================================================
     EEEEE  N   N    OOO    DDDD    EEEEE
     E      NN  N   O   O   D   D   E
     EEEE   N N N   O   O   D   D   EEEE
     E      N  NN   O   O   D   D   E
     EEEEE  N   N    OOO    DDDD    EEEEE
================================================
To subscribe to this newsletter on information technology and society, send the message "subscribe e-node" to enodelist@garnet.berkeley.edu. To signoff the list, send the message "unsubscribe e-node" to unsubenode@garnet.berkeley.edu
        Vol 2, No. 4
        July 1997
THE HYPOCRISY OF ISP WELFARE AND
THE MYTH OF THE CYBER FREE MARKET
by Nathan Newman, Progressive Communications,
newman@garnet.berkeley.edu


With the Clinton administration's announcement of a drive for a "free trade zone" in cyberspace, it might be the time to ask how long we are going to keep the ISPs and other Internet corporations on welfare? And how long to we have to hear hypocritical drivel about the success of the "free market" of cyberspace even as those engaged in the hype lobby for continued subsidies and government regulation that benefits them.

    Since the privatization of management of the Internet from 1992 to 1995, the industry around the Internet has been trumpeting their success as proof of how unregulated market competition had helped explode the size of Internet participation. No players have done more to trumpet the success of this new "free market" than the independent Internet Service Providers, or ISPs in the incessant lingo of the industry.

    From veteran Whole Earth Networks to upstart Netcom to giant American On-line, these Internet providers not only beat back proprietary networks like Microsoft's initial foray, they delivered to their customers (local businesses and mostly upper-income individuals) an unlimited "all-you-can-eat" flat-rate price for service that made the high prices for long distance phone service seem laughable in the face of the new technology. Internet phone calls, made essentially for free over the Net, began to bypass traditional long distance phone services and the Net seemed to promise limitless connections at a price the mastodons of the old regulated phone system could only dream about.

    FCC decisions in May 1997, however, would undermine the "free market" bravado of the ISPs as these Internet free-marketers made loud, extremely public appeals for the Federal Communications Commission to protect them from market prices in order to "save the Internet" (and their own profit margins.)

    The ISPs along with AT&T, Apple Computer, Netscape, Microsoft, Compaq Computer, IBM, and a host of other computer companies demanded and won continued FCC intervention to prevent market pricing on local telephone company services used by ISPs to reach their customers in the first place. Since the initial breakup of AT&T back in 1983, the FCC has exempted Internet providers from paying the same kind of per-minute access charges to local phone companies that long distance companies have to pay to connect their customers. This has allowed Internet providers to pay the flat business rate to local phone companies that ordinary local business customers pay--which in turn has allowed them to offer flat-rate service for the Internet to their customers.

    What this means is that in connecting a customer to an Internet provider, FCC-regulated payments by ISPs to Pacific Bell, as one example, average only $0.00073 per minute (less than one-tenth of a penny) versus $0.014 per minute paid to the local phone company for handling connections to a long distance carrier--a 95% discount for Internet usage. This is all despite the fact that the costs for handling each kind of call on a per-minute basis are exactly the same for the local phone company.

    Even worse for the local phone companies is the fact that Internet calls average much longer than either local or long distance phone calls. According to Pacific Bell, 30% of the total time of customers' use of the phone system generated by dial-up Internet traffic comes from calls lasting 3 hours or more and 7.5% came from calls lasting 24 hours. This compares to an average voice call lasted only 4 to 5 minutes. Pacific Bell cites one Silicon Valley ISP hub where traffic levels in late 1996, driven by a single ISP, undermined service in the whole area. The ISP represented only 3.6 % of total office lines, but accounted for about 30% of use during the busiest hours of the day. The result was that 1 out of 6 phone calls were being blocked due to the congestion. Pacific Bell claims it spent $3.1 million cost to fix that one hub alone, and estimates that it will have spent $100 million on Internet traffic upgrades in 1997 and will spend $300 million on upgrades by the year 2001--while Pacific Bell maintained it would earn only $150 million from additional revenues in that period due to ISP traffic.

