April 1, 1997
Competition, Innovation and Deregulation
Speech by Joseph Farrell, FCC Chief Economist,
at Merrill Lynch "Telecommunications CEO Conference",
New York, March 19, 1997.
Thank you for inviting me here. My predecessor in this job, Michael Katz, told some of you that
the country thrives because there are two of each telecommunications company: the one we hear
from in Washington, almost bankrupt, desperate for relief or a subsidy; and the one you hear
from in New York, rich beyond belief, its prosperity exceeded only by its future prospects. It is
true that I've heard more optimism about the industry these two days than I have in months in
Washington. In particular, I've heard a lot more optimism about competition, and so it's fitting
that I'm going to talk about what can happen as competition develops. In fact, my subtitle might
be "Letting the BOCs out of the Bag."
I know you were expecting the chairman. Reed is not only the chairman, but also a great public
speaker, so you must be disappointed. Last time I substituted for him, in Chicago, I read his
speech. (Actually, several people afterwards complimented me on how well I read aloud, which
I found vaguely insulting: do I look as if I might have trouble reading?) I generally begin my
speeches by warning that my views are not necessarily shared by the Commission or any of the
Commissioners. Of course, in Chicago I said the opposite. But today these are my thoughts, not
I also know you are hoping to get a sneak preview of the access reform and universal service
numbers. These are big numbers and deserve a lot of discussion, which I assure you they get.
(Actually, what's striking is that all the numbers are either billions of dollars or fractions of a
penny.) But I'm not going to give you numbers today. Instead, I want to talk about something
even more important for investment opportunities in the long run. Harder to quantify, harder to
turn into a spreadsheet for your clients, but the whole ballgame in the long run. That is
deregulation, and particularly deregulatory incentives for innovation.
This will require a real shift of "corporate culture" among regulators. Unfortunately, I don't
believe this industry is one where immediate complete deregulation would be a good idea.
Market power is too great, and, even more importantly, the development of competition requires
enforced cooperation from the incumbents. So we can't just RIF the FCC (and the PUCs): we
have to have a gradual and judicious deregulation. This means that regulators have to get good at
knowing when not to regulate. That's not easy substantively, and it's not necessarily easy
The tradition seems to be that everything in telecom is regulated unless it's provable that it
shouldn't , or needn't be. Indeed, I understand that a federal court recently stayed an FCC order -- no, not that one, but the one that de-tariffed the long-distance industry. Apparently we hadn't
proved our case that the government has no need to be involved!
As I say, when to deregulate (and how much) is not an easy question, but here's a suggestion -- of
course, one that begs lots of questions. Regulation should be applied when otherwise there is a
serious threat of sustained, substantial harm, and when regulation clearly makes it better. I
stress the three S's because any good economist or antitrust lawyer can always think of something
wrong with unregulated behavior in any market. I'd also say that one should perhaps look
particularly hard at harms that affect the growth of competition (entry barriers) or that exploit
low-income or defenseless consumers of essential services. A certain amount of gouging of rich
consumers, or consumers of inessential services, may not be so terrible, especially if competition
is capable of riding to the rescue.
Deregulation of innovation may be particularly important. Innovation in telecoms may be a new
service using existing facilities, or it may involve new facilities. I'd like to mention briefly the
new-service issue, then spend most of my time on new facilities, partly because facilities-oriented regulation is newer and attracting more attention these days.
Traditionally, ILECs face suspicious regulators when they want to introduce a new service, or
change a price structure, or reduce a price. Sometimes the suspiciousness is substantive; other
times, there is no clear reason to be suspicious, but because there might have been, changes
require approval, and approval takes time. Of course, it takes more time when regulators are
over-busy and under-staffed. More worrying, it takes more time when many ILECs file many
applications for many innovations. Our goal should be to reduce these delays and burdens. And
I'm happy to say that in a Third Report and Order, gift-wrapped along with the Christmas Eve
Access Reform Notice, the FCC considerably streamlined the process of approving new service
tariffs. I hope this helps a lot. I hope we can go further quite soon.
Now, what about innovation and unbundling? Innovation is driven partly by incentives and
absence of needless hurdles for incumbents, and partly by incentives and opportunities for non-incumbents. Both of these are important. Sometimes, in some industries, innovation is driven
by large incumbents, who alone have the expertise to innovate or the complementary assets that
let them exploit bright ideas. Then it's important not to mar their incentives to innovate. On the
other hand, sometimes innovation is driven by a multitude of unheard-of entities, everyone with
a bright idea. When it's feasible, it seems to be important to open things up so that everyone with
a bright idea can get it tried out in the marketplace.
Look at the computer industry. Lots of progress was made, partly because there were
unregulated market incentives, even while the industry was dominated by large vertically
integrated hardware-and-software firms. But, as Andy Grove has often stressed, things really
took off after this "vertical model" gave way to the "horizontal model," in which many niche
players have access to unbundled complementary inputs: software providers like Intuit or for that
matter Microsoft don't have to build computers or even strike special deals with hardware
providers like Compaq or Intel. This ability to be a specialized player is one of the things that
should come out of the unbundling and development of interconnection in telecom.
