Citizens Budget Commission
TELECOMMUNICATIONS INFRASTRUCTURE
AND NEW YORK'S COMPETITIVENESS
A Background Paper by the
Citizens Budget Commission
Founded in 1932, the Citizens Budget Commission (CBC) is a nonpartisan, nonprofit civic organization devoted to influencing constructive change in the finances and services of New York State and New York City governments. This report was prepared under the auspices of the CBC's Competitiveness Committee, which we co-chair. The other members of the Committee are Jonathan Ballan, Deborah Buresh, Herman R. Charbonneau, Denis V. Curtin, Karen Daly, Morton Egol, Patricia O'Donnell Ewers, Bud H. Gibbs, William H. Hayden, H. Dale Hemmerdinger, Margaret C. Henry, William R. Howell, Peter A. Joseph, Ellen Oran Kaden, Walter T. Kicinski, Marianne E. Kozlowski, William P. Lauder, Frank J. McLoughlin, Felix A. Orbe, Laurence G. Preble, Carol Raphael, Arthur H. Rosenbloom, Heather L. Ruth, Lee S. Saltzman, Joseph V. Salzano, Richard B. Teiman, W. James Tozer, Ronald G. Weiner, Robin L. Wiessmann, and Lawrence B. Buttenwieser, ex-officio.
The Competitiveness Committee was created in early 1997 with a mandate to follow up on an important finding of the CBC's December 1996 report, Budget 2000 Project. The Commission found that New York City and New York State suffered from high tax burdens on residents and businesses; the Commission, in its report on Tax Policy, made recommendations to reform and to reduce State and local taxes to make New York more competitive. However, the CBC also recognized that an area's economic competitiveness is not determined by taxes alone. State and local governments enhance the competitiveness of their jurisdictions by delivering high quality services and providing well-functioning public infrastructure. The charge of the Competitiveness Committee is to determine New York's competitive position in terms of service delivery and infrastructure provision, as well as tax burdens, and to make recommendations for improvement.
The Committee began its work by assigning priority to three areas of infrastructure-telecommunications, ground transportation, and airports. Papers have been commissioned by experts in each area. This report addresses telecommunications infrastructure, and is the first to be released to the public.
This report was prepared by Jed Kolko of the Harvard University Department of Economics, under the supervision of Charles Brecher, Executive Vice President and Director of Research at CBC. They were assisted by the cooperation of staff from Bell Atlantic, and the Committee thanks CBC Trustee Frank McLoughlin for helping make available unpublished data and for his insights into the ways in which less than fully comparable data from other jurisdictions could be interpreted meaningfully. The report was prepared for publication by Nicolette Macdonald, Publications Coordinator. An electronic version of this report is available on the CBC's website at http://www.cbcny.org.
David Greenbaum
Hector Prud'hommeNovember 2, 1998
NATURE OF VARIATIONS IN TELECOMMUNICATIONS INFRASTRUCTURE AMONG CITIES
Reasons For Variations Among Cities
Measuring Variations in Telecommunications Infrastructure
Cities for ComparisonThe Telephone System
The Internet
Cellular Telephone Service
Cable-Company Internet Access
Some Summary Points
Table 1: Internet Access Technologies, 1998
Table 2: Telephone Access Lines by State, 1996
Table 3: International Measures of Infrastructure Stock, 1996
Table 4: Indices of Local Telephone Company Infrastructure Modernization, 1996
Table 5: Competing Local Exchange Carriers Operating in Selected States, mid-1998
Table 6: Complaints about Local Telephone Companies Reported to the FCC, 1996
Table 7: Service Indicators for Business Customers, 1995
Table 8: International Telephone Service Quality, 1996
Table 9: Business Telephone Rates in Selected U.S. Cities, January 1998
Table 10: Business Telephone Charges in Selected Countries, 1996
Table 11: Location of Major U.S. Internet Peering Routers, January 1998
Table 12: Intercity Internet Bandwidth for Selected Areas, December 1997
Table 13: High-Speed Internet Usage in Selected Countries
Table 14: Internet Domains in Selected U.S. Metropolitan Areas, January 1998
Table 15: Internet Usage in Selected Countries
Table 16: Number of Internet Service Providers Offering Designated Services in Selected Areas, December 1997
Table 17: Average Internet Download Times from Major Websites, December 1997 and January 1998
Table 18: Business ISDN Rates in Selected Cities, January 1998
Table 19: Internet Access Monthly Charges in Selected Countries, August 1996
Table 20: Cellular Phone Use in Selected Countries, 1996
Table 21: Cellular Phone Monthly Charges in Selected U.S. Cities, January 1998
Table 22: Annual Cellular Phone Cost in Selected Countries, 1996
Table 23: Availability of Internet Service via Cable Modem in Selected Areas, March 1998
Figure 1: Summary of New York's Strengths and Weaknesses
Background
The fiscal health of New York City is tied closely to its economic health; municipal government depends primarily on locally raised revenues to support itself. To attract and retain economic activity the city must be competitive; in an open society people and firms can and will move to the places that they find most attractive.
In assessing the competitiveness of urban areas, employers weigh many factors, including some over which local government has little influence; for example, climate and the availability of private amenities. However, three related aspects of state and local policy do bear significantly on firms' location decisions-taxes, public services, and the infrastructure.
The research activities of the Citizens Budget Commission (CBC) consider each of these three aspects of competitiveness. Prior work has developed a program of tax reform for State and City government that would enhance competitiveness. This program includes reductions in the State's gross receipts tax, reductions in the City's and State's personal income tax, eliminating the City's unincorporated business tax and other unique business taxes, and equalizing local real property tax rates among commercial and residential property.
In order to provide a competitive package of public services with the resulting reduced revenues, New York State and New York City must achieve higher productivity in their delivery of services. This will require better utilization of technology, especially information technology, in delivering services and changes in the way public employees are deployed and compensated. To encourage such positive changes, the CBC monitors the performance of municipal agencies and undertakes research to develop recommendations for better use of technology in specific State and municipal agencies. For example, a recent report identified ways to improve municipal revenue collection using information technology that would both lower collection costs to the City and save individuals and private firms substantial time and money in making their payments and resolving disputes.
The third and most recent component of CBC's research program is an examination of the ways in which State and City government can improve the competitiveness of public infrastructure. A critical first step in this analysis is to gain an understanding of where New York stands currently. One can recommend convincingly that taxes be lowered in part because tax burdens are significantly lower in other cities; similarly, recommendations about investments in infrastructure ought to be informed with information about the relative condition and performance of these facilities.
Accordingly, the CBC initiated studies of the competitiveness of three components of New York's infrastructure: mass transportation, airports, and telecommunications. Each, in its own way, is vital to the effective performance of business activity. This paper focuses on telecommunications infrastructure; two future papers will deal the other two selected infrastructure elements.
Nature of Telecommunications Infrastructure
In this paper, telecommunications is defined to include only two-way communication; one-way exchanges such as radio and television broadcasting are not included. This telecommunications infrastructure can be analyzed in four categories: the conventional telephone system, the wireless or cellular system, the Internet, and the cable television system as a potential means for interactive communications. Each category of infrastructure can be assessed in terms of the scale and condition of the physical facilities, the quality of the services provided, and the cost of the services to consumers.
In assessing New York's telecommunications infrastructure, the most important points of comparison are urban areas that compete for similar types of economic activity. Domestically, greatest attention is given to the rival cities of Chicago and Los Angeles; internationally, greatest attention is given to the other "world cities" of London and Tokyo. In addition, data was sought from other areas that observers identify as possible models for future communications infrastructure: the San Francisco area for its Internet access, Hong Kong and Singapore for their telephone systems, and Helsinki for its wireless and other infrastructure.
All the relevant and desired data were not consistently available. In considering the findings presented below, readers should be aware that:
- Data often are not available for cities or metropolitan areas, so inferences sometimes must be made from data relating to states or nations.
- Available data generally relate only to publicly owned or regulated infrastructure; additional facilities built by corporations or universities for private purposes are not included necessarily in the analysis.
- Domestic and international sources typically collect their data using different methods and definitions, so comparisons across the data sources should be made cautiously.
Nonetheless, the available data can be used to describe reasonably well the competitive status of New York City's telecommunications infrastructure. This is the first such study available to the public.
Findings: The Telephone System
The newly developed Asian business centers of Hong Kong and Singapore have telephone system infrastructure that is superior to New York's in several ways: the average annual level of investment is higher in Hong Kong ($568 per access line) and in Singapore ($231) than in New York ($118). In Hong Kong a larger share of the system's wires are fiber-optic cable as opposed to older copper cable than in New York (data are not available for Singapore). The cost of service is significantly lower in both Asian cities than in New York. The reliability of service is notably greater in Singapore than in New York. However, New York is better "wired" in the sense that it has more telephone access lines per capita (.65) than Hong Kong (.55) or Singapore (.51). New York also may benefit in the long run from having more competition for telephone service (several firms now compete with Bell Atlantic) than is the case in the still largely monopolistic systems in Hong Kong and Singapore.
Relative to domestic competitors, New York has a better physical infrastructure. The state has more access lines per capita (.65) than California (.64) or Illinois (.62). New York also has a larger share of its system composed of fiber-optic cable (10 percent) than California (6 percent) or Illinois (9 percent). All access lines in New York City are connected to digital switching equipment, but the share is lower in Illinois (83 percent) and California (79 percent). Annual investment levels also are higher in New York ($118 per access line) than in Illinois ($103) and California ($114).
New York's competitive disadvantages relate to cost and quality of service. Not only are costs lower in Asian centers, they also are much lower in the selected domestic comparison cities. The monthly business charge for basic service in New York ($16.23) is nearly twice that in Chicago ($8.57) and well above Los Angeles ($10.32), and per call charges also are significantly lower in those areas. Evidence of poor quality of service in New York is somewhat dated (with most data from 1995), but it includes higher rates of reported trouble per thousand lines (52 for New York, 31 for Chicago, and 24 for Los Angeles); much higher frequency of consumer complaints (more than 100 times greater in New York than in California), and more frequent unscheduled service outages (New York's rate was nearly triple that of California). Evidence of recent improvement includes an independent residential consumer survey placing Bell Atlantic sixth among 13 major local phone companies and ahead of the system serving Illinois.
Findings: The Internet
The United States faces little international competition as the center of Internet facilities and use; it alone has about one-half of the world's individual Internet users and one-half of the world's Internet hosts (computers connected to the Internet). Internet users in American cities have a wider choice of Internet Service Providers (ISP) than do firms in foreign cities because of heavier usage and less regulation; for example, over 100 ISPs compete in New York, Los Angeles and Chicago, but there are no more than eight in Hong Kong or Singapore. Because Internet facilities and content are concentrated in the United States, download times tend to be quicker in U.S. cities than elsewhere; for example, average download times for major websites averaged about 13 seconds in New York versus 19 in Tokyo, 29 in London and 105 in Hong Kong.
