The
Next Frontier for Openness: Wireless Communications
Eli
M. Noam
Professor
of Finance and Economics, and Director, Columbia Institute for
Tele-Information, Columbia University
Paper
for the 2001 Telecommunications Policy Research Conference
Alexandria,
Virginia
October 26, 2001
Abstract
For wireless communications, the FCC has fostered competition rather than openness. This has permitted the emergence of vertically integrated end-to-end providers, creating problems of reduced hardware innovation, software applications, user choice, and content access. To deal with these emerging issues and create multi-level forms of competition, one remedial policy to will suffice: a Carterfone-style separation of the link between equipment and network provision for wireless.
The Wireless Policy Mess
Today
with economic and political uncertainty abounding we are in one of those great
divides, where we either can go back to familiar territory or go forward to the
unknown, and overcome remaining barriers to openness. Openness is more than competition. Competition means the ability of companies to contest each other
and to seek customers’ business. This
can result in efficiency and enhanced consumer welfare. But it can also result in a competition
among bundled product packages instead of competition on a product-by-product
basis. Openness, on the other hand,
means the ability of competitors to access consumers directly rather through
their own rivals. This is particularly an issue in network industries, and has
been a constant theme of regulatory battles for over a century. In
telecommunications, product and service markets were closed to competitors for
a long time. For example, rival equipment makers existed domestically and internationally
but could not reach customers of AT&T'’ network. Telephone networks were opened first to customer equipment. Then
it was the time of long distance, international service, and network equipment.
Then, local telecommunications. It has
now been partly extended to internet service over cable TV networks. But it has
not yet reached wireless communications.
Here, competition has been fostered but not openness. To the contrary,
most trends of wireless policy have been in the opposite direction. But this is not well understood.
In
consequence, this article takes issue with the basic philosophy of the FCC’s
wireless policy. What makes this policy
disturbing is: (a) that it does not seem to benefit from the positive impulses
of other policy reforms; (b) that it cannot be explained as based on political
constraints, as in the case of broadcast TV; and (c) that it largely focuses on
carriers’ flexibility rather than users’ choice menu.
It
might be objected at the outset that wireless has been proceeding in the US by
leaps and bounds, so why criticize an evident success story? To be successful in wireless has not been
difficult. Wireless is a success story
everywhere in the world. Growth rates
in Cambodia, Bulgaria, and Zimbabwe are phenomenal, too, and not necessarily
due to enlightened government regulation, but rather due to a universal demand
for ubiquitous communications and a supply of advanced signal processor chip
technology. Governments can hardly claim
credit for these developments, though that has not stopped them, of course.
In
the US, if anything, government has been the brake, not the engine. First, the original duopoly system for the
analog first generation took many years to establish, and during that time the
US lost its original technology head start to the Scandinavians, who still
benefit from those early years. It has
been estimated that the delay cost the US $20 billion. This is an underestimate, since it counts
only lost carrier revenues, not lost consumer welfare, productivity gains,
exports, and jobs. In Finland, meanwhile, Nokia accounts for over one quarter
of all national export earnings.
When
it came to the second generation of mobile communications, US policy was again
painfully slow and complex. Some of
this cannot be helped, being due to a lawyer-saturated environment, but partly
it was self-inflicted. The reason was
the creation of a system of spectrum allocation of mind-boggling complexity,
whose aftermath is still winding its way to the Supreme Court, almost a decade
later. This system centered on spectrum
auctions with upfront payments, a system much beloved to game-theorists free
background in the wireless environment, by property-rights ideologues, and by
government officials eager to fill the empty coffers of government with
windfall revenues. They sold long-term assets for current consumption. They sincerely believed that they had
created an efficient form of allocation when in reality they established an
extraordinary form of entry barriers to new entrants, and an upfront entry tax
on established companies which were given the choice, in effect, to pay or
die. I have discussed this in greater
detail elsewhere.[1]
The
creation of this system generated costs of delay far in excess of any
efficiency gains. It permitted most of
Europe and Japan to forge further ahead in technology, applications, and
consumer satisfaction. Many millions of
i-mode or short messaging users roam the streets of Japan and of Europe, but
not of the US. This is not to claim
that the US is a backward country when it comes to wireless. It is not. But one
must note that in almost all other fields of communications the US is heavily
dominant. Why not in mobile wireless?