    Now, you don't have to take Pac Bell's numbers as gospel to recognize that with phone rates designed by regulators to yield minimal profit on basic service (with profit to be made on toll calls and extra services), long local Internet calls are a disproportionate drain on resources with little additional revenue, thereby sucking investments away from the rest of the network. Defenders of the ISP subsidy argue that local Bell companies benefit by the addition of dedicated Internet phone lines, but this is a bit like arguing that what the phone companies lose in costs on each Internet line, they can make up in volume.

    Worse than the actual costs of the upgrades for ISPs is the fact that those investments are being made in traditional analog voice phone lines and switches, instead of the phone system moving the ISP phone traffic onto high-speed digital switching systems right at customers homes, an approach that would be more efficient and create the basis for upgrading all data traffic. Most of the Baby Bells began offering such high-speed digital services for ISPs in 1997, but the Internet providers have little incentive to pay for such services as long as they can convince the FCC to allow them to use the local phone lines like ordinary business users.

    And in May 1997, the FCC, under intense lobbying from both computer companies and Internet users, agreed to continue the ISP exemption from access charges, with essentially minor concessions given to the local phone companies in raising all charges on second phone lines, the logic being these would likely be used for Internet connections. Some of the smaller Internet providers complained that the additional charges on all their incoming phone lines would hurt them, but larger ISPs like America Online declared victory: "We will see an increase in our charges, but we do see that on balance we need to accept the additional charges because they are flat and they are nominal," said Jill Lesser, America Online's deputy director of law and public policy. "A permanent access charge would have been orders of magnitude worse for AOL. Even at one cent per minute, we would have incurred a charge that would have been in the neighborhood of $100 million and which we would have had to pass on to the customer. So when you look at an increase that is 1/10 of that, that's a fairly modest increase." The broad coalition of computer companies had successfully protected the subsidized status of Internet providers.

    The irony of the whole decision is that the Internet industry has pictured the privatization of the Internet as the "end of government subsidies" where the free market had successfully stepped into the gap. The reality, as this decision highlighted, is that the profits of the private Internet industry have derived substantially from the cannibalization of past and present investments in the local phone infrastructure. Local phone users, mostly lower-income users without a computer in the home, are seeing investments diverted to industry and higher-income Internet users that could have been targeted for upgrading the overall network or delivering new technology for schools, hospitals or other public places serving the whole public. Instead, the specific private subsidies for the Internet industry have helped fracture planning for the overall local phone system and blocked the general upgrading of data traffic.

    Where federal investments and regulations once fueled overall economic and technological advancement in regional telephone networks, these new "market competition" policies end up sucking funds from the infrastructure serving low-income and local users to subsidize those using the Internet for national and international purposes. And the forced segmentation of "competition" into their own boxes of long distance, local service, ISP and other regulated divisions have so fragmented phone service as to make comprehensive investments for upgrading the overall system nearly impossible.

    Now, if this had been a small sin to help the Internet get off the ground, it might be a minor, even admirable hiccup in regulatory history, but this is the pattern dating back to the first attacks on the integrated AT&T Bell system. And with competition and "deregulation" of telecommunications becoming the metaphor and model for other network-based industries like electricity, it is worth understanding that the original success of MCI and Sprint was based not on being more efficient than AT&T but on regulatory subsidy and infrastructure cannibalization. MCI became a billion-dollar company based on the FCC forcing the Bell System to give it access to its networks, even as MCI was paying little to help maintain that local infrastructure that made its business possible. MCI and Sprint were paying only half the charges than AT&T's own long distance service paid to maintain local phone service - meaning that MCI and Sprint had what amounted to an automatic 20% price discount in offering long distance service. Contra the mythmaking of history, these "free market" successes were free riders who could literally be less efficient than AT&T yet, thanks to regulations supposedly supporting the "market," have cheaper prices.

    The fact is that competition on day-to-day prices undermines long-term investments in infrastructure that have historically been served better by regulated monopoly. Many proponents of competition pooh-pooh concerns over investments in telecommunications infrastructure, noting that in the early decades of this century, full-throated competition led to a massive expansion of phone service across the country.

    Which is true. But.

    After AT&T's original patents derived from Alexander Graham Bell expired in 1893, full-scale warfare broke out between the Bell system and 3000 independent phone companies to compete in building infrastructure across the country, with AT&T retaining only half the market of a vastly expanded 6 million phones by 1907.