Within the United States, however, New York is not the premier city for Internet infrastructure. The unique physical components of the Internet are "routers," the large computers that connect Internet lines and sites. There are 14 major routers in the United States; five are in the San Francisco area, three are in Washington, DC, two are in Chicago, and New York is one of four cities with just one router. New York also trails both the San Francisco area and Chicago in the capacity or "bandwidth" of the major cables connecting it to the Internet-7,041 megabits per second versus 8,688 and 7,915, respectively. As with telephone service, New York also suffers a price disadvantage; for example, the monthly charge for eight hours of Internet use per day using a high speed modem (ISDN) connection in early 1998 was $433 in New York, compared to $263 in Chicago and $287 in both Los Angeles and San Francisco.
New York's advantage with respect to the Internet relates more to content than physical infrastructure. There are more Internet domains (computer addresses) in New York (132,414) than any other metropolitan area; the San Francisco area (102,308) ranks third behind Los Angeles (124,037), and Chicago ranks fifth (43,337) behind Washington, DC (43,416).
Findings: Wireless Phone Service
In most countries wireless phone service is a complement to the standard service, and the system depends on the standard telephone grid for connections. The unique infrastructure of wireless service is a network of antennae for local distribution. There are no systematic measures of the capacity of antennae systems, but objections to their installation or construction are common. While companies in New York City have encountered some such resistance, evidence suggests the opposition is greatest in smaller cities and suburban areas.
Cellular phone service is more widespread in several nations including Finland (29.2 subscribers per 100 residents), Hong Kong (21.6) and Japan (21.4) than in the United States (16.5). However, the U.S. is ahead of Singapore (14.1) and the United Kingdom (12.2).
Within the United States there is variation among urban areas in the price for cellular services. As of early 1998 New York was characterized by high rates, especially for business customers with high usage levels. For example, for a customer using 500 minutes per month, the lowest available rate in New York City ($180) is nearly twice that in Chicago ($100) and well above San Francisco ($130). Internationally, New York's rates are well above those in Singapore, but more competitive with those in Hong Kong and Japan.
There are no reliable, publicly available data relating to variations in the quality of cellular service. Interruptions in calls are a frequent complaint among users in New York City; this may be a consequence of its unique features including the "canyons" between its skyscrapers, but other features in other cities also can create problems for radio wave transmission.
Findings: Interactive Cable Television
The United States is better positioned than most nations to use cable television as a means of Internet access because a relatively large share, about two-thirds, of its households already subscribe to a cable service compared to about 30 percent in Japan, 15 percent in Hong Kong, 10 percent in the United Kingdom and under 5 percent in Singapore.
The use of cable television infrastructure for Internet access varies widely among areas of the United States. Among large cities, only Los Angeles has cable-modem access widely available; this service is available only in selected suburbs of Chicago and the San Francisco area. One provider in the New York area has announced plans to make the service available in the near future.
Policy Implications
New York has aspects of telecommunications infrastructure in which it excels and others where it lags major competitors. Initiatives in the future should be guided by an understanding of the sources of these strengths and weaknesses.
Strengths
New York has developed an advanced telecommunications infrastructure for three reasons: (1) strong demand for telecommunications services, (2) an ability to meet demand efficiently because of the city's dense development, and (3) competition in providing new services.
With the largest urban economy in the United States, New York's demand for telecommunications infrastructure is the strongest in the country. This demand attracted new construction of infrastructure by the private sector.
The demand for telecommunications infrastructure in New York exists because some firms demanding telecommunications infrastructure want to be in New York. Not all firms demand telecommunications infrastructure equally. Three types of firms generate strong demand-design and manufacture of high-technology equipment, financial services, and electronic media and commerce.
The firms directly involved in high-tech manufacturing want to be near similar firms. In the U.S., that critical mass of high-tech manufacturing exists only in the San Francisco area. It is unlikely that any firm producing technology would locate in New York, because New York lacks the critical mass of firms found in Silicon Valley. New York was never in the running for these technology manufacturing industries-Boston's Route 128 was the closest competitor to Silicon Valley, and was surpassed by it years ago.
Financial service firms historically concentrated in New York in order to be close to its stock exchange. They have remained and expanded in New York because of important benefits of agglomeration, cultural attractions and other amenities. The strong demand from these major financial institutions was an important reason for the significant pace of infrastructure investment by local telephone service providers.
The third category of firms-electronic media and commerce-have been expanding in New York. These firms want to be near similar firms and near their clients. Being near similar firms creates a skilled labor pool from which to draw, a network of suppliers and service providers (like lawyers and accountants), and a network of people with whom to share ideas. These firms also want to be near clients, who may be publishing companies or financial firms. Finally, since these firms tend to employ young, skilled workers who are less likely to be tied to a location for personal or family reasons, these firms seek locations with amenities.
New York offers proximity to similar firms, to suppliers, and to clients, as well as an improving quality of life. The public sector can reinforce these advantages in a range of ways. Policies that encourage the formation and development of small business generally, such as tax reductions, zoning exemptions, regulatory relief, and transportation access, attract young high-tech firms in particular. New York can create more demand for telecommunications infrastructure with policies that make the city more attractive to new businesses generally. Similarly, policies that make the city more attractive for new residents help sustain a talented labor force; phasing-out the inequitable subsidy granted longer-term residents through local rent regulations would be one such positive step.
The density of development in New York's central business district provides an advantage for telecommunications infrastructure development. Less wiring is necessary to reach customers in a small area, making some aspects of installation more efficient.
Competition is the third element behind New York's strengths. New York has been the leader in telecommunications competition in the U.S. It was the first city to allow telephone competition; now it is the site of intense competition to build new infrastructure. Even before telephone competition, New York benefited from NYNEX's tendency to invest heavily in advanced infrastructure, especially for its large clients in the financial services sector; nonetheless, new entrants are the best assurance that the growth in infrastructure will continue to match the growth in demand.
In the future, competition will expand to include different modes of communication, as telephone companies start to offer video services over phone lines, or as ISPs start to offer telephone service over the Internet. Fostering all of these forms of competition will keep New York advanced in telecommunications infrastructure.
Weaknesses
Although New York has in many ways a relatively strong telecommunications infrastructure, it also has some weaknesses. Specifically, its telephone infrastructure is less developed than that of Hong Kong and Singapore; its Internet infrastructure hardware is less extensive than that in San Francisco; until recently it has had poorer telephone service quality than other leading domestic cities; and its charges for telecommunications services are higher than in some other leading U.S. cities and most foreign competitor cities.
New York's telephone infrastructure is less developed than those in Hong Kong and Singapore, but it should not be a goal of public policy to try to eliminate this gap immediately. Hong Kong and Singapore's lead is an example of technological "leapfrogging," the adoption of the newest technologies by countries that until recently had been less technologically developed.
Leapfrogging arises for two reasons. First, new technologies may be substitutes for older technologies that are still useful, even if they are no longer cutting-edge. The gap between the cutting-edge (digital switching, fiber-optic cable) was greater in Hong Kong and Singapore than in the U.S., where a workable system exists, so Hong Kong and Singapore invested more in cutting-edge technology than the U.S.
The second reason for leapfrogging applies to infrastructure networks in particular. It often is cheaper to build an infrastructure network from scratch rather than trying to upgrade existing networks, particularly if the old network is embedded in a set of other physical structures. In the U.S. and other countries that have been rich for decades, most of the telephone modernization has involved retrofitting older networks, which is more expensive than laying new networks where none existed, as in newly rich countries in Asia and elsewhere.
The term leapfrogging implies that Hong Kong and Singapore's current lead in telecommunications infrastructure is temporary. The next technologies will be relatively more attractive, and will be adopted more rapidly, in countries that invested less heavily in the latest round of technology. Therefore, the fact that New York's telecommunications infrastructure is less advanced at the moment than Hong Kong's and Singapore's is not the product of a policy or regulatory failure and should not be a target for policy or regulatory reform.
Another infrastructure deficiency is New York's lack of Internet infrastructure hardware relative to the San Francisco area. Again, this is more the result of long-term economic processes than of policy or regulatory failures. The concentration of Internet resources in the San Francisco area has much to do with the concentration of computer hardware and research firms there. Since these firms are the heaviest demanders of Internet infrastructure, San Francisco has become the world's center of Internet routers and bandwidth.
For New York to attract the Internet infrastructure that the San Francisco area has, New York would need a computer and software industry to rival that area's. Historically, that was never a possibility. San Francisco will remain the geographic center of the Internet. New York's strategy should not be to imitate San Francisco, any more than it should be to imitate Hong Kong and Singapore. New York's economy is based on industries that attract less Internet infrastructure investment, and that difference is beyond the scope of public policy, at least in the short term.
New York's historic weakness with respect to telephone service quality is becoming less pronounced. The most recent data (1998) show improvement in consumer satisfaction. Local telephone competitors have based most of their advertising campaigns on providing better service (as opposed to advertising lower prices or different technological offerings), and it is likely that greater competition will lead Bell Atlantic to improve service quality beyond the levels achieved by NYNEX.
Telecommunications charges also should be determined by the market, if there is competition. The fact that cellular phone charges are higher in New York than elsewhere, and that the cellular market is competitive, suggests that some of the high charges are due to the higher cost of doing business in New York. However, the relatively high prices for standard business and ISDN services cannot be explained solely by New York's cost of living. These charges are over 50 percent higher than charges in the other American cities, a gap far larger than the difference in cost of living. These high charges are regulated by the New York Public Service Commission; it should reassess the policy considerations and cost data underlying the rates or permit competition to play a stronger role in determining prices.
A century and a half ago, the opening of the Erie Canal turned New York into a center for national trade, and this solidified New York's position as America's leading city. A few decades later, the railroads made Chicago the nation's crossroads. Another fifty years on, California's dams and aqueducts brought water to Los Angeles, allowing the city to support the spectacular growth that transformed it into the capital of the Pacific. Throughout history great cities have built great infrastructure; in America, it is also true that great infrastructure helps to build great cities.
The latest chapter of this history is the story of telecommunications infrastructure. For decades businesses have become more and more reliant on telecommunications, and the recent growth of the Internet has accelerated this trend. Imagine a breakdown, just for a few hours, of the telecommunications system. Stock markets would close, deliveries would be stalled, sales forces would have nothing to do. It would feel like an emergency, but dialing 9-1-1 would do no good. Business has become dependent on telecommunications infrastructure. A city with an inadequate telecommunications infrastructure cannot thrive. Existing businesses, fed up with failed telephone calls and slow Internet access, would eventually leave. New businesses, particularly in the fast-growing industries most reliant on telecommunications, would never arrive. These businesses would go some place else, some place with a range of telecommunications services of high quality at reasonable rates.
This paper examines New York's competitiveness with respect to telecommunications infrastructure. To date there have been no systematic comparative studies of the telecommunications infrastructure of different cities. Anecdotal evidence exists, of course, but this ought not be the basis for formulating public policy. This paper compares New York's infrastructure with that of New York's main competitors, domestic and abroad.