This is, after all, a which should be a natural field of leadership, given the
American love affair with mobility and dependence on the automobile. The one different variable is policy. The
one pleasant exception is the emergence of a standard –CDMA—that has been a
real advance. But even this case
benefited more from US defense technology than communications policy.
In
the now emerging third generation of wireless, US policy is again slow and
uncertain. Less spectrum is allocated
in America for cellular use than in Europe or Japan. Allocation of spectrum has been a near-farcical process of
bargaining among entrenched industries and bureaucracies. Fortunately, the
Europeans and Japanese have encountered problems of their own which permit us
to pretend that we have engaged in a process of grave policy deliberation,
instead of simply being unable to get our house in order. It should also be noted that one of the main
problems Europeans have encountered is due to the auction with upfront payment
process, a successful US export that had received the eager attention of
European budget officials.
Each
of these setbacks can be explained.
Collectively, they raise the question if we are proceeding with the
right strategy, or whether we have the fundamentally wrong approach. It is rare
to find European telecommunications policy being more pro-openness and
pro-consumer choice than American one[2],
but this is the case for wireless communications.
American
telecommunications and information policy has been at its strongest when it
focused on consumer choice and on the lowering of entry barriers. This translated to a willingness to let
control over communications shift from the core of the network to the
periphery, and for the core of the network to be competitive. The internet is the classic manifestation of
this philosophy. Its success, in
contrast to government-sponsored, centralized, PTT-driven videotex operations
such as the Minitel, BtX, Captain, Prestel, etc, has demonstrated the
fundamental strength of this model.
It
is therefore regrettable that it appears that the FCC has not applied the
lessons from past successes to wireless, too.
But it is never too late. A new
crew is at the oars and tiller, and it might take a new look before it becomes
responsible for yet another $20 billion plus in foregone future benefits.
The Problem of Vertical Integration
Mobile
communications are becoming the front-line communications device for most
people. Ubiquitous, always on. In the attacks on American cities on
September 11, it was used from airplanes, from under the rubble, as a substitute
for congested landlines. Already, over 120 million Americans subscribe to
cellphone service, almost as many as to wireline telephony. It will soon overtake the latter, as it
already has in several countries.
Wireless is moving into internet access, transactions, and media
content. It is becoming too important
to ignore.
American
cellular communications have ceased to be an open entry field. Carriers have consolidated to 6 national
footprints – AT&T, Cingular, Nextel (in regulatory terms not quite a
cellular provider), Sprint, Verizon, and VoiceStream. These companies use, between themselves, 5 different standards,
one of them the analog AMPS that is on its way out; and four digital standards
--TDMA, CDMA, GSM, and IDEN. These
carriers are widely expected to further consolidate into four firms. They are likely to continue using several
different standards, with only limited transferability of handset equipment
among carriers.
The
major problem with the emerging wireless environment is that it is vertically
integrated in ways that have become unthinkable in other media. Could one imagine a telephone carrier that
can limit user access only to its own internet portal that can select the
accessible websites that can control the type of telephone equipment its users
are attaching, and the software that these users are downloading? These
limitations have not been particularly noticeable in the past, where cell
phones could be thought of as some kind of advanced cordless phone for the car. But cell phones are now becoming much more
than that, and for more like computer terminals on then go, and for more
people.
The
main characteristic of the wireless business is that, the customer is a
contractual subscriber who is served vertically by a wireless carrier that
provides a full set of services. The basic components of a wireless operation
are graphed in the following:
Graph 1.
The
user reaches his carrier via his terminal (handset) and the carrier’s allocated
frequencies. The call is then routed via
a fixed network to the recipient. More
recently, this has been extended to route internet-style communications to a
wireless portal that links into transactions and content.
The key component to this system is the spectrum allocation. It enables the carrier to control downstream the terminal equipment and access of a subscriber, and leverage this position of “owning the customer” upstream to the other steps of this chain.
