Copyright 2000 Federal News Service, Inc.
Federal News Service
March 28, 2000, Tuesday
SECTION: PREPARED TESTIMONY
LENGTH: 3685 words
HEADLINE:
PREPARED TESTIMONY OF TIMOTHY J. REGAN VICE PRESIDENT AND DIRECTOR OF GOVERNMENT
AFFAIRS CORNING INCORPORATED
BEFORE THE SENATE
COMMERCE COMMITTEE SUBCOMMITTEE ON COMMUNICATIONS
BODY:
(NOTE: TABLES NOT TRANSMITTABLE)
Introduction
Mr.
Chairman, my name is Tim Regan. I am a Vice President of Corning Incorporated. I
understand that today's hearing is about the deployment of broadband to rural
America. Obviously, this is of great interest to me as a representative of
Corning. We are the original inventors of optical fiber, and of course, are
anxious to see the technology deployed to all Americans, especially those in
rural America.
But, I think it is important to address the question of
broadband deployment to rural America in the context of the
deployment to the nation as a whole. My argument is very simple. Broadband is
not being deployed to residential customers in America, regardless of whether
they are located in urban, suburban, or rural America. Business customers are
getting it, but residences are not.
I know that you might find this
statement somewhat astounding because you hear a lot about the so-called
broadband deployment. Cable modem service, ADSL service (i.e.,
asynchronous subscriber line), and various wireless data services all claim by
some, most notably the FCC, to be broadband. Without getting into semantics, I
will argue in my testimony that these capabilities are more properly described
as higher-speed data service, not broadband service. I will also describe in my
testimony recent economic research that Coming has commissioned to determine why
broadband capability is not being deployed to residential customers. In short,
the study identifies both financial and regulatory barriers to deployment.
Regulation changes alone are insufficient to get the job done.
What's Broadband
The first issue, of course, is the question of
what is broadband. The answer is not obvious.Oddly enough, the term "broadband"
really comes from an older age - the analog age. In the analog age, the
information-carrying capacity of a network was defined by the width of the band
of spectrum used to carry a signal. The wider the band, the greater the
information-carrying capacity. Thus, the term "broadband" was used to
characterize a system capable of carrying a considerable volume of information.
In the analog world, a standard television video signal that requires 6
megahertz per channel was considered to be broadband. Voice at 4 kilohertz was
thought to be narrowband.
In the digital world, the notion of broadband
really doesn't apply. The information carrying capacity of a digital network is
described as a bit transfer rate. As you know, digital signals are represented
by a series of on and off signals that are characterized by pulses of electrons
or photons. Transmissions in the digital world appear more like Morse code.
If we use standard television video as a service to characterize
broadband, as we have done in the analog world, a bit transfer rate of 4 million
to 90 million bits per second would define broadband. An uncompressed standard
television video signal requires 90 million bits of information per second to
transmit. It can, however, be compressed to 4 million to 6 million bits per
second using what is called MPEG-2.
Data has become a very important
form of information in the digital world. Remember that computers were
originally called data processing machines. In the computer data world, the
connections between computers are quite robust. A standard has evolved known as
Ethernet, developed by IBM over two decades ago. It provides for the
transmission of 10 million bits per second between computers on a local area
network. Today, the Ethernet standard has been upgraded to a 100 million bits
per second.
Frankly, I think the term broadband is so imprecise, it is
probably useless at this point.I think the better way of engaging the public
debate is to identify bit transfer rates Americans will need to gain access to
audio, video, and data applications. Table 1 below, which was taken from an
article written by a Microsoft orificial, describes the transmission speeds
necessary to gain access to a variety of applications.
If you think that
Americans will need access to information in all its forms - audio, video, and
data - it is easy from Table 1 to see that a capability in excess of 20 million
bits per second downstream and 10 million bits per second upstream, even using
the most advanced compression technology, is necessary. Let me explain with some
examples of the bit transfer speeds necessary to do audio, video, and data:
Plain old telephone service requires 64 thousand bits per second both upstream
and downstream. Standard television using MPEG-2 compression technology uses 4
million to 6 million bits per second per channel downstream. Since there are on
average 2 1/2 television sets in every household in America, three channels at
4-6 million bits per second each is needed. HDTV using the most advanced
compression technology requires 20 million bits per second downstream. And, 10
million bits per second both upstream and downstream - the so-called 10 Base-T
Ethernet standard - is required to give people the same data speeds at home that
they get at work in order to facilitate telecommuting.
I realize that my
bit transfer speed prescription sounds like a lot. But, I believe it is what
will be needed.
Let me clarify one point though. My comments about
broadband should not be construed as criticism of ADSL or cable modem service.