    But it was an infrastructure that frustrated most of the customers, since they could not call friends in the same city if they belonged to competing networks and would be unable to call whole cities if those towns were controlled by networks hostile to the hometown service. The Bell system was the only service that provided anything approaching a comprehensive long distance phone network. For the rest, competition made most of that expanded infrastructure unavailable across lines of hostile businesses--a state that led to pressures towards consolidation and regulated utilities. As AT&T began to also purchase other phone companies, AT&T officials reversed Bell policy and accepted government regulation of the industry in order to maintain high- quality technology and uniform pricing.

    A 1913 consent decree with the Justice Department officially put AT&T purchases of other phone companies under the regulation of the government and required non-Bell companies to be connected into AT&T phone lines, all in the context of negotiated agreements that turned AT&T and the independents from competitors to collaborators in maintaining the phone infrastructure. State utilities commissions strongly supported the movement to consolidation and 1921 federal legislation, the Willis-Graham Act of 1921, placed AT&T under the jurisdiction of the Interstate Commerce Commission (whose jurisdiction was handed to the FCC in 1934) and exempted it from antitrust restrictions on purchasing other telephone companies. AT&T would purchase 223 independents in the next thirteen years.

    Latter-day market advocates argue that all that was needed were regulations requiring mandatory interconnection between services, and the country could then have preserved the benefits of both competition and interconnection (much as is promised today with phone and Internet competition). The problem with this retrospective viewpoint (and present advocacy) is it ignores the basic economic implications of Metcalfe's Law--the rule-of-thumb that the value of a network increases not arithmetically but geometrically with the number of participants in that network. What this means is that the economic value of interconnection for small networks to much larger systems is astronomically high, while the main value of the investments in infrastructure by large networks is precisely the fact that they can offer such a large geometric network value where smaller networks cannot. Mandate interconnection and much of the value of that larger network's infrastructure (and the incentive to create it in the first place) disappears.

    Regulation of customer phone service rates may be eliminated under "deregulation" but government regulation is still pervasive in establishing the rates paid for interconnection between different business networks, an intervention that will either be too high to encourage new entrants to the marketplace or, more likely given larger networks' preference for no interconnection (i.e. an infinite price), result in a price set too low for the larger network to maintain the quality and breadth of its infrastructure for all users. In such a situation, the most profitable position is to be a smaller network servicing high- income, high-profit individuals or businesses who can, as needed, reach the low-profit customers of the larger network due to mandatory interconnection regulations.

    This is the position of cannibalization where Internet providers are presently positioned. A number of services like Netcom have already begun working to concentrate their customer base in higher-income and business users to maximize their profits. Others will continue to use subsidies from the local phone companies to expand their customer base but that will last only as long as the subsidies from the local phone companies continue; at that point, low-profit customers will be dropped fast.

    What is lost in this whole system of ISP welfare is any broad planning to assure that all citizens will have access to the next generation of high-speed connections to telecommunications services. Over twenty years ago, AT&T began moving towards converting the whole phone network to high-speed digital connections but the breakup of the Bell system undermined that planning, leaving the US with the same old analog connections leading to the home. The fracturing of the Internet due to privatization has led to the "World Wide Wait" we all love and cherish where no comprehensive planning is possible to give Internet users the same instantaneous "dial tone" connections we once took for granted on the old integrated Bell phone system.

    It is no coincidence that the enthusiasm for "markets" in the Internet is promoted by those receiving this welfare and regulatory support, but what is disturbing is how many ordinary Internet users have bought the myth (also known as the lie) that the Internet's recent expansion was based on the "free market" rather than based on government policy. Government design and subsidies created the Internet and government-mandated subsidies from local phone companies to ISPs have been the heart of its expansion.

    This all may be wonderful use of government power, but let's all remember the government's role now when those receiving welfare today cry "market competition" tomorrow in trying to block government mandates and spending to assist lower-income working families in getting access to the system.


ENODE: to loose, untie a knot; to solve a riddle.
E-NODE is a monthly column about the Internet. To subscribe to this newsletter, send the message "subscribe e-node" to enodelist@garnet.berkeley.edu. To signoff the list, send the message "unsubscribe e-node" to unsubenode@garnet.berkeley.edu
ENODE is brought to you by Progressive Communications, a policy research and computer consulting firm.