In
consequence, we are quite used to the notion that the carrier:
·
Selects, markets, and
approves the customer handsets and connects it to its network
·
Provides, selects, and
adopts many of the features, capabilities, and content resident on the handset
·
Operates the wireless
portion of the communications path
·
Operates or provides the
local fixed line distribution
·
Operates or selects the
long distance and international carrier
·
Selects, for areas in
which it does not provide service itself, a partner mobile carrier that
services the subscriber, at rates negotiated and billed by itself
·
Provides
software-defined functionalities on the network
·
Selects and approves
services resident on the network and provided by itself or by third parties
·
Controls the access to a
wireless portal, its content and features, of the providers linked by that portal,
and of the placement of these links
There is nothing wrong with a carrier offering all of these components in a bundled fashion. However, when there is no alternative to taking the entire bundle or major parts of it, several problems are inevitable. One can readily recognize good old friends, issues that have bedeviled fixed line telephony and cable television:
- The reduction or lack of customer choice in applications and content inherent in a vertical integration with no or limited alternatives
- The reduction in innovation of service provision due to the closed nature of the applications and software that can be offered by third parties
- The absence of choice for customers to use, where more advantageous, alternative wireless arrangements are possible, such as wireless LANs, other carriers for roaming, or stronger signals of another carrier
- Market power with respect to vendors of m-commerce, and requirements on such vendors to become business partners
- Selectivity over content, which would be particularly troubling if the wireless medium would become a mass medium with video, audio, and text
- Restrictiveness in the inter-carrier transfer of instant messaging.
These problems will now be analyzed in greater detail.
The Problems of Wireless Non-Openness
- Reduction of Choice among cellular
service providers
Right
now, cell phone users enter into a service agreement with a single
carrier. That carrier accepts all of
their calls, or reaches them in the case of incoming calls. Where the user is outside the service
territory of the carrier, the user gets serviced by another carrier in a
“roaming” arrangement. The
roaming-partner carrier is selected by the primary carrier in a commercial
agreement (a “preferred” roaming arrangement).
The call could also be picked up whichever carrier is around (a
“general” roaming partner, typically a set of carriers, with prices set
industry-wide). The third type of
arrangement is based on signal strength, where the roaming goes to the
strongest signal in that area, unless there is a primary or general roaming
agreement, which would override.
Whatever the arrangement, the user has no choice in the matter, in
contrast to the arrangement in GSM countries, where a caller can select the
roaming carrier and override its primary carrier’s choice.
In
the US, this choice is further limited by the different wireless protocols used
by carriers. A user of a carrier
operating on the CDMA standard cannot roam, in technical terms, on a TDMA or
GSM carrier. A limited number of handsets
can use both TDMA and GSM since they are related. But on the whole, the ability to switch to a carrier using
another standard is minimal. In
contrast, in GSM countries users can easily take their handset to any other
carrier.
Furthermore,
it is impossible to subscribe to more than one carrier using a single
handset. For example, if a user spent
much of his time in both New York and Atlanta, and no company serviced both
cities, he might want to subscribe to companies in both cities rather than pay
expensive roaming charges. However, there is presently no practical possibility
to switch between two carriers. In
theory, something exists called “dual NAM” that would permit dual-carrier
subscriptions. In practice, however,
phone inquiries to several major carriers did not reveal the availability of
such arrangements. This contrast with
the situation prevailing in GSM countries, where user can have the “SIM” cards
of several carriers and inserts one of them into the handset when she wishes to
use that carrier.
Also
in theory, a reseller or reseller group could resell the services of more than
one carrier or service type. This assumes that permission would be granted by
the carriers whose service is being resold, which is not likely if they refused
to permit such choice for their direct customers.
This lack of choice has real implications. Roaming calls are quite expensive, and are not part of the subscriber’s “bucket” of minutes. They are a major moneymaker for carriers.