These are wonderful technologies. They enable the delivery of data at
substantially higher speeds over the existing infrastructure that has been
deployed by ILECs and cable operators. These services provide a useful
transition to full broadband.
The FCC has stated in its Section 706
proceeding that broadband is 200 thousand bits per second - or 1% of my
prescription. I do not see how the FCC can defend such a low standard in light
of the speeds described in Table 1 above as necessary to transmit the
applications we know of today, never mind the limitless array of new ones that
will be created once the infrastructure is deployed.
The FCC and others
have defined broadband at such a low level because they fundamentally
misunderstand the nature of the future network. It has been described by the FCC
as a superhighway. And, consistent with this analogy, the connections to the
home are simply narrow on and off ramps.
This is the wrong analogy. The
network of tomorrow, which will be dominated by data not voice, is not a
highway. It is a series of bridges.
The bridges connect islands of
intelligence computers. After all, this is what the Internet is. It is a network
of computers, and each computer has the capacity to store and process hundreds
of millions of bits of information.
Today, these islands of intelligence
are for the most part connected by very narrow bridges, a copper pair that can
transmit only 56 thousand bits. Even with these very narrow bridges, we have
been able to realize tremendous economic benefit from connecting these islands
of intelligence.
Fed Chairman Alan Greenspan best characterized the
impact of this connectedness in October last year before the Business Council
when he said:
"Your focus on technology - particularly the Internet and
its implications - is most timely...The veritable avalanche of real-time data
has facilitated a marked reduction in the hours of work required per unit of
output and a broad expansion of newer products whose output has absorbed the
work force no longer needed to sustain the previous level and composition of
production. The result during the last five years has been a major acceleration
in productivity and, as a consequence, a marked increase in the standards of
living for the average American household (emphasis added)."/1
Tremendous economic prosperity has been realized over bridges that
connect the computers at 56 thousand bits per second. Can you imagine what will
happen when we can connect these islands of intelligence by bridges that can
carry over 10 million or 20 million bits per second?
The question before
us is how to build these bridges as soon as possible. The problem for rural
America is particularly acute because the cost of building these bridges is 2-3
times higher than it is for the rest of the country.
How Do We Build the
Bridges?*
Obviously, to deploy this new technology will require
considerable investment on the part of all telecommunications carders. The
problem is, the dynamics to finance this investment have not been unleashed.
In fact, we have witnessed some unusual behavior. Incumbent local
exchange carriers (ILECs) continue to deploy copper wire rather than new
technology like fiber optics to provide service to new residential customers
(i.e., "new builds") and to rehabilitate deteriorated plant that is serving
existing customers (i.e., "rehabs"). They are spending approximately
$9 billion deploying copper to serve new builds and rehabs in
the residential market.
This reality was evidenced in a recent article
in The Wall Street Journal which stated:
"Global sales of communications
wire, from fiber-optic and coaxial cable to old-fashioned copper, rose 6% to
$14 billion last year...Here's the most surprising part: The
bulk of the industry's sales continues to come from the same type of wire
Alexander Graham Bell developed in 1879 to transmit voice signals - copper
(emphasis added)."/2
The fiber optics industry is somewhat puzzled by
this investment behavior because fiber optic systems solutions today are at
relative cost parity with copper. The cost parity between fiber optic and copper
solutions for residential customers is well established. Last August, Matthew
Flanagan, President, Telecommunications Industry Association, submitted comments
to the FCC attesting to this fact. As evidence, he submitted sworn affidavits
from four different telecommunications engineering experts who all supported the
cost parity claim./3
Because we are somewhat puzzled by this investment
behavior, we commissioned a study by three Ph.D. economists, Drs. Kevin Hassett
and J. Gregory Sidak, who are associated with the American Enterprise Institute
for Public Policy Research, and Dr. Hal Singer who is associated with Criterion
Economics. The study concluded that the ILECs and the CLECs are acting very
rationally in delaying their decision to invest in new technology to serve
residential customers.
They identified both financial and regulatory
explanations for the delayed investment behaviors.
From a financial
perspective, this delayed investment behavior is explained by a rather new model
for explaining investment behavior known as the Dixit-Pindyck model. This model
shows that when faced with certain conditions, a prudent investor will maximize
his return by delaying investment in next generation technology. These
conditions include a sunk cost investment, a high degree of market or technology
uncertainty, and the absence of robust competition. Under these three
conditions, which are all prevalent in the residential telephone market, a
carrier is better off delaying a decision to invest in new technology./4 Since
ILECs are required to provide telephone service, they invest in copper solutions
which are suited for just plain old telephone service.
The study goes on
to conclude that the incentive to delay for ILECs is intensified by the
so-called unbundling rules which require incumbents to allow their competitors
to use parts of the incumbents' network at a regulated rate. This rate does not
provide a sufficient return on investment to justify investment is new
technology.
