In the Matter of the Residential		)
ISDN Service Offering of			)  Case No. PUC961696 
Bell Atlantic, New Jersey, Inc.			)


February 7, 1997

Q1. Please state your name and business address.

A1. My name is James Packard Love. My business address is P.O. Box 19367, Washington, DC 20036.

Q2. Please summarize your educational qualifications.

A2. I obtained a Masters of Public Administration from the Kennedy School of Government at Harvard University in 1981. In 1985 I received a Masters Degree in Public Affairs from the Woodrow Wilson School of Public and International Affairs at Princeton University. At Princeton I was in the program in Economic Policy Analysis.

Q3. By whom are you currently employed and what is your current position?

A3. I have been employed by the Center for Study of Responsive Law (CSRL) since 1990. At the Center I am the Director of the Consumer Project on Technology (CPT), the Taxpayer Assets Project (TAP) and I am also Director of Economic Studies.

Q4. Please briefly describe your previous employment history.

A4. From 1988 to 1990 I was employed as Senior Economist for the Frank Russell Company, a large pension fund consulting firm. In this capacity I evaluated large real estate deals for pension fund clients. I developed a portfolio reporting system used by IBM to track $2 billion in real estate investments, and a return attribution system used by Digital to evaluate a complex portfolio of hybrid debt and equity instruments. I also worked on the development of new products and services. Prior to the Frank Russell job I spent one year, from 1987 to 1988, as Visiting Lecturer at the Rutgers University Business School, where I taught the core microeconomics course. From 1986 to 1987 I was Lecturer and Research Associate at Princeton University. While I was in graduate school I was employed in a research capacity for the National Bureau of Economic Research. I have also worked for the State of Alaska Department of Revenue, and as a consultant on a number of projects. Prior to Graduate School I was Executive Director of the Alaska Public Interest Research Group (AkPIRG). In Alaska I was a lay practitioner before the Alaska Public Utilities Commission (APUC), the Alaska Pipeline Commission (APC) and the Federal Energy Regulatory Commission (FERC).

Q5. Briefly describe your work at the CSRL.

A5. Over the past six years I have worked on economic regulation, intellectual property rights, antitrust, problems and opportunities presented by new technologies, international trade agreements, and other areas. For the past two years I have spent a significant portion of my time on telecommunications regulation, including problems with the high prices charged for residential ISDN services.

Q6. What is the purpose of your testimony?

A6. The purpose of my testimony is to describe the public policy considerations that should guide the Board in setting the BA NJ residential ISDN tariff, to address the issue of network congestion, and to set out a reasonable pricing methodology.

Q7. What are the important public policy considerations for residential ISDN pricing?

A7. The reason ISDN pricing is important is the Internet. The explosive growth of the Internet is the most important development in communications since the deployment of television. Like every new information technology, it can be described through analogies, but it has unique characteristics which set it apart. The Internet is a communications technology that is fully interactive. The intelligence of the network is widely distributed. Personal computers are used to navigate the Internet, and to run Internet applications. The Internet provides an "open" platform for modern communications. It has fostered both a vast sea of non-commercial communications, and a highly competitive commercial sector.

A few comparisons to television are useful. While television is a very effective medium for broadcasting information to mass audiences, the Internet has excelled at facilitating communications between individuals, and as a platform for publishing information to more focused interest groups. While it is costly to broadcast television programming, it is very inexpensive to publish information on the Internet. The lower barriers to entry for Internet publishing allow virtually all Internet users to be both consumers and producers of information. While some critics fear that television has replaced reading, the Internet has emphasized the written word. While television is considered by many to be a highly manipulative medium, the Internet has created an environment for debate and sharing of information that is decidedly less manipulative, and far more democratic.

It is also important to recognize the enormous pace of innovation in Internet communications tools. Graphical Web browsers like Netscape or the Microsoft Explorer were only developed a few years ago, and new versions are introduced every couple of months with new features. RealAudio and other audio broadcasting technologies, including Internet telephony, are being introduced or improved at a dizzying pace. Searching sites like Yahoo, Lycos, and AltaVista, to mention just a few, are springing up every day. Java and other new programming languages are making it easier to create and distribute new programs. Free online dictionaries and thesauruses, Web sites that draw street maps of any U.S. address, user-designed custom "newspapers," and countless other innovations are being tested and introduced far more rapidly than ever before.