The
main problem here is not technology but resistance to competition. Once a user can switch freely among
carriers, where will it end? A user
might regularly drive through some areas where the signal of his primary
carrier is missing, and then select another carrier that performs better. Next,
a user might switch to a carrier who offers her the lowest rate during that
time period. Soon, the user would be
able to engage in “least cost routing”, LCR, as in “always best connection”
(ABC). This means that there might be
automated competition for every call, as opposed to the present system of
competition for the subscription.
- Absence of Choice Among different
Wireless Services
In the past, cellular phone service constituted an end-to-end service, separate from those of others. However, other wireless services are also being offered. Paging has long been a widespread service, and smart paging via narrowband PCS has gained increasing popularity. An example is the BlackBerry pager for always-on email. Some such services are being offered on cellphone terminals, but only using its cellphone frequencies, as opposed to being able to switch to the service provided by another paging company. Furthermore, a cell phone terminal could conceivably be used as a terminal for a cordless phone at home or at the office, directly without going through the wireless network. Similarly, it could be used as a “walkie-talkie” between several other cell phones in a neighborhood, again without going through the actual network. (This is a popular feature provided by Nextel for its own subscribers). It could be a terminal to the type of data services pioneered by Ricochet. The cell phone terminal could also bypass the wireless network through wireless local area networks (WLANs). Or, the cell phone terminal could be used as a radio receiver for broadcast programs, a scanner for police frequencies, an advanced pager, a ham radio, marine radio, etc. It might be used in a peer-to-peer fashion, by passing carriers altogether. It is time to think of what we now call the cellphone handset as a future general multi-purpose, multi-purpose wireless terminal. Not as an end point of a specific wireless networks but as the starting point of use applications, using whichever wireless system fits best.
Such multi-purpose terminals would be a threat to most cellular carriers. To see that, let us consider the case of wireless Public and Private Wireless LANs that are emerging as so called “hot-spots” on college campuses, airports, office parks, coffee house chains, apartment house complexes, and planes and trains. (see Bertil Thorngren). These networks, operating on unlicensed spectrum, already reach wireless speeds of up to 11 Mbps two-way communications, and can service, in principle, any type of wireless device, whether laptops, PDAs, pagers, or mobile phones. They follow the 802.111.b standard advanced by Apple, or the Bluetooth standard whose range is more limited, on the emerging HyperLan2 standard. In 2001, there were over 3,500 such hotspots around the world, and estimates for the year 2006 are as high as 110,000.
These WLAN’s are expanding into short-range “home networks” as well as wide area wireless internet service providers (WISPs). These advantages are cheap and easy installation, use of unlicensed spectrum (i.e., without the cost and delay of a licensing process), and flexibility to change to the next level of technology. Entry barriers are low and could include hotels, colleges, airports, shopping malls, etc. Disadvantages of WLANs are lower security, the need to coordinate billing and roaming, and the low staying power of new entrants. (Two early entrants --
MobilStar and Ricochet, have gone out of business. (The latter aimed at a national coverage.)
At the same time, the cellular carriers’ 3G plans are also being contested from below, from upgrades in the second-generation technology known as GPRS, EDGE, and others, generally called “2.5 G.” These technologies raise the data rate for mobile operations to speeds not greatly lower than those realistically expected by the third generation UMTS.
Hence 3G operators are in a bind: they often paid high prices for their new licenses; their average use revenues (ARPU) is lower than in the past due to competition; and their new data business might be crippled by a combination of WLANs and 2.5 G. Such a combination with the flexibility of software-defined radio technology and unlicensed spectrum, might give use to the next generation of wireless – “4G” – that would be characterized not so much by superior technology but by more flexible one.
Thus new services are a major threat to carriers planning to offer 3G services. In Europe, these carriers are spending vast amounts of money for licenses and infrastructure, yet the most lucrative markets might be cream-skimmed by low-cost, unlicensed, high-speed WLAN providers. It is not too much to state that the very survival of some of the traditional carriers depends on making their 3G investments a success. Their strategies will include offering WLANs themselves. But where that market is open and unconstrained, competition by new WLAN entrants will be a major threat.
The first defensive strategy would be to induce the government to raise the entry barriers by requiring the entire panoply of spectrum licenses, upfront auctions, and national/regional bidding instead of local ones, to raise cost and induce delay.