Specifically, someone who wants to innovate at the switch level, or by loading xDSL onto a
regular copper loop, or by offering a different, equally remunerative but possibly more appealing
pricing plan to end-users, probably shouldn't have to build a lot of loops in order to try her idea
in the market, any more than Scott Cook of Intuit, or a start-up shareware entrepreneur, has to
build PCs. Nor should such an outside innovator have no option but to sell her ideas to a
dominant incumbent: every venture capitalist knows that is no way to realize an invention's true
worth. (Alexander Graham Bell couldn't even get $100,000 from Western Union for his
Thus it's important for innovation that we make unbundling work. But some have suggested that
regulatory unbundling requirements, applied ex post to network elements that are created by the
incumbent's innovative efforts, reduce the incumbent's incentive to innovate that way.
Congress didn't ignore this issue. Look at section 251(d)(2) of the Act. I haven't heard very
much public discussion of that, and I think that's a pity. So let me put forward a few things to
think about. I'll stress that these are not policies, not even proposals, but, as Woody Allen once
said, "concepts, and I'm looking for funding to turn them into ideas." Also, you should think of
them as my concepts, not the Commission's or any Commissioner's. If you want them, or their
cousins or their opposites, turned into the Commission's ideas, start talking publicly about this
* Allow a lead-time for the developer of a new network element before it must be made available
to competitors. This after all is as close as you can come to a direct translation into our context
of standard intellectual property policy -- patents and copyright policy. Efficiencies of certain
innovations must be shared with competitors, but not right away.
Of course, when an innovation is patented and thus temporarily kept from competitors, the pre-existing state of the art is still available to them. Similarly, when an ILEC invests in a "parallel,
better" infrastructure, such as fiber to the home, perhaps it should be able to keep that from
competitors, for a while or even for as long as it wants, provided that the old infrastructure is still
available on an unbundled basis, is properly maintained, and so on. But how does that proviso
get enforced and ensured? This leads to another idea:
* Maybe, rather than requiring the ILEC to continue to lease the copper out to CLECs once it's
using the fiber, we should think about allowing the ILEC to solve the problem of its bottleneck
control by divesting ownership of the copper to a competitor. Then there would be two
competing facilities-based providers of, in this example, loop infrastructure to a particular
neighborhood. If a third firm wanted to come in and get unbundled loops, it would have a choice
of where to get them. It wouldn't be at the sole mercy of an entry-resisting monopolist. So
might the whole regulatory infrastructure of requiring ILEC unbundling -- for this overbuilt
element, in this local market -- be loosened, relaxed, even removed, if we thought that two direct
competitors is better than regulation?
* Similarly, once a CLEC overbuilds facilities to serve a neighborhood, there is, again,
competition in the provision of network elements, in the sense that a carrier wanting to take
unbundled elements has more than one seller to go to. Should this be enough? Should the ILEC
then be released from some of its duties to unbundle, with the expectation or hope that the
competitive provision will lead the ILEC and the CLEC between them to be willing to offer
unbundled facilities even to competitors? This has happened in the long-distance market, where
competition leads the facilities-based carriers to unbundle capacity at highly competitive prices
even to their arch-enemies such as GTE, SNET, and potentially the BOCs.
But the long-distance market has more than a duopoly. Is duopoly, a market structure
with two directly competing firms, adequate competition for this purpose? I think we can predict
that duopoly won't lead to the kind of long-run competitive pricing that the FCC and states have
aimed for in their interconnection policies. In fact, even the fairly rivalrous long-distance market
does not achieve that. Duopoly will sometimes lead to prices lower than long-run costs, and
sometimes -- probably more often, if entrants are rational -- to prices that are higher. Is that good
Or, rather, is that better than regulation, which itself is highly imperfect? Perhaps not, if
the duopoly is highly collusive. But probably yes, if it's moderately rivalrous. Should we figure
out ways to experiment with this? Or should we just take the plunge and say that a duopoly, with
no obvious reasons why more competitors can't enter, is not a market that we must regulate?
Once one CLEC has overbuilt in a market, showing it can be done, should we presume that any
serious sustained abuse would induce further entry?
Important follow-on incentive questions are raised by this concept-or-proto-idea. What
incentives would such a rule create for ILECs? Would they find the prospect of not having to
unbundle on regulated terms an important carrot that might induce them to cooperate with the
CLEC who wants to build? (Would this cooperation take anticompetitive forms? Would it
consist of more cooperation in providing "stepping stones" for such a CLEC?) And what
incentives would it create for CLECs? Would a CLEC, observing that other CLECs probably
could not get network elements at exactly regulatory prices out of the duopoly, find that an
important inducement to build? And would other CLECs then be persuaded to build? If so, is
that good, or would it lead to an absurd cascade in which some local markets fortuitously have
half a dozen overbuilds and others have none?
Let me close by stressing once more that these are concepts in search of intellectual and informational "funding" to be turned into ideas. But I think if you turn your attention to the prospects for such incentives for innovation, and to the prospects for deregulation in general, you may get out ahead of the number-crunching Wall Street crowd.