The parts of an ILEC's network that must be unbundled and
resold to competitors are known as unbundled network elements, or "UNEs." The
FCC has defined the price for the sale of these UNEs as TELRIC, or total element
long run incremental cost. TELRIC attempts to value the various network elements
based upon their forward-looking costs. The FCC believes that TELRIC replicates
how competitive markets actually operate by approximating what it would actually
cost an efficient, competitive firm to produce LINEs.
The study
concludes that TELRIC pricing creates a disincentive to invest in new
technology. It states:
"Most observers believe that mandatory unbundling
(at TELRIC) limits the upside potential of any new investment project and that
the expected return to investment in some projects may fall below the firm's
cost of capital .... This disincentive to invest has been emphasized in the
public debate over telecommunications policy by both incumbent local exchange
carders (ILECs) with respect to the local telephony networks, and by AT&T
with respect to proposals that unaffiliated Internet service providers be given
the legal right of mandatory access to AT&T's cable-television networks."/5
In other words, the rate of return provided for TELRIC pricing is
inadequate to give carders an incentive to invest in new technology.
Other experts, including Kathleen Wallman, former Chief of the FCC's
Common Carder Bureau and Deputy White House Counsel as well as Supreme Court
Justice Breyer, have observed this disincentive. Ms. Wallman stated in a speech
to state regulators:
"Do we really mean to say that any carder that is
thinking of building a new broadband network should count on being able to
recover, from day one of the operation, only the forward looking cost of their
brand new network? I don't think so. No rational, efficient firm would take that
deal. And that would be our collective loss, not just theirs./6
Similarly, Justice Breyer reinforced this observation in last year when
he noted that "... a sharing requirement may diminish the original owner's
incentive to keep up or to improve the property by depriving the owner of the
fruits of value-creation investment, research, or labor."/7
The point
is, the new economics as characterized by the Dixit-Pindyck model combined with
the unbundling rules at TELRIC create a powerful disincentive for ILECs to
invest in new technology. This disincentive is reflected in the stock price of
incumbents, including AT&T, when they make decisions to invest in
infrastructure. Their stock price falls. In January, it was eported in The Wall
Street Journal that the share prices of SBC, Bell Atlantic, and GTE fell when
Paine Webber downgraded the firms because they "...may have to make additional
investments to deploy high-speed Internet-access services..."/8 With this
explanation, it is clear that both financial and regulatory changes are
necessary to give carders an incentive to invest in new technology, especially
broadband technology. The important point to remember is that both financial and
regulatory changes must be made.
Both financial and regulatory changes
have been proposed by Members of this Subcommittee.
Senator Rockefeller
recently proposed a bill to provide financial incentives for rural deployment of
higher-speed data and broadband service. Senator Brownback has proposed a bill
to eliminate the regulatory barriers to deployment of pocket-switched,
higher-speed capability without repealing the inter-LATA restrictions or the
unbundling requirement for the existing copper loop. Both proposals are
necessary to get the ball rolling. I applaud their efforts.
Conclusion
Mr. Chairman, in conclusion, I think my testimony can be summarized by
two points: First, broadband is not happening. Second, to accelerate broadband
technology deployment both financial and regulatory changes are necessary.
Thank you for your time and attention. I stand ready to address any
questions you may have.
FOOTNOTES:
1 Remarks by Alan Greenspan,
Information, Productivity, and Capital Investment, Before the Business Council,
Boca Raton, Florida, October 28, 1999.
2 Mark Tatge, "Wire Makers Thrive
Despite Advent of Wireless Phone", The Wall Street Journal, February 16, 2000,
p. B4.
3 Matthew J. Flanagan, re: Implementation of the Local
Competition Provisions in the Telecommunications Act of 1996, CC Docket No.
96-98, Telecommunications Industry Association, letter to Federal Communications
Commission, August 2, 1999, which states at p. 6-7 that "In his Declaration, Mr.
Cannata from Marconi Communications, demonstrates that POTS can be provided over
a fiber-to-the-curb ("FTTC") system at 98 percent to 103 percent of the cost of
providing POTS over a copper system using a digital loop carrier ("DLC/copper").
He notes further that the FTTC system can be upgraded to provide high- speed
data (i.e., 10/100 Base T) by incurring a 16 percent incremental cost compared
to a 40 percent to 50 percent incremental cost to upgrade DLC/copper to provide
Digital Subscriber Line (xDSL) service. Finally, he demonstrates how a further
upgrade to provide VI-ISquality broadcast video can be deployed for an
incremental cost of 44 percent over FTTC for POTS, which again compares
favorably to the 40 percent to 50 percent incremental cost associated with the
ADSL solution.