Many public officials, experts, and citizens recognize the importance of the Internet as a model for communications. Not only is the Internet a dynamic platform for successful new information services, but it addresses important social needs. The blind, the hearing impaired, and others have found the Internet a vast resource for communications that are more accessible than older technologies. The Internet is a useful tool that can also be used to make government more open and responsive to ordinary citizens. For example, consider the vast increase in the number of citizens who can browse the Congressional Record or read copies of legislation from THOMAS.LOC.GOV, one of several Congressional Web sites.

Q8. What about distance education?

A8. Distance education may revolutionize the education system. Many experts are talking about new "just in time" education systems, where persons can find and use training or educational materials on an "as needed" basis for a wide range of applications. My wife is currently working with a university in Malaysia which is using the Internet as a platform to expand educational opportunities in that country. Traditional distance education programs have relied upon video cassettes, audio tapes, and printed materials. The Internet makes it possible to introduce real-time lectures, have students interact with their peers, and significantly broaden their access to learning materials, very economically. The Clinton Administration is talking about a major effort to fund the development of new Internet distance education "tools" for the United States. This is a field that is rapidly developing.

Q9. What does all this have to do with ISDN?

A9. The problem in the United States - and everywhere else - is the limited bandwidth available for residential consumers at a price they can afford. At present, applications are being developed that can work on analog modems, but innovation is severely limited by the slowness of modem access. The problem that government officials have to address is this limited bandwidth. Indeed, the problem of limited bandwidth for residential consumers was the subject of a recent day-long forum before the Federal Communications Commission (FCC), at which I was an invited panelist. My comments in that forum are attached to this testimony. (Exhibit 1 contains a prepared statement as well as excerpts from the FCC transcript).

In the beginning of the Bandwidth forum held in January 1997, FCC Chairman Reed Hundt made some opening remarks. It was interesting that Chairman Hundt spoke extensively about the views of leading computer and software companies at a forum designed to address local exchange telephone company policies. A transcript of the forum is available on the Internet at: (Excerpts of which are attached as Exhibit 2). Here are some of Mr. Hundt's comments:

Bill Gates [Microsoft] and Andy Grove [Intel] had a fascinating dialogue in which Andy -- and Les is here to explain this today I hope. Andy said, if I remember correctly, the following. And I'm remembering this by reading the quote. "As exciting as the Internet is though, there's one big problem. Telecommunications bandwidth. And Gates responded, bandwidth bottlenecks. No question. That's the biggest obstacle to where we'd like to take the PC." We actually found this particular colloquy to be immensely interesting and we are curious to know whether it is a synonym for a colloquy that we are having here in public in a variety of proceedings. In our colloquy, we talk about the exclusive facility bottleneck problem in communications. We talk about whether or not the last mile of the local loop, the spread of wire from the consumer to some centralized aggregation point. We talk about whether or not that is a bottleneck. We talk about that in antitrust terms. We talk about it in competition terms. We also talk about whether the current local exchange market is a bottleneck. That is the key issue in access reform. When we're having that conversation, what we'd like to find out today is are we also having a conversation on the same topic that Andy and Bill were talking about. Do they mean the same thing that we mean? If we don't mean the same thing, let's find out the difference. But if we do mean the same thing, then it should be the case that this obstacle to where we'd like to take the PC in fact is the topic that we are spending hundreds, indeed thousands of hours on right here at the FCC. And so now we've discovered that maybe our traditional public policy pursuits are directly related to the future of our information economy in an even bigger way than we may previously have thought was the case.

At the FCC bandwidth forum, there was a presentation by MR. Stagg Newman from Bellcore, in which he addressed many of the same issues:

MR. NEWMAN: Thanks, Bob and Elliot. I'd like to thank you for inviting me here . . . . What I'd like to talk about today is the bandwidth bonanza and basically there's no question in my mind that there will be a tremendous band width bonanza for corporate America. We will see the same type price performance improvements we're already seeing, that we've seen in the computer world. That is a doubling every -- you can argue how many months, 18 months, 12 months, 24 months, of the price performance characteristic for obtaining bandwidth. But what happens to the mass market? The mass market is the key issue I think and that's a real challenge for all of us, both from a technology side, from a regulatory side and from an operations side. Said another way, Bill Gates was quoted recently over the Internet, of course, or misquoted, as saying in the year 2000, 90 percent of all the users of the Internet will still access it over dial up modems over the telephone network at 28.8 kilobytes per second or in reality usually less for those of you who have been on certain networks recently. Unfortunately, our analysis is he's probably right.