The second defensive strategy would be to reduce the ability of users to connect into WLANs and to partner in build up and roaming. This would be done by preventing users to connect their wireless devices into both the carriers’ cellular networks and into the WLANs. They could accomplish this by refusing to connect such multi-mode equipment to their networks.
Here, until very recently, the FCC was on their side. For a variety of reasons, the FCC refused to license equipment that could function for multiple different services. As Michael Marcus observes, a ham radio could not be used as a marine radio, etc. As it happens, however, the FCC has reversed this course that had been embarrassingly at variance with its after-lunch rhetoric of flexibility. On September 13, 2001, the FCC adopted rules on software-defined radios (SDRs).[3] These rules have moved a great step forward in flexibility. They permit radios whose operating parameters are determined by software—which could be almost any radio these days, if the rules are interpreted generously-- to operate on multiple standards and services. They also separate the radio transceiving RF functions from applications software, thereby enabling flexibility on the applications side.
This initiative has been much to the credit of the FCC. It has the potential to subvert the present carrier-centric regulatory approach. With software-defined radios receiving legal backing, multi-service transceivers will inevitably emerge. Under present rules, however, the carriers would not have to connect such equipment. They ultimately control which equipment can access their networks, and the potential for such equipment to be left stranded will retard its development. This will be developed further below.
- Control Over the Approval of Handsets
Reduces Innovation and Choice
At
present, the approval of handsets by carriers and by the FCC is a two- or even
three-stage process. The FCC (and
similar regulatory bodies elsewhere) issues specifications regarding the radio
(RF) and health aspects (SAR)of equipment.
This includes frequency, power, and radiation. Some of this is based on self-certification (by way of verification,
or declaration of conformity after testing by accredited labs). In other cases, involving higher risk
equipment, equipment must be submitted for testing by the FCC or licensed
private Telecommunications Certifications Bodies. There are also issues of non-access to certain frequencies for
transmission. For example, only
licensed pilots can buy transmission equipment for aviation frequencies.
At present, the approval of handsets by carriers and by the FCC is a two- or even three-stage process.
These stages involve the air interface standards that govern the transmission from the handset to the base station, such as CDMA (technical standard IS 95), TDMA (IS 136), I-Den, and GSM. These standards are set by a variety of manufacturer-driven groupings. The decision whether to approve a particular handset for connectivity, however, lies within the discretion of the carrier, since that carrier is entirely free, in the US, made the FCC’s PCs tales to select its standard. In Europe, in contrast, any equipment that complies with the GSM specifications will be connected to the network. There is no carrier discretion. In the US, the industry association CTIA often certifies a manufacturer’s equipment to the industry, but each carrier can add its own requirements and flavor of specifications. In consequence, large carriers also test and approve equipment for connection to their network. Hence, the mere adherence by a manufacturer to the standard specifications in the US is not enough. It must also find favor with the carrier. There is no right to use equipment to connect to a cellular network.
The carrier’s business calculus on what equipment to approve is based on a variety of factors. Since in the US, in contrast to Europe or Japan, the carrier rather than the consumer buy most handsets, low cost is a major factor, as would be serviceability, ability to maintain a limited inventory, and independence from a single source. In addition to reducing the choice available to users, this system also makes manufacturers somewhat dependent on large carriers. The handset makers also tend to be major suppliers of network equipment. They would not lightly put used equipment into the marketplace that would be disfavored by the carriers as threatening their basic business by facilitating access to services such as WLAN that compete with the business of their best customers.
- The Usefulness of User Equipment is
Limited by Closed Operating Systems Software
Beyond the question of whether multi-services equipment can be licensed and connected is the question of control over the nature of the terminals themselves. As handsets become smarter, they begin to resemble small computers. To function, they now begin to incorporate operating system software. At present, there are four initiatives in that direction. Participants include:
· Microsoft (not surprisingly), with its Stinger, in cooperation with Samsung and Sendo, a UK-based handset maker.
· Symbian, a joint venture of most cellphone makers resisting a Microsoft dominance, with its Epoc. They include Ericsson, Nokia, Motorola, Matsushita, Sony, Psion (software), Kenwood, and IBM.