The first section provides background on telecommunications infrastructure. Because telecommunications technologies, regulatory policies and market environments are complex and changing rapidly, discussion of infrastructure policy sometimes requires use of new and unfamiliar terms. This background section provides the necessary context to understand the data and the analysis. The second section explains how telecommunications infrastructure varies across locations, and why these differences matter to businesses. Also, the second section describes the approach to, and the difficulties of, measuring telecommunications infrastructure. The third section presents the data and highlights New York's position relative to the other cities. The fourth section discusses the policy implications for New York.
Telecommunication is the transmission of data using some kind of electromagnetic system. The data may be sounds, like a human voice, pictures, text, or numbers; the distance over which the data are transmitted may be across the street, across the city, or across the world. Typically, telecommunications refers to the two-way exchange of data. A radio, in this sense, is not considered a form of telecommunication because the communication goes only one way. You cannot send a song over the radio back to the deejay at the station.
The main forms of telecommunications at the moment are the telephone (both wired and wireless), the Internet and, in some places, cable television. That statement is qualified with "at the moment" because telecommunications is perhaps the most rapidly changing industry in the modern economy. Only five years ago few people had heard of the Internet, and wireless technology was new. Five years from now the fax machine might be obsolete, the Internet and wireless as common as the telephone is today, and yet-unincorporated firms might be the leaders in new technologies.
This introduction to telecommunications covers the telephone, wireless communication, the Internet, and cable television. It emphasizes physical infrastructure and regulatory issues.
The telephone is the most familiar kind of telecommunication. In part, this is because it is nearly ubiquitous in rich countries, is over a century old, and is the basis for newer forms of telecommunications.
The telephone system consists of a network of local telephone grids, plus long-distance and international connections that link the local grids. In the United States, these local telephone grids are operated by publicly-regulated private companies, which, until recently, had a monopoly over their local areas. The local grid consists of wires-copper wires, mostly-connecting nearly everyone in the service region to each other and to the telephone company's central offices (COs). These offices contain switching stations, which is where the wire leading to one user gets connected to the wire leading to another user when one places a call to another. Higher-capacity cables, usually fiber-optic, connect the COs. The formal name for the telephone system is the Public Switched Telephone Network-"public," meaning that the local telephone company is required to offer access to anyone at an agreed (regulated) rate, and "switched," meaning that any user of the network can be connected not only to the phone company, but ultimately to any other user of the network.
The wires and switches in a local area are called a Local Access and Transport Area, or LATA. InterLATA traffic, the formal term for long-distance calls, travels over higher-capacity wires that connect LATA's together. Local and interLATA service differ in three ways: physical infrastructure, pricing, and competition. First, the physical infrastructure of the local system consists largely of older copper wires, while the interLATA wires are being upgraded to fiber-optic cables. (Remember those ads that Sprint used to run about hearing the pin drop during a long-distance call? Those ads were promoting their network of fiber-optic cables.) Second, the pricing of interLATA calls (and international calls) far exceeds the cost of providing those calls. In contrast, local phone service in most of the U.S. is charged at a flat rate, rather than charged by the minute, and local service is subsidized for rural and low-income persons. The result of these pricing differentials is that the interLATA and international telephone markets are easy targets for competition, which is the third major difference from local service. To understand this fully, a quick detour to regulatory issues is necessary.
Regulatory issues
Telecommunications regulation in the United States has had two recent milestones. In 1984, AT&T signed an antitrust consent decree that broke up the existing national telephone company into "Baby Bells," smaller companies responsible for providing local telephone service over specific service areas. The Baby Bells were restricted to local service-that is, they could not offer long-distance service-and AT&T refocused its efforts on long-distance only. Competition was enhanced in the long-distance market, and firms like Sprint and MCI began to offer long-distance service in competition with AT&T.
Regulation of the Baby Bells is handled by the states. State public utility commissions set rates for local and intra-state calls.
The regulatory context again changed dramatically under the 1996 Telecommunications Act. The 1996 Telecommunications Act eliminated two barriers that had been central to telecommunications regulation: The barrier between local and long-distance, and the barrier between telephone and other forms of telecommunication. The barrier between local and long-distance was broken in that local telephone companies would be permitted to offer long-distance service, and local service would be subject to competition. Firms can compete in local telephone markets in two ways. They can build their own infrastructure networks and serve their subscribers using their own wires, or they can buy time in bulk on the main company's network and resell this time to their subscribers. The first group is called facilities-based carriers; the latter is called resellers. Together these competitors are called Competitive Local Exchange Carriers (CLECs), and the former monopolies are called Incumbent Local Exchange Carriers (ILECs).
This deregulation in America is part of a worldwide trend: The World Trade Organization called for the dissolution of national telephone monopolies in 1998. Significantly, Japan is dividing NT&T, the world's largest telephone monopoly, into two national providers and an international service provider. But generally the impact of this deregulation has not yet been felt; the primary effect in the U.S. has been the merger of Baby Bells (Pacific Bell with Southwestern Bell and NYNEX with Bell Atlantic [1]). In OECD countries, 99 percent of telephone subscribers are still customers of a telephone monopoly. [2]
Some areas have felt the benefits of competition. Telephone service providers have focused initiatives on large businesses and big cities in order to gain or retain a large share of these markets. Whereas the gap in telecommunications services between large cities and other places had been shrinking, partly due to regulatory efforts to equalize service, big cities are now redeveloping a lead in telecommunications offerings due to competitive initiatives.
The 1996 Telecommunications Act also eliminated the barrier between telephone and other telecommunications markets. Telecommunications companies are free to offer a range of telecommunications services; most notably, cable television companies are permitted to offer Internet access and even telephone service. While there may be technological difficulties in entering other telecommunications markets, firms are no longer legally forbidden from offering as wide a range of services as they are able.
Why, in countries with market economies, did telephone service remain a monopoly for so long? Telephone service depends on an expensive network of infrastructure that, once built, costs little to operate. In economic terms, telephone service involves a very high fixed cost with very low marginal costs. It is most efficient for one infrastructure network to provide telephone service for everyone rather than companies setting up competing, and possibly redundant, infrastructure networks. Sometimes a company will build its own infrastructure, such as Sprint did with its fiber-optic network and Internet companies are now doing (see below). Usually, though, competition consists of new companies leasing space on an existing network, which is owned by an established company that does not welcome competition. It also requires agreement on technical standards; for instance, competing local companies are required to allow customers to keep their old phone number when they switch to a new local provider. Working out infrastructure leasing arrangements and technical standards is difficult.
Two major additional regulatory principles are universal service and equal access. Universal service is the provision of telephone service for all people in the local company's service area. The Federal Communications Commission (FCC) requires local telephone companies to subsidize telephone service for rural customers (to whom it may be expensive to lay telephone wires) and for low-income customers. Equal access means that long-distance providers must be able to offer "1+" dialing; that is, once a customer chooses a long-distance company, the customer should not have to dial an access code each time she or he makes a long-distance call. Nearly all areas of the United States have achieved universal service and equal access.
Mechanics
The most important way in which the telephone differs from other kinds of telecommunication is in the mechanics of transmission. The human voice is a far more difficult kind of data to transmit electronically than either numbers, text, or pictures. Technically speaking, there are two kinds of data: analog and digital. Analog data are continuous or continuously varying, while digital data are expressed in discrete pieces. Numbers, text, and pictures are usually digital forms of data, but the human voice tends to be transmitted as analog data. Rather than being separated into pieces, as text might be, the human voice is transmitted over the phone as a continuous stream of data, just as one would hear it in person.
Because analog data are not divisible into small pieces, when a phone call takes place, a connection is made between the caller and the callee that lasts for the duration of the call. That is, a set of physical wires linking the callers is engaged during the call and cannot be used for anything other than that call. If there is a lull in the conversation, the line cannot temporarily be used to transmit another phone call. The wires constitute a "dedicated communication path." This technology is called "circuit-switching" because the entire circuit (that is, a wire) is tied up for the duration of the phone call. Circuit-switching technology typically uses copper wiring, which is the predominate form of circuitry within the local telephone systems.
The technology for converting the human voice from analog data into digital data now exists, but the invention of this technology came long after the invention of the telephone and the development of national telephone systems. Until recently, digital transmission of the human voice sounded choppy and disconnected.
The alternative to circuit-switching is called "packet-switching," which is the routing process used by the Internet and will be discussed at length below. For now, it should suffice to define packet-switching as the transmission of data that has been broken down into electronic pieces that are reassembled at their destination. Unlike circuit-switching, packet-switching allows the sharing of a single line for the simultaneous transmission of data of different kinds and from different sources.
The data-carrying capacity of a circuit is referred to as its "bandwidth." Technically, bandwidth is the range of frequencies that a transmission channel can support, but one can think of bandwidth simply as the size of a circuit. Local telephone circuits are usually copper wires, which have a lower bandwidth than the fiber-optic cables that carry many long-distance and international calls (and some Internet traffic, too). High bandwidth cables (sometimes called "broadband") are more expensive to lay and operate than low bandwidth cables and therefore tend not to be used with circuit-switching technology, since circuit-switching requires the dedication of an entire line for the duration of the call.
These distinctions can get confusing, but they are the most helpful way of understanding the world of telecommunications. To clarify, analog and digital are kinds of data, circuit-switching and packet-switching are routing processes, while copper wires, fiber-optic cables, and other kinds of circuitry are kinds of infrastructure. Analog data requires circuit-switching. Digital data can be sent using either kind of routing process, but packet-switching is more efficient. Each routing process can work with a range of infrastructure, so one cannot neatly pair up each routing process with a particular kind of infrastructure. What makes telephone service distinct from other kinds of telecommunications is the data and the routing process, not the infrastructure; only telephone service employs analog data and circuit-switching, while using various kinds of infrastructure.
Most explanations of the Internet begin by trying to describe it physically. What distinguishes the Internet from other kinds of telecommunication, however, is not its physical structure. Basically it shares the same infrastructure as the telephone system uses. The above discussion of data, routing processes, and infrastructure should help make this clear. The Internet is essentially a special routing process called packet-switching.
Packet-switching is a process by which data being sent from one point to another are broken into pieces, sent separately across a network of wires, and reassembled into coherent data at its destination. The data must be capable of being broken into pieces-in other words, the data must be digital, which includes text, numbers, pictures, some sounds, and, as technology improves, the human voice and video. These pieces are called packets, and each one travels across the network of wires with a set of instructions about where it is supposed to end up.
When some data are sent across the Internet, the packets travel jumbled together, regardless of their origins and destinations. Using the Internet, unlike the telephone, does not require a dedicated communication path from the source to the destination. The Internet, therefore, is a far more efficient way of sending enormous amounts of data. A road traffic analogy may be useful: Sending a message using the telephone is like having to reserve a dedicated lane of traffic exclusively for yourself for the entire duration of your trip. So long as you are traveling, no one else may get in any lane on any road that you use over the course of the trip. When you arrive, the roads you used are then freed for the next user. Sending a message using the Internet is similar to how road traffic actually works, with many cars sharing a road, since at any given time a car only needs a bit of space, even if your car ultimately covers a great distance.
