Market Share of Competitive High-Speed Internet Access Providers

Source: Congressional Budget Office based on Federal Communications Commission, High-Speed Services for Internet Access (various years).

Note: Competitive providers (for example, satellite Internet suppliers) are providers other than the local cable and telephone companies.

Two developments could prevent that outcome. First, new competitors could enter the market. Second, competition in the larger telecommunications markets could force the telephone and cable companies to behave competitively in the broadband market, even without a larger number of competitors in that market.

New Entrants into the Residential Broadband Market

New competitors could enter the residential broadband market by using either wired or wireless technologies to connect their customers to the Internet. Much of the existing competition in the market comes from firms that use wired technologies. Until 2003, the FCC required incumbent local telephone companies to lease frequencies on their telephone lines to competitive companies that wanted to provide DSL service. In February 2003, the commission announced the start of a three-year transition to gradually eliminate those leasing requirements. The final outcome of that transition is as yet unknown because the FCC's decision is being challenged in the courts and the commission had planned in any event to revisit its decision in three years. However, even before the regulatory change, some competitive DSL providers went bankrupt because of the slowdown in investment in telecommunications equipment beginning in 2001. The remaining competitive DSL firms face uncertain prospects, and their share of the market is quite small.

The most prominent technological alternatives in the category of wired high-speed Internet access are new fiber-optic cable networks and the use of existing electric power lines. Despite widespread interest in fiber optics, the latest estimates by the FCC suggest that only 15,000 households and small businesses use it to connect directly to the Internet, and few analysts anticipate rapid deployment of fiber networks in the near future. Moreover, the most likely agents of fiber deployment are telephone companies, which means that such a development would add capacity to the market but not a new competitor. Power lines that transmit electricity are already connected to most residences and small businesses and thus appear to offer a potential broadband connection. In April 2003, the FCC began a major study of that form of broadband communications. However, whether the use of power lines to provide high-speed Internet access is technically viable is unclear. Also uncertain is how the cost of that alternative compares with that of cable modem or DSL service.

Wireless technologies do not at present offer much competition to the local telephone and cable companies in the broadband market; in 2002, they accounted for less than 2 percent of residential broadband connections (see Summary Table 1). The FCC continues to make additional portions of the electromagnetic spectrum available for use, and private firms regularly announce new initiatives that use wireless data technology. To date, however, wireless broadband alternatives--including the use of satellite systems--serve only niche markets and limited geographic areas. The latest wireless broadband technologies are designed to overcome many of the problems that constrained earlier wireless systems, but the industry's history of past high expectations and subsequent limited success or even failure suggests that a wait-and-see attitude might be appropriate.

One popular new wireless technology, Wi-Fi, is successful within its rather narrow domain. Designed for connecting local computer networks to large numbers of users without the need to run costly wires, Wi-Fi has since been adopted by consumers (largely as a way of sharing their single DSL or cable modem connection among multiple home computers) and by institutions with large existing facilities, such as universities or hotels (as a way of providing broadband at low cost to peripatetic users). However, few firms have figured out how to profit directly from providing Wi-Fi service (although manufacturers of Wi-Fi hardware are profiting from its popularity). Furthermore, a Wi-Fi base station ultimately requires its own wired or wireless link to connect to the Internet.

Competition in Related Markets

Limited prospects for growth and competitive threats in their core business markets could spur local telephone and cable companies to compete more vigorously in the broadband market. As of December 2002, competitive providers in the telephone service market--which in many cases include the same cable television companies that compete with the phone companies in the broadband market--accounted for 10 percent of all residential telephone lines and 13 percent of all telephone lines nationwide. New competitors have also entered the core markets of cable television providers. From June 1998 to June 2002, the number of households that received multichannel video programming from a satellite service doubled, rising from 9.2 million to 18.7 million (or 18 percent of all households with televisions). By contrast, subscriptions to cable video service increased by only 5 percent during that period. To defend their core markets, both local telephone and cable television companies have begun to lower their prices for customers who are willing to buy a bundle of services--for example, telephone, cable television, and broadband access--from the same provider.

Supply and the Performance of the Market

Available indicators of the broadband market's performance are mixed. Price competition is not vigorous, at least not by the standard currently set in the market for cellular telephone service. (Like the broadband market today, it was at one time limited to only two providers but now has four to six competitors in every major market in the country.) Systematic data on broadband prices are sparse; however, existing information suggests that although DSL service providers have dropped their prices slightly (once rebates and other special offers are taken into account), cable providers have raised theirs to match the initially higher DSL rates. Recently, some of the largest DSL providers, including Verizon and SBC, lowered their prices to levels substantially below those of several of their major cable competitors. That move has not yet drawn widespread response from the cable companies.