· eCos. This is one of several Linux-based open-source operating system concepts, this one from Red Hat, 3G Labs, and others. At present it is more of a concept than a reality. Manufacturers other than Qualcomm have been reluctant to engage in this non-proprietary environment.
· J2ME. (Java version 2, micro edition). This concept, too, encourages openness, though it is not non-proprietary in the way that Linux is.
The important thing to understand
is that the choice of the operating system pre-determines the openness of the
applications software. With Stinger or
Epoc, the carriers and manufacturers would be able to control the types of
software applications loaded onto the handsets. With eCos and J2ME, they could not. Hence, open approaches would be preferred by those carriers eager
to promote a great diversity of applications.
They would favor manufacturers adopting these operating systems. Others,
perhaps more concerned—legitimately-- with quality control of applications,
security against viruses, and with the increased bargaining strength it gives
them over applications developers and third-party applications service
providers, would not favor equipment that is open to software, and hence would
favor manufacturers who select the closed operating software.
As wireless networks begin to offer increasingly higher-level services, the question of who may load what applications onto a handset, and what network-based service interfaces these applications may access becomes important. Is a user restricted to only the applications that are offered by his primary service provider, or may he load other applications? Furthermore, can these applications have full access to the functions of the network and the handset? These questions may be best illustrated with three examples, due to Kevin Kahn.
1. Suppose a brokerage wants to offer a handset application that uses the screen and the alerting (sound or vibration) capabilities of a handset to provide a service to its client. This application requires that the code at the handset have access to the APIs that access the handset functions and it also requires that the service have access through the network to get messages to that application. Does deployment of such an application require the cooperation of the wireless service provider? In today’s Internet, deployment of such an application on an end system PC would not require any support from the ISP – will the same be true for wireless?
2. As a second example, suppose a company wishes to deploy a universal messaging application that can alert users to any messages, email or voice, that they have on any of a number of message queuing services. This application requires access to the alerting indicators of the handset and it may also require access to the voicemail service inherent in the wireless providers service so that messages queued there can be included in the new service. Do the necessary APIs and addressing paths exist to allow such an application.
3. As a final example, suppose a vending machine with an infrared or Bluetooth interface that wants to interact with a handset to communicate with a backend billing service to handle the vending. This requires a digital certificate to be sent from the machine through the handset to the backend service followed by an authorization certificate being sent back via the same path to the machine. Again, this requires the application to have access to specific functions of the handset (probably the IR port, the screen to present selections, and the buttons to select). Can this application be deployed without cooperation of the wireless service provider?
- Closed Portals Reduce User Choice
Under
the presently evolving system, the user reaches a wireless portal, whence she
can be connected to a variety of other sites.
The selection and placement of these links, however, is under the
control of the carrier. Other portals
might be accessed, but that requires additional clicks. This situation is very similar to the one
discussed for cable television’s access to portals other than those of the
cable company or its partner. It has
led, in the US, to requirements on Time Warner Cable to open its system to several
service providers/portals. Similar
rules are being contemplated for the entire cable industry. These issues are well known and require no
recapitulation. Virtually the same arguments on both sides apply also to cell
phone access to portals, and through them, to the broader internet. It should
be noted, however, anticipating the conclusion of this paper, that they are
much easier to resolve to the wireless medium.
- Transactions and Content are Limited.
The
wireless carrier’s portal is not a common carrier. Hence, the selection of
websites, e-vendors, and content providers is entirely that of the
carrier. Its selection would be based
on its own economic, cultural, and political considerations. Being a selector, it would also incur some
legal liability, which would further increase caution.
Even where there might be a wide set of links, such as in the case of Japan’s NTT DoCoMo and its i-mode portal, such openness can be affected by differentiating policies. DoCoMo has preferential arrangements with a small number of its partner sites, and handles their billing through the users’ phone bill. In contrast, the majority of other sites must arrange for their own billing, putting them at some disadvantage. This provides DoCoMo with some leverage over its primary partners.