Mr. Jacobs from Coming Incorporated shows in his
Declaration similar results with respect to broadband solutions. His analysis
shows that an Ethernet fiber-to-the-home system (EFTTH") using multimode fiber
can be deployed at 7 percent less than ADSL over copper, and EFTTH is
substantially more capable. The EFTTH system can deliver POTS, 10/100 Base T
data, and VHS-quality broadcast video, which cannot be done on an ADSL system.
Mr. Tuhy from Next Level Communications states in his Declaration that
"fiber-based narrowband solutions for local access serving residential end-users
can be deployed at cost parity with copper-based solutions as measured on an
installed first cost basis for newly constructed or totally rehabilitated
outside plant." He makes a similar statement with respect to broadband. He notes
that Next Level Communication's FTTC system "can be deployed to provide
integrated voice, data, and video for the same cost as a copper-based solution
with an ADSL overlay for high-speed data." This assumes new builds or total
rehabs as well as first installed cost comparison. Finally, Mr. Sheffer from
Coming Incorporated addresses the rural deployment issue in his Declaration. He
cites a proprietary Bellcore (now Telcordia Technologies) study prepared for
Coming showing that the cost of narrowband fiber-to-the-home (" H") at
$2,370 per home passed beats narrowband DLC/copper at
$2,827 per home passed. In other words, narrowband FTTH is 16.2
percent less costly than DLC/copper in a rural setting. More surprisingly,
broadband FTTH also beats narrowband DLC/copper by 7.5 percent (i.e.,
$2,616 per home passed for broadband versus
$2,827 per home passed for narrowband). Again, this analysis
was based on new builds and total rehabs and the cost comparisons were done on
an installed first cost basis.
4 Kevin A. Hassett, J. Gregory Sidak, and
Hal J. Singer, An Investment Tax Credit to Accelerate Deployment of New
Generation Capability, February 28, 2000, p. 7, which states: "A simple example
can make the point more intuitive. The traditional view is that one should
invest in any project that has a positive net present value of cash flows.
Recent advances in economic theory have shown, however, that this rule is not
always correct. On the contrary, it is often better to wait if at all possible
until some uncertainty is resolved and cost reduction can be achieved. Consider,
for example, a finn that traditionally offers telecommunications services
through copper wire. The firm must decide whether to install a new advanced
broadband line that costs, say, $100 today but has an uncertain
return tomorrow. Suppose that, if the demand for high-bandwidth services is
high, the finn stands to make $400 profit. If, on the other
hand, there is a bad outcome and the demand for the new services is low, then
the new "pipe" will be underutilized, and the firm will gain nothing from owning
it. If the probability of either outcome is 0.5, then the expected net present
value of laying the new broadband line is, ignoring discounting, calculated as
follows: (0.5 x $400) + (0.5 x $0) -
$100 = $100. We can summarize this simple
decision problem in the following table.
Because the project has a
positive expected cash flow, one might think it optimal to install the cable
today. But it is not. If the firm delays making the investment, it can reduce
the risk by observing the experience of others and capturing the gains
associated with deploying reducing-cost technology later. The value of waiting
is that the firm can decide not to make the investment if the bad state occurs.
We can summarize this subtler decision problem in the following table:
By waiting, the firm would increase its expected return by
$50. If the firm invests today, it gives up an option to invest
tomorrow that is worth $50. The firm is better off waiting
because it can avoid the loss of $100 by not purchasing the new
cable in the bad state. Note that the two examples would have the same expected
return if the firm were allowed to resell the advanced broadband line at the
original purchase price if there is bad news. But that salvage scenario is
patently unrealistic for two reasons. First, many pieces of equipment are
customized so that, once installed, they would have little or no value to anyone
else. Second, if the demand for high-bandwidth services is indeed low, then the
advanced broadband line would have little value to anyone else. For these
reasons, the investment in the equipment is "irreversible" or sunk in the sense
that it has virtually no value in an alternative use.
5 Id., p. 3-4
6 Remarks of Kathleen Wallman at the annual convention of the National
Association of Regulatory Utility Commissioners, Boston, Mass., Nov. 11, 1997.
* AT&T Corp. v. Iowa Util. Bd., 119 S. Ct. 721, 753 (1999) (Breyer,
J. concurring in pan and dissenting in pan) (citing 1.H. Demstez, Ownership,
Control, and the Firm: The Organization of Economic Activity, 207 (1988)). 8
Stephanie H. Mehta, "Local-Phone shares Fall Amid concern Over Firms' Need to
Invest, Rising Rate", The Wall Street Journal, Jan. 13, 2000, at B4.
END
LOAD-DATE: March 29, 2000