On a similar theme, Mr. Newman said:

Now, what are the obstacles to getting there? At the risk of offending many very good engineers who are trying to solve difficult technical problems, I think the critical problem is the mass market access. The Ciscos, the three coms, et cetera, of the world will make their routers and switchers get faster and faster. In fact, a Cisco router circa 1997 will have ten times the band width of the largest voice switches in the networks today. So the backbone switching technology will get faster. The backbone fiber optic technology we and others are working on will give us a 20 to 100-fold increase in that very shortly through wavelength multiplexing. We'll solve the traffic congestion problems. They are challenging problems, but those are being solved. The competitive marketplace will solve those. The real problem is how do you solve the mass market access? And that's what we'll spend the rest of the time looking at. What are the options today? Today you've got the Internet backbone network, Internet access points and ISPs. How do you get on and off the highway? Well, if you're a corporate user or if you don't mind paying $200 a month, you can get a private line from your home or your office into the network. And the private lines are getting faster, the price performance is improving all the time. So again, no problem here as long as you're willing to pay a fair amount of money. But if you're not, what are the alternatives? POTS? You know, 28.8 bit per second, limited to 64 kilobit, but in many applications those won't really be anywhere close to that. ISDN which can get you up to 128 kilobits per second. They both have the advantage that they work over today's copper network just by putting something into the copper loop. Then there's a whole family of what are called digital subscriber lines, so-called XDSL, ADSL, VH or Very High speed digital subscriber lines, et cetera. And I'll talk more about those.

Mr. Newman's talk, and subsequent presentations by experts from the local exchange telephone, cable, wireless and competitive access providers made it clear that the among the alternatives to POTS described by Mr. Newman, only ISDN was ready now for large scale deployment. Mr. Newman described the severe wiring problems with cable networks, and the high cost and immature state of ADSL deployment.

Mr. Newman also explained to the FCC how cable modem technology is likely to be much slower in practice than the high "burst" speeds that have been widely discussed. Mr. Newman showed a video of a cable modem, which uses shared bandwidth. When operating without network congestion, the cable modem downloaded a file in 10 seconds. With network congestion, the download took 4 minutes (See Exhibit 2).

At some time in the future, maybe 5 to 10 years from now, we will see some non-trivial deployment of cable networks, ADSL, or wirelesses systems for faster connections to the Internet. But for most people, the only practical solution in the near term is BRI ISDN. Policy makers should look to the future, but they must also make the present work better.

ISDN offers vast benefits, if it is deployed at a price anywhere near its actual costs. Apparently, ISDN can be deployed to nearly every home in the BA region, right now. It will take many years before most BA residents will be able to get anything better than ISDN. We could wait for alternatives, but there are substantial costs associated with delays.

The market for information technologies is international, and the United States is competing with other counties to define and shape a new industry. Five years ago the Internet looked nothing like it does today. Five years is a eternity on the Internet. We shouldn't spend the next five to 10 years using analog modems if there is a better digital technology that can be deployed cheaply today.

Unfortunately, most incumbent local exchange carriers have priced residential BRI ISDN services far above costs, and have practically killed the residential ISDN market. We are beginning to see funny responses. Analog modems, which operate at about one fifth the speed of a residential BRI line, are selling like hotcakes because residential consumers cannot afford the high residential ISDN tariffs. Microsoft, which is a company that takes the bandwidth problem seriously, is reportedly modifying Windows NT to allow users to "bond" several analog telephone lines together to get more bandwidth. Users will be able to bond several POTS lines together to get the speed of a single BRI ISDN line. This is incredibly wasteful, from a social point of view, but it makes sense to the users because the incumbent LECs have priced ISDN lines so highly.

Q10. Have all LEC's priced ISDN excessively?

A10. There are huge differences in the prices that residential consumers are being asked to pay for ISDN. Some independent telephone companies' prices are close to POTS rates. For example, the Northern Arkansas Telephone Company (NATCO) charges $10.90 per month for POTS, and $17.90 per month for residential BRI ISDN - a difference of $7 per month. In California, the Roseville Telephone Company charges $20.35 for POTS, and $29.50 for residential BRI ISDN - a difference of $9.15. Neither company asks for per-minute usage charges.

Ameritech charges between $28.05 and $34.50 per month for residential ISDN in Illinois, Wisconsin, Ohio and Michigan, without per-minute usage fees, and $94 per month plus per minute usage charges for Indiana.