The absence of openness resembles the “walled garden” arrangements of some Internet portals provided by cable companies. Correspondingly, we can term this arrangement the “walled airwave” system.
Implications
for Public Policy
The previous section has identified the potential for real problems. But the recognition of such issues does not mean that regulatory approaches are needed. A vigorous competition among mobile carriers could overcome most issues and generate unbundling through market forces. At the same time, the ability to exercise market power with respect to mobile commerce providers or wireless LANs might be common to all mobile providers and more profitable than a more open system. In such a case, market forces might not lead to unbundling.
The knee-jerk response to the problems identified in this paper is that competition will take care of it. But suppose that carriers would be consistently worse off by offering consumers the choice of moving easily around to other carriers or service providers. Such competition would reduce prices and profitability. It would, on the other hand, grow the market. But it is quite likely that each carrier would be better off servicing a less competitive slice of a smaller market, rather than engaging in greater competition in a larger market.
It is not clear why a carrier A would be the first to offer such choice to its customers. After all, it would provide an exit to its own customers, without a potential compensating gain from the customers of the other carriers B and C. The main reason would be to hope for enough users of B and C to switch their subscriptions to A in order to have the choice of not using A. This can hardly be a strong selling point. Furthermore, any choice requires the consent and cooperation of B and C, which might not be forthcoming once they realize that they are opening the door to a mutually destabilizing competition. They will be concerned with reputation effects if they are blamed in users’ mind with poor performance caused by an element not under their direct control. And they might be able to use bundling as a way to price discriminate, as George Stigler has pointed out in a different context. The likelihood of oligopolistic behavior within a small group of carriers is high. As the number of competitors shrinks, each has less to gain and more to lose by maverick behavior. It is also an inhibitor for any software developer to take initiatives for new applications if the market is largely closed, and this further reduces the attractiveness of any non-conforming behavior be a carrier.
Where market forces do not work, would regulation? Let us look at several potential points of intervention and evaluate their need.
A schematic view of an unbundled wireless network environment is provided in Graph 2. It shows, at each stage of the chain of wireless provision, alternative providers. We will consider “openness” at several points, and conclude that only one of them – the openness of the terminal equipment to access multiple providers of wireless services and providers-- is critical. A subsidiary second opening –unlicensed spectrum—supports such policy. This is discussed in the following.
Graph 2
I. The Separation of the User Equipment (UE) From the Carrier
Such a policy would amount to a ‘Carterfone’ policy for users’ wireless equipment. Following that decision, the FCC permitted users to attach equipment chosen by themselves to the telecom network. While the carrier could still offer and market its preferred equipment, it could not exclude other equipment, as long as it conforms to certain technical specifications pertaining to the RF transceiving function and non-discriminatory industry specifications for air interfaces standards. These specifications could not close equipment third-party applications or access to other network protocols offered by other types of providers, as long as it conforms to the FCC’s software defined radio rules.
While a fully bundled service could be offered by a carrier as before, the carrier could not prevent a user of selecting, for any given call, another wireless service provider or use the equipment for other communications purposes.
The significance of such arrangement is that equipment will be offered by the market that adds features, and, more importantly, permits a user to select service providers depending on circumstances. For example, a user in a shopping mall, campus, office building, or airport could connect to a wireless LAN. A user encountering a circuit busy could switch to another carrier. A user seeking to receive synchronous music, radio style, could do so by accessing a specialized broadcaster.
This choice would reduce the need for most other access requirements, since the user would not be tied to a single carrier with significant costs of switching to another. This is partly embodied in the GSM standard which provides some user selectivity over carriers, though the approval of such alternatives remains with the primarily carrier, which also handles the billing.
This approach would be similar to that adopted by the FCC for CPE following the Carterfone decision in 1968. It followed Cassandra warnings of impending network chaos, but has worked spectacularly well.
- Access to Unlicensed Spectrum.