How did this routing process, which would become the Internet, arise? The U.S. Defense Department wanted to send messages that would be safe from interception or military attack. Since packet-switching splits a message into tiny pieces, sending a message using packet-switching is like shredding your document, sending the shreds, then reassembling it at the destination. The shreds alone could not be deciphered without knowledge of the re-assembly process.
The other feature of the Internet that made it suitable for military use was that the "routers"-the computer stations at which packets are directed to their next stop-are arranged in a web-like network rather than in a hierarchy. Whereas a phone call involves a specific connection from the caller to the caller's telephone company's central office, then to the callee's telephone company's central office, then to the callee, an Internet message can travel many different ways among the web of routers. This design protects the network from military attack. Suppose an atomic bomb knocked out one router; then Internet traffic would still move, but along different paths connected by other routers.
Physical infrastructure
The Internet's physical infrastructure consists of leased telephone lines, some other lines, and routers at key connection points. These routers are big computers that direct Internet traffic along the lines. There are 14 major routers in the United States. Most of the lines that constitute the Internet are leased from telephone companies by Internet service providers (ISPs); some of the lines were laid by the ISPs themselves. By leasing and laying these wires and cables, ISPs create networks along which data can travel. ISPs range in size from big national firms to small neighborhood companies. The big national ISPs (UUNet, BBN Planet, and Earthlink, for instance) lease and lay high-capacity fiber-optic cables that connect cities across the country and the world; this network of leased and laid cables is called the Internet backbone. These big ISPs also own routers that direct traffic along their own networks. Smaller ISPs lease some local telephone wires and enter agreements with larger ISPs to connect to the Internet backbone. And some ISPs, like America OnLine (AOL), also offer content to which only their subscribers have access.
All ISPs use the same packet-switching technology (called TCP/IP) to route data. The major routers are where the networks of different ISPs connect, or "peer," with each other. At these major routers, data can be transferred from one ISP's network to another. Via these major routers and the ISP's own routers, data can be sent between any points in the world. When people describe the Internet as a "network of networks," they are referring to the manner in which the big ISPs each have developed networks that are connected. Some ISPs have set up smaller private routers in order to peer with each other and avoid the congestion at the major routers.
When people use the Internet, they are connected to it through an ISP. Some users, such as small businesses and home users, use a modem to connect to the Internet. When these users send or receive data over the Internet, they are sending and receiving data through the local telephone company's network, since the modem dials up the ISP on a standard telephone line. Thus, an e-mail message travels from the user's computer, over the standard telephone line through the telephone company's switching station, then to the ISP, at which point the data are routed using packet-switching as outlined above. Other users, like universities and corporations, contract with an ISP to have a special leased line that connects them to the Internet. These leased lines are identified by their data transmission capacity, or bandwidth (see table on Internet access options below). For these users, the telephone company leases high-capacity cables connecting the business and the ISP through the telephone company's wire center. Data travel quickly between the business and the ISP, which is in turn connected directly to the main cables of the Internet. With either kind of Internet connection-dial-up modem or leased line-the end-user deals directly with the ISP, but the telephone company provides the essential physical connection.
In summary, the infrastructure of the Internet relies on the telephone system in two ways. First, ISP's lease local, long-distance, and international telephone lines in order to build their own networks. Second, Internet users must rely on the local telephone system for their connection to an ISP, even if the telephone company provides this connection behind the scenes and charges the ISP directly.
Internet usage, access and speed
Telephone conversations consist of data being transmitted in real time-the speed of the human voice. On the Internet, though, packets are sent out in a single burst (unlike the phone, where sounds come only at the speed of speech). Internet users would like their transmission to be instantaneous, but one of the most common complaints about the Internet is the sluggishness of data transmission. Several factors affect this speed. They are best explained with an illustration.
Suppose I am using the World Wide Web. The Web is an application for the Internet that allows users to view multimedia information. Data on the Web might contain text, pictures, sound, video, or a combination. (There are many other Internet applications, such as telnet and FTP.) I would like to look at the website for Amazon.com, an Internet bookstore. To do so, I type in the address for Amazon.com <http://www.amazon.com> in the appropriate space on Netscape, which is the software I use to look at the Web. Hitting return sends the information that I want to look at Amazon.com's website. This request travels from my home computer, through my modem, over the standard telephone line to the telephone company's switching station, where it is then directed to my ISP, which is Harvard University. Harvard is a small ISP that contracts with a big ISP-AT&T. A relatively small router at Harvard directs my request to AT&T's nearest (and larger) router, which is also in Cambridge. Amazon.com, however, is connected to the Internet through a different ISP called UUNet. So my request travels along AT&T's Internet backbone to New York, where there is a major router at which AT&T's and UUNet's backbones connect. My request is transferred at that router to UUNet's network, along which it travels to Seattle. Finally UUNet transfers my request to a cable linking Amazon.com directly to the UUNet network. My request ends up at the big computer (called a server) at Amazon.com. In response to this request, the Amazon.com server sends back a huge amount of data; all the text and pictures that constitute the introduction page to their website (called the "home page"). That huge amount of data travels back to me, possibly along the same route or perhaps along another, since there are other major routers where the AT&T and UUNet backbones connect. On a good day the home page pops up on my screen in under five seconds. I then might send back some data, such as the following request: Please list all the poetry books by Elizabeth Bishop you have for sale. They answer, and so on.
More than anything else, what determines the speed of this data exchange is the endpoints. The nature of the link between my computer and my ISP is probably the greatest bottleneck in this process. ISPs offer all kinds of access; there are dial-up connections with a regular modem and a regular phone line, which are the slowest kind of connections. Those speeds are now 28,800-33,600 bits of data per second (referred to as 28.8K or 33.6K). Other technologies are faster, and they consist of more advanced modems that transmit data faster over the same telephone lines, and different lines from my house to the ISP. The first category, advanced modems, includes the Integrated Services Digital Network (ISDN) and the Asymmetric Digital Subscriber Line (ADSL). The second category, different lines, includes direct fiber-optic links like T-1 and T-3 lines. On the Amazon.com end, they probably have a T-3 or other high-capacity cable from their company to UUNet, their ISP. What slows them down might be the speed at which their server can pump out data in response to all of the requests they get.
Table 1: Internet Access Technologies, 1998
Another potential congestion point is the routers that send packets to the next destination, like the router in New York that connects the AT&T and UUNet backbones (as well as those of many other ISP's). These routers need to identify and pass on an enormous amount of data in all directions. Computer companies are actively trying to develop new technologies to ease congestion at these routers.
The location of these routers can affect the speed and quality of the Internet at different locations. According to a respected technology magazine, in 1996 Minnesota was cut off from the Internet for 13 hours because all of its Internet traffic depended on a particular router that malfunctioned temporarily. [3] Furthermore, an Internet consultancy reports that Internet transmission times can vary across cities by a factor of four depending on where routers are located relative to each city. [4]
The main cables and wires of the Internet generally are not sources of congestion. It may even take the packets of data more time to "transfer" in New York from AT&T to UUNet than it does to travel along UUNet's cables from New York to Seattle. The cost of leasing or laying cables is very low relative to the cost of upgrading or installing routers. So, continuing the road traffic analogy, the expressways move very quickly. Congestion is likelier to occur at expressway interchanges and likeliest in the driveways and parking lots.
These patterns of usage and congestion help explain why Internet users are concerned with the kind of access they have. Internet users are forever wishing for faster modems, faster connections, and faster servers. These patterns also explain why the telephone companies are important players in Internet access, even in cases when they do not provide Internet service to an end-user directly. But the regulation and pricing of the Internet differs considerably from the regulation and pricing of telephone service, and therein lie important economic and policy issues.
Regulation and pricing
Unlike telephone service, the Internet is a new, rapidly growing, dynamic industry with many competitive firms and little regulation. The number of Internet hosts (hosts are main servers that connect to the Internet; Harvard has several, for instance, and Amazon.com has at least one) is growing 60 percent per year worldwide. In contrast, the number of telephone subscribers is growing at under 5 percent per year. [5]
Regulation of the Internet has been confined to content and end-user products, rather than regulation of the underlying infrastructure. Much of the public policy debate about the Internet has focused on defining obscene content and on ensuring privacy and confidentiality. The main regulatory issue, the Microsoft antitrust case, is about the market for Web browsers, the outcome of which will have little impact on the Internet's infrastructure of wires and routers.
Another important contrast to telephone service is that it is prohibitively difficult to measure Internet usage. Since telephone service employs circuit-switching technology, each telephone call uses an identifiable resource-the telephone lines-for a measurable period. With packet-switching technology, measuring usage on the Internet would require counting the packets that a user transmits. For now, at least, the cost of counting packets exceeds the benefits that any company or regulator would get from usage pricing. Therefore, most Internet usage is charged by a flat rate: For instance, unlimited hours on the Internet over a standard modem line for $20/month. For modem users (that is, home users and some small businesses) there is also the charge for using a standard phone line, which may charge for the duration of the connection.
For long-distance and international data transmission, telephone service is always charged according to distance and duration of the call, whereas Internet usage is generally flat rate. This difference has encouraged companies to develop technologies that allow the Internet to assume some of the functions of the telephone.
For instance, faxing a document means using telephone technology (circuit-switching) to send digital data. Just like a telephone voice call, a fax transmission requires a dedicated communications path for the duration of the call. Faxing uses telephone infrastructure. What separates a fax from a voice call is the kind of data that is transmitted. As discussed above, it is far more efficient to send digital data using packet-switching technology, and the pricing of e-mailing a document versus faxing a document reflects that. E-mailing a document around the world is essentially free, in terms of marginal cost; faxing that document around the world results in international phone charges.
A further example is using the Internet for telephone calls; that is, using packet-switching technology to transmit the human voice, which (as discussed above) depends on technology that transforms the human voice from analog data to digital data. That technology is being refined, and Internet telephony service is a small but booming business.
A necessary caveat: Most of the conventional wisdom about the Internet, be it about congestion, access speed, technological developments, or future applications, should be accepted cautiously. The Internet is so new, and so rapidly changing, that no one knows the direction it will take, even in the short run. Part of the difficulty in reporting facts about the Internet also comes from its decentralized structure; there is no central authority administering, measuring, or developing the Internet. Data on the Internet are therefore less reliable, or simply unavailable, relative to data on telephone service and infrastructure. What makes the Internet such an exciting place commercially and technologically makes it a frustrating target for research.
To understand wireless telecommunication, it is helpful to go back to the distinction between data, routing processes, and infrastructure. Wireless is characterized by its infrastructure; rather than fixed wires and cables, it comprises radio signals, antennae, and satellites. Over this infrastructure one can send either kind of data (analog and digital) using either kind of routing process (circuit-switching or packet-switching). Thus, wireless telecommunication can mean mobile telephones, but it also includes wireless fax or wireless Internet access.