In states where state regulators have protected consumers, prices are also much lower. For example, in Tennessee, where a member of the state regulatory agency was reportedly an ISDN user, ISDN is priced at $13 over the POTS rate, with no per-minute charges. In other Bell South States, the flat rate price of residential ISDN is $60 to $80 dollars - far higher than Tennessee, but still a fraction of the BA flat rate offering for New Jersey.

In Texas, New Mexico, Delaware and Utah, state regulators have sought rates below those which the LECs wanted to offer. In California, PacBell charges $13.50 for POTS, plus $7 more for BRI ISDN. PacBell gets 1.05 cents per minute in usage fees per B channel, from 8 am to 5 pm, but off-peak hours are unmetered. PacBell wanted to impose off-peak metered rates, and a higher monthly fee, with a 20 hour allowance for off-peak usage, but faced strong resistance from ratepayers and the California Public Utilities Commission. It now appears as though the California Commission, if it affirms the ruling of the administrative law judge, will approve a monthly residential ISDN tariff of less than $30 per month with a 200 hour allowance for off-peak usage. This is far below the pricing of the comparable Bell Atlantic offering. In Utah, the PSC recently approved a residential ISDN call pack system for US West that is similar in design to the Bell Atlantic offering, but much cheaper rates were decided upon after a full evidentiary hearing. The 50 hour option was $30, the 100 hour option was $40, the 200 hour option was $54, and the flat rate was $64. The Bell Atlantic flat rate offering is nearly four times as expensive as the rate proposed by the PSC for Utah. US West is appealing the Utah ruling.

Among states with flat rate options, Bell Atlantic offers the highest flat rate in the United States. When looking at the various call-pack options, Bell Atlantic finds itself far more expensive than the US West call-pack options.

In general, however, it seems as though many state regulators have not given residential ISDN much thought in non-contested rate hearings; and we caution state regulators from reading too much into the early pricing comparisons. In general, we think that the premium for residential ISDN should be based upon the long run incremental cost of the service.

Q11. What evidence is there about the cost of ISDN?

A11. I haven't examined the BA cost studies, but we do know that BA has filed several ISDN Centrex tariffs for businesses. These tariffs price ISDN much closer to POTS. As the Commission is aware, BA does not ask for per-minute charges on voice ISDN Centrex services, nor does it collect per call charges when calls are made within a "virtual" office. BA has priced ISDN Voice Centrex, which is the same service as Centrex for data, closer to POTS, because it faces competition in the PBX market. In some states, BA has reportedly claimed that usage costs for voice Centex service are less than $1 per month. The Commission can and should tell consumers what BA has claimed in the past about its ISDN Centrex costs.

We do have some publicly available information on Bell Atlantic's non-traffic sensitive costs. In an October 18, 1995 filing with the FCC, Bell Atlantic reported that its monthly loop costs were $14.56 for POTS and $16.77 for ISDN, a difference of $2.21 per month, or 15 percent. (Exhibit 3, Bell Atlantic data request, CC Docket No. 95-72, October 18, 1995). In addition, Bell Atlantic reported NTS non-loop costs for ISDN of $4.18 per month. On page 4 of that pleading, Bell Atlantic said:

Shown in the answer to question 1 and as discussed in Bell Atlantic's Comments and Reply Comments in this proceeding (copies attached), the difference in the costs between dial tone lines and Basic Rate ISDN loops is minimal. The small difference that does exist is due to an additional ISDN plug-in card which is required to transmit ISDN signals.

Bell Atlantic's estimate of $6.31 for the incremental NTS cost of ISDN is actually higher than for the other major LECs. In the same proceeding, US West estimated that its NTS cost of ISDN was only $1.18 higher than POTS. (Exhibit 4, June 29, 1995 filing by US West in CC Docket No. 95-72). Data on the NTS cost of ISDN are also reported somewhat differently in the FCC's December 24, 1996 Notice of Proposed Rulemaking, Third Report and Order, and Notice of Inquiry, on page 36, Table 2 (Exhibit 5, FCC 96- 488). The following data were reported:

Table 1

Ratio of Costs of Standard Analog Service to BRI ISDN Service
LECOutside Plant (loop only) costs All NTS costs
Bell Atlantic1:1.01 1:1.36
Pacific Bell1:1.051:1.13
US West1:0.801:1.07
Average Ratio of Costs1:0.96 1:1.124

Source: FCC 96-488, December 24, 1996 NPRM, Table 2, page 36

It is interesting to note that Ameritech is the LEC reporting the greatest NTS costs for ISDN BRI, while its residential prices are far lower (except for Indiana) than the residential rates charged by Bell Atlantic and the other LECs for which data are available.