The
key source of leverage for carriers is the high entry barrier for new and
future entrants in service provision, due to the spectrum auctioning system
with its advance payment feature. Given
the difficulty in freeing additional spectrum and the high cost of acquiring
it, it seems unlikely that there would be new entrants emerging to challenge
the reduced group of carriers. Therefore, government should additionally
provide adequate spectrum on a license-free basis, with users and service
providers paying for usage rather than for ownership, in the way that
automobiles pay for the use of highways. This has been developed in detail by
the author in other papers.[4]
Once such spectrum is available, and once users’ terminals can access service
providers such as WLANs operating on such spectrum, users will not be
constrained by the limited choice of maybe four cellular carriers that could
still collectively be restrictive.
- Access to Alternative Wireless Portals.
The
third access issue is that to the wireless portal. The issues here are similar to those discussed for the cable
industry. The similar arrangement would
mean that the wireless carrier would let the user pre-select its primary
portal, or that several such portals would be accessible at no extra effort, or
that the two upper layers of the carrier portal would be open to third
parties. This approach would mirror the
open access of the internet, and the approaches now being applied to AOL Time
Warner, and considered by the FCC in its proceedings.
Content openness may be the easiest type of openness to consider since it is essentially a browser level openness. The question can be reduced to whether the user can enter an arbitrary URL to a network portal to access content (independent of any business deal between the wireless provider and particular content providers) and whether browser plug-ins can be created and downloaded to render the resulting content if required. This issue is analogous to the walled garden discussions that have occurred in the wired internet.
This problem would largely go away if the users could access, through their handsets linked to other carriers and wireless providers, also other portals and websites.
- Openness of the carrier’s network
The fourth element of openness relates to services offered by third parties and requiring presence in the wireless network itself. The options are either to keep wireless networks closed to third parties, or total openness, resembling a common carrier access for third party software applications, or a type of equipment collocation that exists in telecommunications. Here, too, the ability to access alternative wireless carriers through flexible handsets would be enough to deal with this issue.
Conclusion
The focus of the FCC policy has been to provide carriers with choice—in the utilization of he licensed frequency, in the technical specifications of it service, in its pricing, etc. There does not seem to have been a similar orientation towards choice of the users, broadly defined as consumers and providers of various attached services. The implicit notion was that by providing carriers with options, and creating competition, users will be well served. And that certainly goes a long way. But carriers are likely to resist offering consumers the choice of moving easily around to other carriers or service providers. Such competition would reduce prices and profitability.
The conclusion of the analysis is that the key point of openness, and arguably the only one needed, is that of openness of user equipment. With this openness achieved, the user would have alternative avenues to spectrum, content, portals, applications, software, etc. A secondary policy would be to assure alternative wireless pathways such as WLANs by providing an adequate amount of unlicensed spectrum.
Why is all this important? The overall goal of the openness approach described above is to establish for the wireless environment the same dynamism shown in the internet with it’s open access terminals, especially the PC, encouraging hardware and software innovation and applications. Right now cellular telephony is a dynamic sector, mostly based on the growth of penetration. Soon, however, this growth will plateau as universal wireless connectivity is being approached. At that point, we need the impetus for further innovation that a more open system provides. For the carriers, the overall positive impact in terms of traffic generation may well outweigh some loss of control. For users, service providers, and technology developers, the advantages of openness might be significant.
American communications policy has fared best when it puts its faith in the dynamism of the periphery of the network, instead of seeking to strengthen the ability of the network core to dominate. Wireless is no exception. And the mediocre results of policies focusing on the core, in contrast to those for other parts of the communications environment, suggest that a reorientation is in order. The key step now is to follow the opening set by the FCC’s for software defined radio by a Carterfone-style opening to equipment that can access multiple wireless networks. With it we can leap frog the “3G” model with it’s carrier – orientation to a “4G” model patterned after the internet
Acknowledgements
I am grateful for the help and comments received by:
James Alleman, Bob Atkinson, Ron Barnes, Brian Bebchick, Kenneth R. Carter, Kathryn Condello, Terry Hsiao, John Lee, Don Nichols, Michael Noll, Michael Marcus, Bertil Thorngren, John Williams, and to Charlie Firestone and the Aspen Institute’s Regulatory Policy Meeting, especially Kevin Kahn and Robert Pepper. Views expressed here are entirely my own.