Data specifically on the broadband investments of the two dominant firms are not available, but the high overall rates of spending by telephone and cable providers along with the rapid increase in the geographic availability of broadband suggest that both types of firms are investing in new broadband capacity. Indeed, between 1996 and 2001, the four largest local telephone companies (companies that were part of the old Bell system) increased their spending on new structures and equipment by 64 percent (see Summary Table 2). Cable companies also raised their investment spending substantially, increasing it by 68 percent over five years. Investments of that magnitude indicate that, thus far, the negative consequences of the market's domination by the local telephone and cable companies have been limited.

Summary Table 2.
Investment by Regional Bell Operating Companies
(Billions of dollars)
			1996		1997		1998		1999		2000		2001
Investment
	Structures and equipment		19.7		20.0		20.2		27.5		35.7		32.4
	Research and development		0.2		0.2		0.1		0		n.a.		n.a.
		Total	19.9		20.2		20.3		27.5		35.7		32.4
Memorandum:
Operating Revenues			78.7		80.4		86.0		90.4		109.2		108.7
Total Investment as a Percentage of Operating Revenues			25.3		25.1		23.6		30.4		32.7		29.9
Source: Congressional Budget Office using data from Federal Communications Commission, Statistics of Communications Common Carriers (various years).
Note: n.a. = not available.

Some observers contend that both the regulations that have required telephone companies (but not cable companies) to open their networks to broadband competitors and uncertainty about whether those regulations will be in effect in the future have discouraged even higher levels of investment. The FCC's recent decision to suspend the requirement that telephone companies share their wired networks with competitive broadband providers, although subject to litigation and state regulatory interpretation, may reduce whatever restraint that rule might have had on telephone companies' investment in the broadband market. Nevertheless, evidence of a rapid rise in investment by the telephone companies during the period in which those regulations were in effect suggests that the deterrent to such spending that the regulations provided has been limited.
 

The Demand Side: Public Goods and Constraints on Content

The claims of problems on the demand side of the broadband market derive from underlying arguments about public goods and constraints on the content that is available for downloading over the Internet. Public goods are those goods that once they have been provided can be consumed by an additional person at no cost and for which it is inefficient to exclude any one consumer from receiving them. (National defense is the most frequently cited example of a public good.) Economists generally believe that private markets will supply too small an amount of a public good if they are left to their own workings.

Communications networks have attributes of a public good because an additional subscriber to a network provides a benefit to the existing subscribers--an additional person to communicate with--that is available to all of the network's members at little or no additional cost. Because a would-be subscriber fails to account for that "network effect" in calculating the benefits of subscribing, too few people will subscribe compared with the number that will maximize economic well-being. In the case of high-speed access to the Internet, however, the availability of traditional dial-up access through telephone lines, a slower but still relatively effective alternative, substantially reduces the extent of losses in economic well-being that might be attributable to network effects.

The observation that broadband access is an input that can be used to produce certain public services is the basis for the claim of a second potential failure on the demand side of the market. Services such as public health or education have some of the characteristics of a public good, and government policies promote their availability in a variety of ways, including subsidization of high-speed Internet access at so-called public portals (for example, libraries). However, the use of high-speed Internet access to disseminate those services does not suggest a market failure in that delivery mechanism, nor does it suggest that subsidies to households and small businesses are necessarily the most efficient way to provide more public goods. Existing subsidies for Internet access are also intended to address concerns about lack of access in low-income areas.

A third demand-side issue is the claim that constraints on the entertainment-oriented content that is available over the Internet (for example, restrictions on the number of their products that movie and music producers make available) represent a failure in the market. That claim is not relevant because questions about efficiency and the regulation of entertainment products lie outside of the market for high-speed Internet access. For example, the desirability of a broadband connection might be increased if restrictions arising from copyright protections could be ignored, but the lessening of those protections might have negative consequences as well--for the development of new content and the efficiency of the markets that produce it.
 

Conclusions

Although the residential broadband market is not characterized by a large number of competitive providers, the forces of supply and demand seem to be working to offer an increasing number of households and small businesses a choice of high-speed Internet service from among a range of providers. The number of subscribers is growing dramatically; in short, the market is booming. Some of the problems that remain, such as uneven availability, are a function of the market's relative youth and immaturity and are not necessarily permanent features. The current domination of many markets by only two broadband providers, however, could turn out to be more long-lived.