Q12. How should residential ISDN BRI Service be priced?

A12. The Commission should price residential BRI ISDN as close to the incremental cost of the service as possible. It is in the public interest to expand deployment of ISDN and other higher bandwidth technologies. Pricing is the major barrier to deployment of ISDN services today.

Bell Atlantic's call-pack options are far too complex, and require consumers to anticipate and pay for usage that they never need. The Commission should reject the call pack approach as needlessly complex, and replace it with three options. First, a fully metered option, with usage fees in the neighborhood of 10 cents per hour, or another number which is no more than 10 to 20 percent higher than their direct costs. Second, a "virtual flat rate" (VFR) option with a 200 hour usage allowance. The VFR option would not constrain normal dial-in users, while preserving an incentive to hang up the line when it is not in use. We believe the VFR option should be priced under $20 per month. Third, BA should offer a flat rate option with no limits on usage. This flat rate should be priced higher than the VFR, but not excessively, based upon Bell Atlantic's true costs.

Bell Atlantic has expressed concern over nailed-up connections, and it is reasonable for the Commission to make a distinction between dial-in usage and nailed-up lines. The $249 unlimited usage fee that BA wants to charge is too high, however, even for a nailed-up connection. Indeed, it is more than half the rate that some BA customers pay for a dedicated T-1 connection, with far less bandwidth. Why should residential consumers pay more than business ISDN Centrex users, or business T-1 users? This is quite punitive.

For now, however, the most important rate for the residential market is clearly for dial-up users. Nailed-up lines are and will be a tiny and rather unimportant part of the residential market, for some time. However, the long term goal should be for every household to have affordable 24x7 connections to the Internet, delivered through a packet switched data network.

Q13. What about network congestion?

A13. There is no evidence that ISDN is causing any network congestion. Nor is there much evidence that personal computers are causing network congestion. While it may be the case that lines for Internet Service Providers (ISPs) are used heavily at certain times of the day, it does not follow that residential users are causing congestion problems.

Telephone networks are fixed cost plants. The cost of the network is determined by the build-out at any given time, regardless of usage. The only "usage" cost is based upon the "build-out" for peak usage. The residential network is built according to certain assumptions regarding usage. The important issue is the proper "capacity" for the system. People who use the Internet as dial-in users have complicated lives. They eat, sleep, go to school, go to the movies, play with their children, and do other things which don't involve the Internet. For this reason the important policy analysis must focus on what percentage of Internet users are connected at any one time and how that compares with the capacity of the network.

What do we know about Internet usage? We know that ISPs do not have one line per customer. They have far less than one line. According to responses from ISPs, most firms serving the home market have between 10 and 20 customers per incoming line. That means simply that only 5 to 10 percent of their customers can be using the service at any given time.

When American Online (AOL) announced its flat rate service, it had about 200,000 modems for 8 million customers, so it could only accommodate about 2.5 percent of its customers at any given time. AOL needed at least 400,000 lines to accommodate the increased demand due to its flat rate service offering. That is why the AOL lines were busy, and it was hard for consumers to connect. But we have to look at this situation to understand what it means for the residential telephone network. If only 2.5 percent of AOL's customers could be online at one time, then 97.5 percent are off-line. AOL's Internet servers were congested, but the AOL users were not causing congestion of the telephone network.

The Bell Atlantic residential telephone network is built for 14 percent of the customers to use the network at any one time. The ISP's which sell to residential users can't accommodate anywhere near 14 percent of their customers.

The length of calls to ISPs are longer, and even the total minutes of calling may be longer than for voice. But Internet users likely use the network differently than voice callers in other ways too. For example, ISPs report heavy loads after 9 p.m., when voice calling dramatically declines. Ultimately, you have the independent and compelling data from the ISPs, which provide powerful evidence that residential ISP consumers as a group never approach the 14 percent connection rate that is the capacity of the residential voice network.

Q14. What about congestion at the ISPs?

A14. ISPs, like other businesses, may use lines intensely. That isn't necessarily bad for BA, because BA is selling telecommunications services to the people who make the incoming calls. For example, BA is promoting the sale of second telephone lines for residential consumers. These POTS lines do not generate much in long distance revenue. They are used mostly for teenagers, modems and faxes. When BA sells an apparently profitable POTS line so a consumer can call an ISP, BA is making money on the sale of the residential POTS line. If the residential consumer could not call an ISP, many consumers wouldn't need the second POTS line. Thus, the ISP is generating a demand for something that BA sells--residential lines. Also, the ISPs are ordering T-1 connections and other services that BA sells. Indeed, it is instructive that as Internet businesses have taken off, the LEC revenues, profits and stock prices have also done very well.

