[1] Most of the data that compare companies refer to NYNEX, since data were collected before the company merger. Accordingly, "NYNEX" is used in the text even though the NYNEX service region is now served by the new Bell Atlantic.

[2] "A Connected World," Economist, September 13, 1997.

[3] "Mother Earth Mother Board," Wired, vol. 4, issue 12, October 1996. http://www.wired.com/wired/ 4.12/features/ffglass.html

[4] "Top 10 Discoveries about the Internet," Keynote, an Internet consulting firm. http://www.keynote.com/measures/top10.html

[5] "Old Champions, New Contenders," Economist, October 11, 1997.

[6] "The Orbiting Internet: Fiber in the Sky," BYTE, November 1997. http://www.byte.com/art/9711/sec5.art1.html

[7] "A Connected World," Economist, September 13, 1997.

[8] BYTE magazine, op. cit.

[9] "Phone Merger Aims to Control Electronic Pipeline of Future," The New York Times, June 29, 1998, and "At Last, Telecom Unbound," Business Week, July 6, 1998.

[10] Frances Cairncross, The Death of Distance (MA: Harvard Business School Press, 1997).

[11] In comparing the domestic and international data, note the apparent inconsistency. In Table 2, the FCC reports the U.S. level of access lines per capita is 580, while the ITU reports the U.S. level as 640. These numbers should be the same, but different data collection agencies use different measurements, and these methodological differences are not described in detail by the agencies. While it is safe to compare pieces of data within the same table, one should not make judgments about data from different sources.

[12] "MediaOne Plans Big Fiber-Optic Network," Reuters, February 9, 1998.

[13] FCC, Fiber Deployment Update 1996, Appendix tables.

[14] New York City data are courtesy of William Floyd, New York office of Bell Atlantic. In 1998 all lines in New York State were served by digital switching equipment.

[15] State of New York, Public Service Commission, "New York Telephone Construction Expenditures," September 17, 1997. The report's findings are based primarily on construction budgets submitted by New York Telephone and other companies.

[16] "Chan on Demand," Economist, March 21, 1998.

[17] Though Japan and Hong Kong invest far more in telecommunications per capita than the U.S. does, the gap is much smaller when expressed as a percentage of total investment (gross fixed capital formation). The ITU reports that Japan's telecommunications investment is 2.4 percent of GFCF, Hong Kong's is 2.3 percent, and in the U.S. it is 1.9 percent. Telecommunications investment relative to GFCF is highest in Finland, at 4.2 percent.

[18] However, the New York investment per access line in Table 3 ($118) should not be compared to the national-level data in Table 4. The New York value refers only to NYNEX investments, while Table 4 refers to investments by all companies, including long-distance and competing local companies.

[19] Conversations with Gail Schwartz, Teleport Communications Group, and Steve Ingish, Metropolitan Fiber Systems (January 1998).

[20] "A Connected World," Economist, September 13, 1997. The United States also had hundreds of independent local telephone companies at the start of the century. What makes Finland unique is that its competing phone companies never underwent the thorough consolidation that the American companies did.

[21] Organization for Economic Co-Operation and Development, Communications Outlook 1997, p. 25.

[22] Economist, op. cit., March 21, 1998.

[23] J.D. Power and Associates, Residential 1998 Local Telephone Service Satisfaction Study, August 6, 1998.

[24] OECD does not report data for Hong Kong and Singapore. Data from the International Telecommunication Union indicate that Hong Kong and Singapore offer very inexpensive telephone service for business subscribers. While it is difficult to compare data from different sources, the ITU data suggest that telephone tariffs in Hong Kong and Singapore are even lower than in Finland.

[25] Information Sciences Institute (at University of Southern California). http://www.isi.edu/div7/ra/NAPs/

[26] The gap between ISDN capacity and ISDN subscriptions is quite large. While ISDN capacity runs around 75 percent of access lines in the U.S., subscriptions amount to around 1.5 percent of access lines. The variation in ISDN subscription varies somewhat within the United States. Measured by the number of ISDN B-channels per switched access line, New York, California, and Illinois are all above the national average. The rate is 1.96 in New York, 1.72 in California, and 1.66 in Illinois.