A note on the infrastructure; wireless can be either fixed or mobile. In the United States, wireless technology is almost always mobile, meaning that one can use a wireless phone from any place within a fairly large service area. Fixed wireless, in contrast, consists of a fixed antenna for a user. Fixed wireless is a low-cost way of providing telecommunications services to rural or remote areas, since the wireless signals replace the long cables that would otherwise have to be strung in order to connect the remote location to the main network.
The infrastructure unique to mobile wireless service is antennae. These antennae transmit wireless phone calls by receiving signals from wireless phones and transmitting these signals to the standard telephone network. The area that an antenna serves is called a "cell," which is the basis for the term "cellular phone"; a cell may consist of anything from a single building to an area miles wide. There is considerable debate over the siting of cellular antennae, which is generally governed by local and state regulations with some oversight by the FCC. Antennae are located atop tall buildings in dense areas. In residential areas, zoning laws often restrict antennae heights or their construction altogether, and they tend to be built in nearby industrial areas or along highways. Some areas-typically suburbs and rural areas-have issued temporary moratoria on new antenna construction, restricting the development of wireless infrastructure there.
Satellites are increasingly used to transmit cellular telephone traffic. Unlike antennae, which are grounded at specific locations, satellites orbit the earth and give cellular telephones (and pagers) greater range than fixed antennae offer. Satellites are proliferating in order to provide wireless Internet access and interactive video in addition to wireless telephony. These satellite wireless services appeal most to multi-national corporations, which need to share data among their offices. [6]
Both fixed wireless and mobile wireless systems are "switched" with the standard fixed-wire telephone network. That means that from a wireless phone one can call anyone else, regardless of whether they are using a wireless or standard telephone. This switching happens because wireless providers can send signals digitally from their own antennae to the local telephone grid.
The recent growth of mobile wireless phones has been remarkable. Over the last decade more than half of new phones worldwide were mobile, and within the next decade wireless may be the dominant mode for voice conversations. [7]
Why is this growth so rapid? In most countries wireless phone service has not had to abide by the regulations that standard telephone service has. Competition in the wireless market preceded by years the current worldwide deregulation of standard telephone service. Particularly in countries where the public monopoly was inefficient in providing telephone service, wireless has been an appealing alternative. As the price of wireless continues to fall toward the price of standard telephone service, even in the relatively deregulated U.S., more and more people are selecting wireless service in lieu of a second standard telephone line.
Still, wireless growth has lagged in the U.S., compared with Europe and Asia. One reason is that, relative to other countries, standard telephone service in the U.S. is inexpensive and reliable. Another reason is a technical concern. Whereas most of the world's wireless systems use digital data transmission, which offers better sound quality and less interference than analog does, the U.S. still is largely analog. This is, in part, due to the higher prices of digital phone equipment (around $200), but also due to a lack of coordination on technical standards for digital service. The popularity of digital wireless in the U.S. is growing, though, with the introduction of Personal Communications Service (PCS), a digital technology. The introduction of PCS resulted in new services (voice mail, etc.) available to wireless subscribers in the United States.
And, ironically, the Internet has kept standard telephone service in competition with fast-growing wireless industry. Wireless service was designed for voice, not data, transmission, and accordingly has a very narrow bandwidth. Most wireless service allows maximum data transmission speeds well below the 28.8K speed of standard telephone lines. Thus, wireless cannot yet compete with standard telephone service for connections to the Internet. This will probably change, as wireless technology improves, but for now the inadequacy of wireless for Internet traffic may have kept wireless from making standard telephone service obsolete.
Many predict that physical wires and cables will stay ahead of wireless for data transmission for years to come. Broadband (that is, high-capacity) satellite systems are being put in place primarily for video broadcasting and in areas where wire-line infrastructure does not exist, but not for widespread Internet access. [8]
Before the 1996 Telecommunications Act, a discussion of telecommunications would not have mentioned the cable television industry. That act removed the barriers that separated telephone provision from other industries, so cable television companies are now permitted to offer telephone service and Internet access. Further, the 1996 act restricts state or local regulation of telecommunication services provided by cable systems.
Cable television is a different kind of infrastructure. It is a fixed network, like the standard telephone network, but with higher-capacity ("broadband") coaxial cables. Unlike the standard telephone network, it does not support circuit-switching, since the cables are shared by many users. One cannot set up a dedicated communication path the way one does with standard telephone service. Thus, the infrastructure is better suited to support packet-switched Internet traffic.
The appeal of cable television provision of Internet access is the large capacity of its cables relative to telephone lines. The speed of cable Internet access is estimated to start at one megabit, approximately 30 times the speed offered by a standard telephone modem. But since the cables are shared rather than dedicated communication lines as telephone lines are, this greater bandwidth will be shared among many users and could get congested.
The barriers to cable Internet access are technological. Although cable television cables already reach most U.S. homes, the cable network is not yet "switched." Recall that the telephone network is switched, which means that any user can be connected to any other user. Because the cable network is not switched, one cannot send a TV program or movie (say, off of videotape) either to the cable company or to another cable subscriber; transmission is only one-way. In order to provide Internet access, the cable companies must install switching equipment in their offices and modems in subscribers' homes.
Because telephone companies already have switching capacity, higher-speed Internet access though ISDN or other technologies can be provided quickly on a customer-by-customer basis. In contrast, cable companies must first add switching capacity to their network generally before being able to offer Internet access. So far, telephone companies have been able to offer high-speed Internet access more quickly than cable companies in most areas.
The recently announced purchase of TCI, the nation's second-largest cable company, by AT&T is an attempt to combine the strengths of both infrastructures. AT&T plans to add switching capacity to TCI's high-capacity cables, over which AT&T will be able to offer telephone and Internet service in addition to cable television. This convergence of technologies gives consumers the immediate benefit of consolidated provision and billing; the potential longer-term benefit is that this new network will support more advanced technologies than exist today. [9]
More broadly, the telecommunications of infrastructure of the future is still anybody's guess. At the moment there is a three-way race between the telephone network's development of new modems and high-capacity cables, the wireless network's development of high-capacity data transmission, and the cable television network's development of switching capacity. Some cities are better served by one kind of infrastructure than others, but ultimately the competitiveness of a city's telecommunication infrastructure will depend on which infrastructure, if any, becomes the dominant one.
NATURE OF VARIATIONS IN TELECOMMUNICATIONS INFRASTRUCTURE AMONG CITIES
Virtually all businesses depend on telecommunications, and for many it is a primary input. As the economy becomes more service-oriented, and as the production of data and information replaces the production of physical goods, telecommunications matters more. Telecommunications are obviously central to high-tech industries like multimedia, electronic commerce, and research, but telecommunications are no less important for sales, for customer service centers, and for routine data processing.
For New York to be competitive in telecommunications infrastructure, the city must have the telecommunications services that businesses need. Not all forms of telecommunications matter for businesses. One-way telecommunications, which allow data transmission only from the central node to the end-user, are primarily for entertainment (like radio and television) and contribute more to residential consumption than business productivity. This analysis will not cover radio or television and will consider cable offerings only to the extent that they facilitate Internet access. The analysis will focus on telephone services, Internet access, cellular phone service, and cable-based telecommunications.
What do businesses demand from telecommunications infrastructure? Businesses care about three things: (1) the availability of specific telecommunications services; (2) the quality of those services; and, (3) the cost of those services.
Businesses vary widely in their telecommunications needs. While large corporations and institutions might need their own leased-line connections to the Internet, a small business might rely on a standard phone line or two. Cities where a wide range of telecommunications services are available-not just at the highest technology frontier-will attract a wider range of businesses. These telecommunications services must not only be available, they must work reliably, and businesses must be able to find the support they need to implement these technologies. Typically, alternatives to modern telecommunications-like regular mail, messenger services, and face-to-face contact-are far slower and more expensive ways to communicate, so interruptions or difficulties in telecommunications service can bring a business to a halt. Finally, many businesses rely on telecommunications as a primary input. Between hardware installation, subscription fees, and other costs, telecommunications can represent a major part of these businesses' budgets.
Reasons For Variations Among Cities
It is not obvious that the availability, quality, and cost of telecommunications will vary among urban locations. After all, you can make a phone call to almost any point on the planet, and the call sounds almost the same. Furthermore, when you call to make an airline reservation or buy some shares of stock over the Internet, the company you deal with could be literally anywhere. The title of a recent book on technology, The Death of Distance, expresses this sense of placelessness. If telecommunications eliminates the effects of distance, then places are no longer more central or more isolated. [10] In this sense, every place is equal.
The variation in telecommunications infrastructure among places is less clear than the variation in transportation infrastructure for two reasons. First, telecommunications involves infrastructure that is often hidden from view and therefore less obvious. Many urban telephone wires are underground, and the switching stations are hidden within the telephone companies' buildings. The data that travel over telecommunications networks are equally invisible. One cannot visibly follow the path that your phone call or e-mail message takes. One cannot see congestion or a breakdown.
Second, the Internet was designed to overcome local congestion or failures. The Internet is a complex web of wires and cables, such that there are many paths between any two points. If a particular path is congested, or inoperable, data traveling along the Internet are routed along another path to the same destination. The slowdown from this kind of detour is usually imperceptible to a user, since it is generally the user's connection to the network (and the connections of the users that the first user is communicating with) that is the binding constraint on data transmission speed. Thus, the Internet was designed to minimize the variation in service among places.
But telecommunications infrastructure does vary among urban areas, sometimes dramatically. Places are not equally well served. There are telecommunications centers and telecommunications backwaters. Location matters, in two essential ways.
First, telecommunications infrastructure consists of a physical network, even though some of it is invisible. This physical network has central nodes, major thoroughfares, slow paths and isolated peripheries. Along this physical network travel physical data. Although data transmission may seem instantaneous, or nearly so, data are electromagnetic signals that travel in a real, physical sense. The time it takes a data packet to travel between two points depends on the distance between those points, just as the time it takes to transport anything else between two points depends on distance.
This hierarchy of places is truer for the Internet than for the telephone system, since telecommunications policy has stressed the importance of universal telephone service. For the Internet, though, some cities are major connection nodes, just as some cities are major airport hubs. In cities that are not major nodes, cross-town Internet communication sometimes gets routed through the nearest major city. Data coming from a more isolated city may have to travel greater distances in order to reach their destination, if the nearest connection nodes are far from that city. For instance, e-mail messages going from one computer in Boston to another computer in Boston often travel through New York or Washington, DC.