It is important to understand that the load on each of an ISP's incoming lines wouldn't change even if the residential consumers called more or less per month. If the residential consumers called only 5 minutes per month, or 500 hours per month, the ISPs would still purchase just enough lines so that their customers would not be put off by busy signals. Indeed, when AOL began, it could only connect 1 percent of its customers at a time. After AOL adds more lines and can connect 5 percent or more of its customers, its incoming lines will likely have a similar per line load as when they only could connect 1 percent.

Q15. Should ISP's pay higher fees?

A15. BA makes much of claimed network congestion allegedly caused by ISPs and Internet users. As a preliminary matter, these issues are more appropriately considered in the access reform proceeding now being conducted at the FCC or in a separate state proceeding on business tariffs. This hearing is focusing on the rate for residential ISDN consumers. Nonetheless, BA has a practical monopoly in the residential market, and they are also an active ISP. We do not think it is wise to allow BA to discriminate against their rivals in the ISP business. Lots of businesses make intensive use of their lines. Our organization has a PBX, and it is often congested during the day. It is difficult to understand what public interest rationale would be met by requiring ISP's to pay higher rates than other businesses. If Bell Atlantic wants to raise the rates for business users who have above average usage, they should lower rates for other below average businesses.

The LECs have a technology which permits them to take data calls off the circuit switched network, and transport calls to the ISPs using a packet switched data network. One technology involves the new Nortel device that is now being widely advertised. The LECs want to raise the rates that an ISP pays for its incoming lines by $10 to $15 per month, to make the Nortel service more attractive to the ISPs. In the FCC bandwidth forum, we proposed an alternative. We think the regulators should remove all usage charges for ISDN calls that are delivered over the Nortel technology. Then the ISPs might have an incentive to pay more for the Nortel technology, because their customers could get cheaper ISDN. This would be a win-win solution. The LECs would sell more residential ISDN, and the data would be taken off the circuit switched network. The LECs haven't shown much interest in this, even though the outrageous ISDN tariffs are based upon the assertion that ISDN use of the circuit switched network is costly.

Q16. What price would allow residential ISDN greater penetration in the market?

A16. The "magic number" for residential ISDN is somewhere under $30 per month, fully loaded, including the SLC charges. Bill Gates thinks residential ISDN should be priced below $20 per month, and so do we (The Road Ahead, p. 101). But at a minimum, the Board needs to keep the cost to the consumer under $30 per month. We think the Board should try to price BRI ISDN line closer to $20 per month, so we can attempt to really expand deployment.

Q17. What about International Comparisons?

A17. I was in Japan in August 1996, and meet with NTT officials, who were involved in efforts to introduce flat rate options for Internet connections in Japan. Many Japanese technology experts believe Japan is way behind the United States in terms of residential Internet connections because they charge about 3 cents per minute for a local telephone call. They were trying to find a way to become more like the United States. However, in Japan there was a great interest in rapid deployment of ISDN lines. There are even ISDN pay telephones everywhere with jacks for connections for personal computers. The monthly price of a POTS line was 2,830 Yen, and the ISDN line was priced at 3,500 Yen per month, a difference of 670 Yen, or about $5.41 per month at today's exchange rate. Put another way, the ISDN line costs less than 24 percent more than a POTS line. By contrast, look at Bell Atlantic's huge mark-ups for residential ISDN.

NTT distributes literature which provides the names and contact information of ISPs providing ISDN connections to the Internet. NTT officials told me in August they wanted to double the number of ISDN lines by the end of 1996. I've never heard any U.S. government regulator set any goals for deployment of digital services. There was one other aspect of the Japanese approach I found interesting. NTT was trying to develop a new high speed data network, and promote flat rate residential options, in limiting hours. But to get the flat rate option, you would have to register two telephone numbers with NTT, and these would be unmetered. It would be interesting to think about this in the context of the Nortel technology, to see if we can begin to think about the elimination of usage charges on digital residential connections to the Internet.

In Europe, ISDN is also being encouraged, and in some countries, POTS and ISDN prices are identical. But many European countries suffer from high usage fees for local calls. In my view, the flat rate local call system in the United States explains why the United States seems to so completely dominate the development of the Internet today.

Q18. Does this conclude your testimony?

A18. Yes.