[27] "USWest Communications In High-Speed Internet Plan," Reuters, January 30, 1998.

[28] The domain data refer to metropolitan areas. "New York" includes the city, Long Island, and parts of upstate New York, New Jersey, and Connecticut. "San Francisco" refers to the entire Bay Area, including Silicon Valley.

[29] A domain refers to the name assigned to an institution. A "host" is a computer that is connected to the Internet. Many hosts can exist under the same domain. International data and domestic data are collected by different agencies using different methods. Thus, it is impossible to translate hosts into domains (or vice versa) for making direct comparisons. It still is meaningful to look at rankings.

[30] These unique data are part of a service offered by Keynote, an Internet consultancy that helps businesses track the accessibility of their websites.

[31] "Results of the Great Web Race Survey," Timedancer consultancy, 1996. http://www.timedancer.com/guide/results.html

[32] Keynote and Timedancer Systems both warn that their data are collected from computers that may have different kinds of connections to the Internet. That means that their comparison might not be across identical agents in different cities, which would be the ideal test. But they do claim that their comparisons give a sense of the relative quality of Internet service in different cities, and reputable trade journals like Boardwatch publish their data.

[33] "Invasion of the Stealth Antennas," The New York Times, March 26, 1997. Also see the Cellular Telecommunications Industry Association website: http://www.wow-com.com/-professional/siting/index.cfm

[34] The domestic city data and the international data should not be compared directly. The data come from different sources, using different methodologies, and they cover different time periods. To compare New York and the U.K., for instance, the domestic data suggest that New York is at the high end of American cities and therefore above the U.S. average; since the international data show that the U.S. and the U.K. have similar average charges, New York's costs are almost certainly higher than those in the U.K.

[35] According to the International Telecommunication Union's World Telecommunication Development Report 1998, in the U.S. two-thirds of households with televisions also subscribe to cable. The share is 40 percent in Finland, 30 percent in Japan, 15 percent in Hong Kong, 10 percent in the U.K., and under 5 percent in Singapore.

[36] Other cities have turned to the public sector to build telecommunications infrastructure. The largest public investment project is in Tacoma, Washington, where Tacoma City Light is building a $100 million fiber-optic cable system. Already the local telephone and cable companies have responded by upgrading their own networks, making it unlikely that Tacoma City Light will be able to compete and recover its costs.

[37] Paul Krugman, Geography and Trade (MA: MIT Press, 1991).

[38] "Offices Plugged in and Ready to Go," The New York Times, February 4, 1998.

[39] At the recent San Francisco Multimedia Summit, multimedia firms listed the kinds of assistance they would like from the public sector. Telecommunications infrastructure improvements were low on the list, far behind lower-tech concerns like parking, transportation access, lower rents, and tax/regulatory relief. "Cozying up to Multimedia," San Francisco Chronicle, February 27, 1998.

[40] Peter Huber, "Gotham's Hidden Infrastructure Boom," City Journal, Spring 1998.

[41] Author's calculation based on County Business Patterns 1995. The San Francisco Bay Area is followed by Washington-Baltimore, New York, Boston, and Los Angeles.

[42] A good history of the Bay Area's advantage in the computer industry is Annalee Saxenian, Regional Advantage: Culture and Competition in Silicon Valley and Route 128 (MA: Harvard University Press, 1996).

[43] Telecommunications providers, like all other firms, pay wages and rents that are set in the market. Wages and rents are higher in New York, and under competitive conditions those high input costs are passed onto consumers as higher prices.

[44] The company has the option to extend the plan for two additional years under certain conditions, including annual rate reductions valued at $25 million in each of the two years.

[45] BA-NY is only allowed to increase its rates for reasons of inflation and "exogenous" factors, and only if these factors exceed $6 million. (An exogenous factor is an external event, such as a regulatory decision, that causes BA-NY's costs to increase.) Thus far, BA-NY has not implemented any rate increases under the plan.

[46] New York Telephone Company, First Plan Year Service Quality Report, State of New York, Department of Public Service, November 6, 1996.

[47] New York Telephone Company, Fourth Quarter 1997 and Calendar Year 1997 Service Quality Report, State of New York, Department of Public Service, January 22, 1998.

[48] Communication with New York State Department of Public Service.