The variability in the telecommunications network is even more apparent internationally. It is much easier to lay telecommunications cables across land than across water. There are more links within the U.S. than across the Atlantic or the Pacific. Since AT&T and NYNEX have been the main providers of global telecommunications infrastructure, the major existing transoceanic links have an endpoint in the United States. This structure of the global network may put New York and other American cities at a competitive advantage to cities in other countries
Second, telecommunications providers-telephone companies in particular-are generally local or regional, rather than national or global. Telecommunications networks in different cities are often owned by different companies. Different telephone companies, under the regulation of individual states, have modernized the telecommunications network at different rates. Modernization includes the replacement of old copper wires with high-capacity fiber-optic cables. Modernization also includes the installation of digital switching equipment, which routes calls more efficiently and permits the introduction of new services like voice mail and call waiting.
Though many Internet providers are national, like America OnLine, the local telephone companies provide much of the infrastructure necessary for Internet access. As discussed above, the key determinant in the speed and quality of an Internet connection is the link that connects a user to the local or national network, the "last mile." The last mile usually depends directly on the telephone company. Thus, your local telephone company determines in part the speed and quality of your Internet connection.
Even when a telephone company is not strictly local-Bell Atlantic, for instance, serves the East Coast from Maine to Virginia--telephone service is regulated at the state level. Two businesses in different states, but served by the same telephone company might face different prices for the same service.
Competition in telecommunications services will shrink these local differences in service and cost. The Telecommunications Act of 1996 allows local telephone companies to compete in other regions and may result in national telephone service providers. Also, the Act shifts some regulatory responsibility from the states to the FCC, reducing another source of geographic variation. On the other hand, the level of competition itself varies from city to city. Businesses in large cities choose among half a dozen telephone providers; in other places, the local telephone monopoly is intact. If competition brings better service, lower prices, and new technologies, the key to a strong local infrastructure may be the encouragement of competition.
To summarize, though it is not always obvious, telecommunications infrastructure varies among cities. The physical structure of networks and the localized telephone market structure means that location matters, and businesses can choose the location that best fits their needs.
Measuring Variations in Telecommunications Infrastructure
Two ways to measure telecommunications infrastructure (or any other kind of infrastructure) are to: (1) measure the stock of infrastructure and, (2) measure the flow of services from the infrastructure. In the case of telecommunications, the infrastructure stock consists of a network of cables (primarily telephone company wires, which may be copper or fiber-optic, and also cable television company wiring), switches and routers, and cellular transmission equipment. The flow of services from infrastructure can be measured in terms of availability of services, customer satisfaction with service quality, and the cost of services.
It is more straightforward to measure infrastructure stock. It consists of physical structures that can be enumerated, and the answer is definitive. There are X number of miles of fiber-optic cable here, another Y miles there, and so on. Measuring availability, quality, and cost of services is more complicated. Availability, quality, and cost refer to intangible services, like transmitting an amount of data, rather than physical structures. And while the physical structures may be standard or at least comparable across different places, it may be more difficult to compare telecommunications services, which are affected by public policy and by local demand.
Despite these difficulties, it is important to measure infrastructure services along with infrastructure stock. What a business ultimately cares about is the availability, quality, and cost of telecommunications services. The size of the stock determines this, in part, along with how well the infrastructure is maintained, how much demand and therefore congestion there is, and how the services are priced. Therefore, the size of the infrastructure stock is, by itself, an inadequate measure of a city's infrastructure. In comparing cities to find the most appropriate infrastructure for their needs, businesses care more about the availability, quality, and cost of the services they will use than about the city's total amount of physical infrastructure.
Just as cities with better telecommunications infrastructure have a better chance to attract and retain businesses, the converse is also true: Cities with a concentration of businesses that rely on telecommunications will attract private-sector telecommunications infrastructure investment. Infrastructure investment is risky and often requires a high threshold of demand to be profitable. New technologies are introduced where previous technologies have succeeded and where consumers are knowledgeable and savvy. Thus, the current usage of telecommunications technology is a good guide to predicting where the next innovation will be introduced. A business seeking a location with good infrastructure knows that telecommunications technology is changing rapidly, and the business needs to predict whether the chosen location will continue to provide quality infrastructure. For a city, then, being competitive in telecommunications infrastructure also means having the high demand for telecommunications today that will attract additional infrastructure investment tomorrow. Measures like current Internet usage and current cellular phone usage serve as indirect predictions of future infrastructure offerings.
Measuring a city's telecommunication infrastructure is a challenge. The available data are those that regulators require, and regulation takes place only at the state and national levels. Local telephone companies are reluctant to share infrastructure stock data that they are not required to report to regulators, so it is often impossible to gather city-level data for indicators, even when state-level data are published by the FCC. Moreover, only some telecommunications providers are subject to these regulatory requirements- Internet service providers may advertise the extent of their networks, but the FCC does not force them to disclose this information. The lack of central control or oversight of the Internet may have spurred it growth and its richness, but makes it difficult to measure its size, speed, or scope.
Another complication is that much infrastructure stock is built and owned by private institutions, such as intra-university networks that connect buildings on a college campus. Though private, this infrastructure offers public benefits, since the public's ability to gain access to the institution's information over the Internet depends, in part, on the institution's infrastructure. There is no readily available data on the amount of privately provided telecommunications infrastructure.
Comparability is yet another analytic problem. Different telecommunications providers offer different services, and even when comparing the same basic technology (like ISDN, or T-1 leased lines), different providers might use different hardware, or offer different levels of technical support. So even a straightforward price comparison for a given service may be misleading. The comparability problem is magnified at the international level, since countries use different technical standards in their own national networks. Finally, even if the data were comprehensive and comparable, telecommunications are changing rapidly. No snapshot of the infrastructure represents the current situation for long, so even the most carefully constructed comparison may lose relevance quickly. With all of these caveats, this paper compares the best available data on telecommunications infrastructure in New York and its competitor locations.
To summarize the main caveats in analyzing telecommunications infrastructure data:
- city-level data are often unavailable, so conclusions must often be drawn based on state- or national-level data;
- data often refer only to publicly-owned or publicly-regulated infrastructure; it is unknown how much infrastructure corporations have built to create their internal networks;
- data often cannot be compared directly, particularly when domestic city-level data come from one source and international country-level data come from a different source; in these cases, the relative position of New York to other cities must be inferred carefully.
Cities for Comparison
The cities to which New York will be compared fall into two groups. The first are the major domestics and international competitors for those advanced business service activities-financial, legal and consulting services-in which New York specializes. The two leading domestic competitive locations are Chicago and Los Angeles. The two leading international competitors are London and Tokyo. These four cities also are used as a basis of comparison for the companion study to this paper on transportation infrastructure.
The second group are less prominent competitors, but they are cities with a reputation for particularly attractive and advanced telecommunications infrastructure. Four areas fall into this category-the San Francisco Bay area, Helsinki, Singapore, and Hong Kong. The distinctiveness of each of these places warrants a brief description.
San Francisco Bay area
The San Francisco Bay Area, which includes the South Bay communities known as Silicon Valley (centered on San Mateo and Santa Clara counties), contains a significant concentration of high-technology industry. This industry has demanded state-of-the-art telecommunications, and as a result the Internet access services in the area are anecdotally acknowledged to be the best in the country. Silicon Valley continues to grow, maintaining its dominance in the production of new technology. The city of San Francisco itself has a highly developed multimedia industry, as well as a concentration of financial firms, which are major users of telecommunications.
Helsinki
Helsinki-and Finland generally-has long been a world leader in telecommunications. Internet use is more widespread in Finland than in any other country, and telecommunications in Finland have been provided competitively for decades. New technologies are implemented in Finland as quickly and thoroughly as in the United States. In particular, the Finnish wireless market is highly developed. Helsinki may reveal itself to be the international standard in wireless services and, therefore, the benchmark against which New York might be measured.
Hong Kong and Singapore
Both Hong Kong and Singapore have established themselves as business centers in Asia, home to the fastest-growing economies in the world-at least until recently. These two cities gained economic prominence during the era of modern telecommunications and, consequently, may be less saddled with older telecommunications infrastructure than many American and European cities. Their basic telephone systems are cutting-edge; exclusively digital switching and a high fraction of fiber-optic cable. This provides an attractive base for high-speed Internet service and new entertainment technologies.
This section presents the available data for comparing New York's telecommunications infrastructure with that of competitor cities. The findings cover telephone systems, the Internet, wireless services and cable television systems.
The competitiveness of an urban area's telephone system can be gauged along several dimensions. The following analysis covers: (1) the stock of infrastructure in terms of number of lines, type of cable and switching equipment, and recent levels of capital investment; (2) the degree of competition in providing services; (3) the reliability of service; and, (4) the cost of standard services.
Access lines
A telephone access line connects a telephone user to the telephone company's central office. The density of telephone access lines; that is, the number of access lines per capita, is the most basic measure of telecommunications infrastructure. The higher the level, the greater coverage the telephone system has, and the greater the availability of telephone service. This measure varies considerably across the United States and among other countries.
Table 2 compares U.S. states in terms of telephone access lines per capita. The first column shows total access lines, both residential and business. Central cities have more access lines per capita because the ratio of businesses to residents is high in central cities, and business lines are a significant fraction (about one-third) of all telephone access lines. Furthermore, because telephone networks are concentrated in urban areas, urban residents tend to have better access to telecommunications and therefore higher rates of residential access lines per capita. Thus, New Jersey, the most urbanized state, tops the residential list, and rural states like the Dakotas are at the bottom of the list.
Table 2: Telephone Access Lines by State, 1996
Business access lines indicate the level of telecommunications infrastructure in place for business use. The third column of Table 2 shows business telephone lines relative to the level of employment in the state.
In the residential rankings, New York places slightly ahead of California and ahead of Illinois by a greater margin. New York's higher ranking suggests greater deployment of telecommunications infrastructure for residential use. In the ranking for business access lines, California is ahead of New York, though New York remains near the top of the list. New York is somewhat ahead of Illinois in terms of business access lines per job.
Table 3: International Measures of Infrastructure Stock, 1996
The international comparison shows that the U.S. has more access lines per capita than the other countries under consideration. (See Table 3.) The telephone density of the U.S. is particularly impressive given its low population density, since it is costlier to wire more dispersed locations. The high level of access lines per capita is due, in part, to the movement to achieve universal telephone access in the U.S. Taking the domestic and international data together, the New York area has done well in connecting its people to the public telephone network. [11] With New York near the top among U.S. states, and the U.S. at the top of the world ranking, New York's density of telephone access lines is among the highest in the world.
Fiber-optic cable and digital equipment
The telephone network has come a long way from the days of operators manually connecting wires into a switchboard to initiate a call, and telephone networks are constantly being upgraded. In recent years, the rising demand for new telephone services (voice mail, call waiting, Internet access) and the desire for greater efficiency has made it essential to upgrade both the wires and the switching equipment of the network.
Telephone companies have replaced old copper wires with high-capacity fiber-optic cables (made of glass) in the heavily trafficked parts of the telecommunications network. The advantage of fiber-optic cable is that it has a much higher capacity for transmitting data than copper wiring does. The interoffice cables-the lines carrying long-distance traffic and even cross-city traffic-are largely fiber-optic now. Slowly, some of the wires connecting users to the telephone company's central offices are being replaced with fiber-optic cable. Upgrading the users' connections is known as providing "fiber to the curb," since the fiber-optic network would then extend to the front door of the user's location. In most places, fiber to the curb is a long way off, since it requires the replacement of an enormous existing infrastructure. And, if technologies exist to allow high-speed data transmission across old copper wires, the need for fiber to the curb may diminish or vanish.
Telecommunications companies have also been replacing analog switching equipment with digital switching equipment. The advantage of digital switching equipment is that it handles the routing of calls more efficiently. Digital switching equipment also permits features like equal access for long-distance companies, call waiting, and voice mail. In large cities, a majority of the switching equipment is now digital; indeed, in some places around the world, switching equipment is exclusively digital.
These two indices of telephone network modernization are available at the state level. (See Table 4.) The first three lines of the table compare fiber-optic cable, expressed in different ways. The first is as a fraction of total cable mileage, to indicate the rate of network upgrading. The second is relative to land area, since the physical distance between a user and a fiber-optic cable affects the telecommunications service that is available. The third is relative to population, in order to reflect the possibility that inadequate fiber provision relative to population could result in over-use and congestion. In all three measures New York is ahead of both Illinois and California. This reflects the aggressiveness with which NYNEX has upgraded its network and results in a more developed infrastructure for New York's users.
Table 4: Indices of Local Telephone Company Infrastructure Modernization, 1996
Companies other than the ILECs (NYNEX, Ameritech, Pacific Bell) also have fiber-optic cable networks. Long-distance carriers and some competitive local carriers add to the infrastructure stock. An example is MediaOne's new fiber-optic network in Los Angeles; MediaOne is a cable television provider and subsidiary of the telephone company USWest. MediaOne plans to connect studios and other large businesses in Los Angeles. As well, there will be a direct connection to SoHoNet, a filmmakers network in London, to facilitate collaboration between the movie industries in the two cities. The Los Angeles investment is planned to cost $250 million. [12]
Still, it is the large local telephone companies that own most of the fiber in the country. These local telephone companies own 74 percent of the fiber-optic cable. Long-distance companies like AT&T, MCI, and Sprint own 17 percent, and competing local telephone companies like Teleport and MFS/WorldCom own the remaining 9 percent. [13] Thus, comparing the infrastructure of the local companies is a fair reflection of the overall infrastructure in each city.
The data on digital switching equipment also place New York ahead of its domestic competitors. In 1996 nearly all (95 percent) of the access lines in NYNEX's New York service region were connected to digital switches, including all of the access lines in New York City. [14] This was higher than the shares for Ameritech's lines in Illinois (83 percent) and Pacific Bell's lines in California (79 percent).
These findings are consistent with a recent report by NYNEX's state regulator. The New York State Public Service Commission found that the state's telephone network is at least as modernized as elsewhere in the country, in terms of fiber-optic cable, digital switching, and availability of high-capacity data lines. [15]
Internationally, Hong Kong, Singapore, and Finland have exclusively digital switching equipment, ranking them ahead of the U.S. generally and on a par with New York City (Refer to Table 3.). For Hong Kong and Singapore, these numbers reflect their relatively recent economic growth and modernization. The more recently a country built its infrastructure network, the more advanced is the technology it initially employs.
For fiber-optic cable, extensive international data are not available. A journalistic source indicates that Hong Kong has more fiber-optic cable (in absolute terms) than Germany and Japan, countries over ten times its size. And one-third of the homes have "fiber to the curb," making Hong Kong's fiber-to-total-wiring ratio near 50 percent, well above the 10 percent figure in the United States. Hong Kong's density contributes to its high level of fiber-optic cable, since it is cheaper to serve a given number of homes or businesses if they are in a single tall building (requiring one large cable) rather than spread across many lower-rise buildings. [16]
So, while New York leads other American cities in terms of network modernization, it lags some foreign locations-Hong Kong and Singapore, for instance-that have built entire modern networks more recently, rather than having to upgrade older networks.
Telecommunications investment
Another measure of telecommunications modernization is the amount of capital investment. While this measure is too general to reflect specific investments, it is consistent with the above evidence on specific kinds of network upgrading. (Refer to Table 4.) That is, domestically New York's rate of investment, nearly $118 annually per access line, is above the rate in California ($114) and Illinois ($103).
Countries vary more widely in their rate of telecommunications investment. (See Table 3.) The U.S. ranks last among the places under comparison. Two reasons are: (1) The better initial condition of the U.S. network (less new investment is needed than in Hong Kong, which developed recently), and (2) the tendency of the U.S. to save and invest a smaller fraction of GNP than other countries do. [17] The large gap between the U.S. and other countries suggests that even though telecommunications investment is higher in New York than elsewhere in the U.S., the New York level is probably still below the foreign cities in comparison. [18] These data are from 1996; the current Asian economic crisis may bring these investment levels more in line with the United States and Europe.
Telephone system competition
The Telecommunications Act of 1996 opened local telephone markets for competition. While the overwhelming majority of telephone customers across the country continue to receive service from the incumbent carrier (the ILEC), there is competition for some services, especially for larger business customers. Some industry executives suggest that the early effect of potential competition has been improved service quality, rather than a wider range of services or a large price reduction. [19] Thus, telephone industry competitiveness is an indicator, though crude, of the quality of telephone service-especially for business customers-in a city.
Within the U.S. there is variation in the degree of competition. New York was the first city to have telephone service competition; now, dozens of carriers compete with the local incumbent in New York and elsewhere. Some competitors, like AT&T/Teleport and MFS WorldCom, offer service in many cities with their national fiber-optic networks. These providers, who have built their own telecommunications infrastructure, are known as "facilities-based." Other competing carriers, known as "resellers," lease space in bulk on the local incumbent's network and provide telephone service over those lines. New York leads both Illinois and California in the number of competing local carriers, as well as in the number of facilities-based local carriers. Though Texas was among the states most resistant to competition initially, Texas now has more competing carriers than anywhere else.
Table 5: Competing Local Exchange Carriers Operating in Selected States, mid-1998
Among major countries, the U.S. has the second most competitive market behind Finland. There, as early as 1930, about 600 telephone companies offered service, many in direct competition with each other. Recently, Finland ended the formal licensing procedure for telecommunications operators, replacing it with a perfunctory registration process. [20] The competitive market in Finland is credited with bringing lower charges, high service quality, and a culture of technological innovation.
Other countries have committed to telephone competition, though not to the extent of Finland and the United States. The U.K. ended British Telephone's monopoly in 1991. Japan is breaking up Nippon Telephone and Telegraph, though NT&T is still largely state-owned. [21] Hong Kong and Singapore are opening up more slowly. Hongkong Telecom agreed in January 1998 to relinquish its monopoly on international calls, but local service remains a monopoly. Singapore Telecom retains its monopoly, though the government may offer a license to a competing company (possibly Hongkong Telecom) soon. [22]
Telephone service reliability
There is variation in reported service quality across telephone companies. Since these companies are geographically based, this variation means that telephone service quality depends on location. Table 6 summarizes complaints about the companies reported to the FCC in 1996, excluding slamming complaints. Slamming means illegally switching a subscriber's long-distance carrier without the subscriber's consent. Customers often cite the ILEC in a slamming complaint even though it may be the fault of the long-distance carrier, so these complaints do not reflect ILEC service quality. The remaining set of complaints is the basis of the table. NYNEX is second from the top of the list of most complaints, more than Pacific Bell and far more than Ameritech. Ameritech-the company serving Chicago and most of the Midwest-had a complaint rate less than one-third of NYNEX's.
Table 6: Complaints about Local Telephone Companies Reported to the FCC, 1996
More specifically, NYNEX had a poor record of serving business customers, relative to Pacific Bell and Ameritech. Table 7 gives the details. NYNEX succeeded in connecting 96 percent of its business customers promptly, compared to 99 percent for Pacific Bell and Ameritech. NYNEX's business customers had more troubles with newly installed lines and more complaints in general; they also suffered more unscheduled service outages over a three-year period.
Table 7: Service Indicators for Business Customers, 1995
Although the data in Table 7 put NYNEX behind the companies elsewhere, the figures still should be interpreted cautiously. Service quality is difficult to measure. Much of the data are reported by the companies themselves, and companies might define "complaints" differently. Also, the data from the FCC are nearly three years old, gathered before the initial impact of competition. Telephone competition is likely to improve service everywhere, especially in the more competitive markets like New York. A more recent ranking of residential telephone service quality, reflecting these changes, ranked Bell Atlantic sixth out of 13 large local telephone companies, ahead of the national average and of Ameritech. [23]
The variance in these rankings (USWest's good showing in the reported complaints, but poor showing in the J.D. Power survey, for instance) precludes drawing a strong conclusion. In the case of NYNEX, though, the company's service record was worse than both Pacific Bell and Ameritech in all three rankings. It is unlikely that this consistent finding, from three different data sources and methodologies, arose out of coincidence. However, more recent data suggest improvement, and New York City may no longer be at a strong competitive disadvantage.
Table 8: International Telephone Service Quality, 1996
Service in New York also may be better than in many foreign cities. (See Table 8.) While subject to the above caveat that definitions differ, the U.S. enjoys better service than anywhere in Europe, and worse than only Singapore and Japan. It is difficult to gauge where New York would fit on an international ranking, but one can infer that service quality in New York is better than most in international locations.
Cost of standard telephone service
The cost of standard telephone service matters for telephone calls and for Internet access, when access is through a dial-up modem. Again, the geographic coverage of most telephone companies implies that differences in telephone companies translate into geographic differences in telephone service.
There are many levels of business telephone service, and no one cost measure can summarize the range of possibilities that a business faces. The data presented here are intended to illustrate cost differentials for a standard service. These standard price differentials form the basis for the pricing arrangements that businesses make with telephone companies and therefore are suggestive of actual differences in telephone service charges.
For almost all of the services listed, New York is the most expensive of the cities under comparison. (See Table 9.) Monthly subscription charges and per call charges are higher in New York than in these cities, and initial connection charges for a single line are higher as well. However, connection charges for three lines are lower in New York than in Los Angeles and San Francisco. In all four cities the monthly subscription charge is below the national average, but New York's cost-per-call and connection charges are substantially higher than the national average.
Table 9: Business Telephone Rates in Selected U.S. Cities, January 1998
One factor mitigating the higher charges in New York is that New York's local calling area is somewhat larger than those in other cities. NYNEX officials argue that a higher proportion of calls made in New York are within the local calling area, so the average monthly business phone bill includes fewer toll calls in New York and is therefore more in line with phone bills in other cities. This argument, though, ignores an important element of the phone bills of small businesses; much of that bill may be for dial-up access to the Internet, which in large cities (where ISPs are numerous) are typically local calls. The phone charges incurred for Internet use are reflected by the cost of a local call. These charges are over one-third higher in New York than in Chicago, Los Angeles, or San Francisco.
Internationally, Singapore, Finland and Hong Kong offer the least expensive service. (See Table 10.) Finland's competitive telephone market has kept charges low. [24] Hong Kong's and Singapore's markets, in contrast, are dominated by near monopolies (Hongkong Telecom and Singapore Telecom) that have recently built efficient state-of-the-art telephone systems. The U.S. is costly relative to other rich countries, and since New York is expensive relative to domestic competitors, New York is therefore expensive relative to its foreign competitors.
Table 10: Business Telephone Charges in Selected Countries, 1996
For the Internet, the concepts of stock and flow of infrastructure have multiple dimensions. Urban areas can vary in the capacity of routers and cables comprising the physical system, the availability of leased circuits, the concentration of Internet domains, the range of Internet service providers, the speed of data transmission, and the cost of service. Available comparative data for each of these aspects of Internet service are presented and discussed below.
Routers and bandwidth
As described earlier, the Internet is, in large part, comprised of leased lines from the existing telephone network. Measuring this infrastructure involves measuring two elements: (1) the size (expressed in bandwidth) of the major leased lines (referred to below as "pipes") that are dedicated for Internet data transmission, and (2) the switching stations, or routers, that connect the largest pipes. A city's Internet infrastructure stock can be expressed in terms of the presence (or absence) of routers and the amount of bandwidth in major pipes that feed the city.
Routers are high-power computers that direct Internet traffic and allow Internet users on different networks to communicate with each other. The major pipes that constitute the Internet (the "backbones") connect at these routers. The location of these routers determines the physical path that data take-and therefore the speed-when data are sent across the Internet.
The presence of a router nearby results in faster data transmission, other things equal, and anecdotal evidence suggests that high-tech businesses do locate near routers, where data transmission capacity is highest. The transportation analogy works well here. The routers are like major airports with the bandwidth analogous to flights in and out. A frequent flyer wants to be near a large airport with non-stop flights going in all directions; similarly, a heavy Internet user wants to be near the routers and bandwidth.
There are 14 major routers in the U.S., most of which are privately built and operated (though some initially received National Science Foundation funding). Five of these routers are in the San Francisco Bay area, which reflects both the high demand for Internet services in the Bay Area and the ability of the private sector to meet this demand with new infrastructure. (See box.) Washington, DC, with three routers, has the major East Coast installations, followed by Chicago, with two. Los Angeles and New York have one each, as do Dallas and Houston. (See Table 11.)
Internet traffic need not travel through the 14 main public routers. As these main routers are becoming more congested, the major ISP's are establishing "private" routers where they can transfer Internet traffic with other ISP's at less congested nodes. They are "private" in the sense that they are not required to offer connections (for a fee) to all-comers. There are also local Ethernets, privately-owned (though sometimes publicly-supported) networks of cable that allow faster data transmission within a city. So the lack of a major router does not necessarily mean Internet isolation.
Table 11: Location of Major U.S. Internet Peering Routers, January 1998
Internet bandwidth, however, does follow the spatial pattern of the major routers. ISP's build their main cables to connect to these main routers; thus, a business located near a major router will also be located near many high-capacity fiber-optic cables. An approximate measure of a city's data transmission capacity is the sum of the inter-city bandwidth of the pipes operated by the major national ISP's. As shown in Table 12, the Bay Area leads in this measure as well. Chicago and New York follow. This is further evidence of the concentration of Internet infrastructure in the San Francisco Bay area.
Table 12: Intercity Internet Bandwidth for Selected Areas, December 1997
Routers are in place throughout the world. Tokyo has four major routers, London has two, and Helsinki, Hong Kong, and Singapore each have one. These foreign routers, for the most part, were established to connect domestic Internet traffic, rather than requiring domestic Internet traffic to "connect" in the United States. [25] Thus, these routers tend to support less bandwidth and less traffic than those in the U.S. Many of the ISP's serving these foreign cities are based in the U.S. and therefore peer primarily at U.S. routers. The number of routers in a foreign city therefore overstates the intensity of Internet activity there.
ISDN and leased-circuit availability
Internet infrastructure goes beyond routers and pipes. The infrastructure also includes the technologies that users require in order to connect to the Internet. Some users, including small businesses, connect to the Internet using a standard telephone line with a dial-up modem. Mid-size businesses and some home offices often use ISDN, which provides faster service that a standard modem, but uses the same copper lines as regular telephones. The other widespread Internet access option is a leased line, most commonly a T-1 or T-3 line. These leased lines offer continuous Internet access at much higher speeds than ISDN. Whereas ISDN service is most commonly provided directly from the telephone company, leased-line service is commonly provided by an ISP.
Within the U.S., ISDN is widely available. In New York, California, and Illinois, about 75 percent of access lines were connected to switches with ISDN capacity in 1996, and the figure likely has grown since. (Refer to Table 4.) That means that the infrastructure is in place to provide ISDN service to at least 75 percent of telephone subscribers. The actual number of subscribers is considerably lower, since a small fraction of potential subscribers actually choose to subscribe. [26]
Actual ISDN usage (and, implicitly, availability) varies widely by country. ISDN usage rates are higher throughout Western Europe, and lower in Asia, than in the United States. (See Table 13.) This reflects a preference for a particular technology in Western Europe rather than a general superiority in infrastructure. Business users in the U.S. rely more heavily on dedicated leased circuits like T-1 and T-3 lines. The U.S. market for leased circuits is over ten times the size of the next largest market, Japan.
Table 13: High-Speed Internet Usage in Selected Countries
The next generation in Internet access technology is the Asymmetric Digital Subscriber Line (ADSL), which offers extremely high-speed data transmission over existing copper wires using special modems. ADSL subscribers enjoy Internet access at speeds up to 7 megabits per second, 250 times the speed offered by a standard (28.8 kilobits per second) modem. As of winter 1998, ADSL was widely available only in Phoenix, but Bell Atlantic, USWest and other telephone companies nationally are planning to offer ADSL in other cities later in the year, at prices similar to ISDN service. Any conclusions about geographic variation would be premature. [27]
Internet domains
Internet hardware is one dimension of its infrastructure; another is the content that travels through the wires and routers. Speedy access on the Internet depends not only on where the network of wires and routers is, but also where the content is.
Furthermore, the builders of Internet hardware-the ISP's leasing and laying pipes, the firms building routers, the telephone companies introducing new access technologies-install new equipment where there is demand. Knowing where the Internet is used is the best guide in predicting where the next hardware improvements will take place.
Measuring Internet usage is much more difficult than measuring telephone usage. Internet technology breaks data into pieces during transmission, and it is prohibitively expensive to monitor the flow of these data pieces. It is necessary to measure Internet usage indirectly, and one way is by the number of Internet domains. An Internet domain is the address that identifies a location on the Internet. The domain ends in a prefix identifying the kind of organization it is: .com for firms, .edu for universities, .gov for government agencies, and so on. For each domain name there may be many computers that provide content over the Internet, and many people who use the Internet from that domain. So the domain does not indicate either the hardware or the number of individuals using the Internet exactly, but it is an approximation. Each domain name, like nytimes.com (the newspaper) or uchicago.edu (the university) is registered to a physical address to identify its location.
Table 14: Internet Domains in Selected U.S. Metropolitan Areas, January 1998
Although San Francisco is the Internet capital of the U.S. in terms of Internet hardware, New York leads in terms of domains. (See Table 14.) Los Angeles and the San Francisco Bay Area are also major concentrations. [28] Chicago, the other domestic competitor location, is a distant fifth.
Domains per capita measures the density of domains relative to population. The San Francisco Bay Area has a greater domain density than anywhere else in the country, due to the concentration of high-tech firms in Silicon Valley. New York's density is relatively low, ranking fourteenth among the 19 largest places.
At the international level, domain data are unavailable, but a similar comparison may be made with data on Internet hosts. [29] The United States is the world's main Internet user, with about one-half of the world's Internet hosts and users. (See Table 15.) In terms of hosts per capita, Finland leads the world. A higher fraction of Finns use the Internet than do the residents of any other country.
Table 15: Internet Usage in Selected Countries
Range of Internet service providers
Internet service providers (ISPs) offer subscribers access to the Internet. By leasing space on the network of telephone wires and cables, ISPs build smaller networks that connect to the larger Internet, allowing their subscribers complete Internet access. Some ISPs are national, like AOL and CompuServe; others specialize in business services and can build fiber-optic cables directly to large institutions; and others are small, local companies offering mainly household and small business services.
ISPs vary in the services they offer, both in terms of the kind of Internet connections they supply (dial-up modems, ISDN, leased lines) and in terms of the level of technical support and consulting. Table 16 shows the number of ISPs offering different kinds of Internet access in selected U.S. cities in December 1997. Businesses in New York have as wide a choice of ISPs as businesses elsewhere, with considerably more choice for the basic technologies of modem and ISDN. For very high-capacity Internet access-T3 access is the top of the line-Los Angeles offers slightly more choice than other cities.
The ISP market is more developed in the U.S. than elsewhere in the world. Businesses in the largest American cities have over 100 ISP's to choose from, while businesses in Hong Kong and Singapore have only a handful. The range of choice corresponds both to the size of the country and to the level of competitiveness in telecommunications services generally. The relative lack of choice in Hong Kong and Singapore does not, however, mean that businesses there are unable to get adequate Internet services, since the range of services available elsewhere is also available in those cities.
Speed of data transmission
One effect of the stock of infrastructure is the speed at which a business can send information over the Internet. Internet transmission speed is a function not only of stock, but also of demand and therefore congestion. One way to measure Internet speed is the time it takes to download the homepage from a website. The download time is the time it takes for a homepage to appear on the user's screen after the user requests to see it. It equals the sum of (1) the time it takes to send the request from the user to the server where the website is stored, and (2) the time it takes to send the homepage of the website (which probably includes graphics, text, and maybe video or sound) from the server to the user.
Table 17 presents average download times from 40 major websites, with measurements taken from cities in the U.S. and abroad. [30] From each city, the same set of websites were downloaded, and the cities were ranked according to the download speed. This list includes all of the comparison cities, domestic and foreign, except Helsinki and Singapore. The fastest download times are not necessarily in cities with the largest stock of infrastructure. While San Francisco has the most infrastructure in terms of routers and pipes, Internet usage is more than commensurately heavy there, so that website download times there are higher than the other cities under comparison. New York fares well in this comparison, ahead of the other cities except Los Angeles, where download times are slightly faster.
In this measurement there is a bias against foreign cities, because the measured websites are American
;){kind=link}
;){kind=link}
;){kind=link}
;){kind=link}
;){kind=link}
;){kind=link}
;){kind=link}
;){kind=link}
;){kind=link}
;){kind=link}
;){kind=link}
;){kind=link}
;){kind=link}
;){kind=link}
;){kind=link}
;){kind=link}
;){kind=link}
;){kind=link}
;){kind=link}
;){kind=link}
;){kind=link}
;){kind=link}
;){kind=link}
;){kind=link}